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

[Proponent]

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.

Could it be that the helium supply is very cold -- below the melting point of oxygen -- and that net expansion and heat leakage were expected to warm it above the melting point, but, during at least some part of the process the heat leakage was less than expected?

[LouScheffer]

Thinking of how the COPV could get cold enough to freeze LOX.

How do they generate the high pressure, cold helium to fill the tank?  They could start with a tank of helium gas, compress it, then cool it, then load it.  The most natural cooling would be with liquid nitrogen, but this would not be cold enough to freeze LOX.

Alternatively, they could have a tank of liquid helium, with an outlet at the bottom.  Then shutting the vent valve (and heating some helium if needed) would pressurize the tank, causing the liquid helium to push out the outlet at the bottom.  This requires no pump and no cooling, so it's mechanically very simple.  This is exactly how NASA pumped liquid hydrogen from the giant storage sphere into Saturn rockets and the Space Shuttle. Presumably the helium would warm up and evaporate on its way to the COPV, or in the COPV if it managed to remain liquid that far. The final temperature in the tank, as a function of time, would depend on the specfic heats of the materials, the fill rate, and whether the tank was submerged in LOX.  But in any case it could still be very cold, perhaps cold enough the freeze LOX.   

In this scenerio, loading the tank more quickly would result in a lower minimum temperature reached by the tank during filling.  This could potentially explain why this failure mode has not been seen before, and how such a fault could be triggered by a change in loading procedures.  This is all speculation, but seems to fit the facts we know.

[Jim]

Thinking of how the COPV could get cold enough to freeze LOX.

How do they generate the high pressure, cold helium to fill the tank? 

In high pressure tanks (6k-10k psi) on railroad cars at the base of the pad.

http://ww1.hdnux.com/photos/51/45/04/10899724/5/920x920.jpg

[Bubbinski]

Would sensors or cameras be able to detect frozen blocks of LOX on the COPVs and warn the controllers for future fuelings?

[Kaputnik]

Would sensors or cameras be able to detect frozen blocks of LOX on the COPVs and warn the controllers for future fuelings?

I think the theory is that the putative lox ice (lice?) crystals would form between the liner and the overwrap material, so it would not be visible from outside.

[LouScheffer]

Thinking of how the COPV could get cold enough to freeze LOX.

How do they generate the high pressure, cold helium to fill the tank? 

In high pressure tanks (6k-10k psi) on railroad cars at the base of the pad.

http://ww1.hdnux.com/photos/51/45/04/10899724/5/920x920.jpg

The simplest possible approach to helium loading would be to just let ambient temperature helium flow into the tank from the high-pressure railroad cars, then let the LOX cool it.  However, it's not clear that the helium in the COPV tanks could cool quickly enough.  For hydrogen tanks as envisioned for cars, thermal equilibrium with the tank can take minutes, and these tanks are bigger.  This would be no problem if the LOX and helium are loaded earlier in the count, but might not cool the helium enough when LOX is loaded at the last minute.  That's why they might load cooled, instead of ambient, helium.

Also, if they load at ambient temperature, the LOX is cooling the helium, so it will never get colder than the LOX.  So how could Elon have suspicions about LOX crystals?

This is why I suspect there is more to loading than just opening a valve to the external helium tanks.

[Comga]

(snip)
Also, if they load at ambient temperature, the LOX is cooling the helium, so it will never get colder than the LOX.  So how could Elon have suspicions about LOX crystals?

Helium above 55 K has a negative Joule-Thompson coefficient.
It is a struggle to understand the implications, which may be that as the He is compressed, it cools.
Can someone with real knowledge and understanding of thermodynamics weigh in on this?

And Jim, are you saying that the He COPV's are filled passively, with just the high pressure in the storage tanks forcing it in?  That would seem to be very inefficient.  What are the relative pressures?

[Jim]

And Jim, are you saying that the He COPV's are filled passively, with just the high pressure in the storage tanks forcing it in?  That would seem to be very inefficient.  What are the relative pressures?

That is the way all gases are loaded on to launch vehicles.  There are no compressors on the launch pads.

[duh]

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)

There haven't been 20 launches with subcooled LOX.

True, but there have been over 20 launches with this design(as far as we know). So, possibly not a design fault, but perhaps the wrong design. But I'll wait for SpaceX to determine cause before claiming anything concrete.

Came across an interesting (to me at least url)

   http://www.reliasoft.com/newsletter/v8i2/reliability.htm

that touches on differences between quality and reliability. I found figure 2 to be of particular interest, including thinking
about it in terms of this "incident" on the pad.

There have been other thoughts expressed in this thread that I could not readily find at the moment that it seems to
me that this url could also relate to.

Just my mythical 2 cents worth. Your mileage may vary. (YMMV)

[scubadown]

Thinking of how the COPV could get cold enough to freeze LOX.

How do they generate the high pressure, cold helium to fill the tank? 

In high pressure tanks (6k-10k psi) on railroad cars at the base of the pad.

http://ww1.hdnux.com/photos/51/45/04/10899724/5/920x920.jpg

The GHe railcars are more like 3500 psig, they may have been de-rated though as they are very old and I think no longer DOT rated. They were used for years on Shuttle before we switched to a LHe system. We would have to have around 25 railcars on the siding of the CCF to start the launch countdown process (enough for three launch attempts).

The process we used in Shuttle was to reduce the railcar pressure to about 300 psig and then pump it up to 6000 psig using Joy and Henderson compressors. Once the railcar was at 300 psig we would switch to a full one and start the process over.

[MarekCyzio]

https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19990069948.pdf

Interesting NASA document on LHe and GHe infrastructure in KSC/CCAFS

[CyndyC]

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

Right after the CRS-7 anomaly, someone on SpaceflightNow commented that Helium is very tricky to work with, that it can leak through solid steel.

A comparison of densities is in the table below, mass in kg being "a measure of the amount of material in an object, being directly related to the number and type of atoms present in the object." It's easy to see how infinitesimal the helium atom is compared to the others. All measurements except for liquid helium are from the same website.

Helium Gas·	0.1664 kg/m3	at NTP or Normal Temp & Press = 20oC (293.15 K, 68oF) and 1 atm
·	0.1785 kg/m3	at STP or Standard Temp & Press = 0oC (273.15 K, 32oF) and 1 atm
Helium Liquid·	147 kg/m3	at 4 K (physics.info)·
·
Aluminum·	2712 kg/m3
Titanium·	4500 kg/m3
Stainless Steel·	7480-8000·
Steel·	7850 kg/m3
Carbon Epoxy Comp·(61%)	1600 kg/m3

http://www.engineeringtoolbox.com/metal-alloys-densities-d_50.html

[Comga]

And Jim, are you saying that the He COPV's are filled passively, with just the high pressure in the storage tanks forcing it in?  That would seem to be very inefficient.  What are the relative pressures?

That is the way all gases are loaded on to launch vehicles.  There are no compressors on the launch pads.

In high pressure tanks (6k-10k psi) on railroad cars at the base of the pad.

http://ww1.hdnux.com/photos/51/45/04/10899724/5/920x920.jpg

The GHe railcars are more like 3500 psig, they may have been de-rated though as they are very old and I think no longer DOT rated. They were used for years on Shuttle before we switched to a LHe system. We would have to have around 25 railcars on the siding of the CCF to start the launch countdown process (enough for three launch attempts).

The process we used in Shuttle was to reduce the railcar pressure to about 300 psig and then pump it up to 6000 psig using Joy and Henderson compressors. Once the railcar was at 300 psig we would switch to a full one and start the process over.

These seem to be conflicting expert opinions.
Which way is it?
Is there a subtlety where both are true?

[virnin]

And Jim, are you saying that the He COPV's are filled passively, with just the high pressure in the storage tanks forcing it in?  That would seem to be very inefficient.  What are the relative pressures?

That is the way all gases are loaded on to launch vehicles.  There are no compressors on the launch pads.

In high pressure tanks (6k-10k psi) on railroad cars at the base of the pad.

http://ww1.hdnux.com/photos/51/45/04/10899724/5/920x920.jpg

The GHe railcars are more like 3500 psig, they may have been de-rated though as they are very old and I think no longer DOT rated. They were used for years on Shuttle before we switched to a LHe system. We would have to have around 25 railcars on the siding of the CCF to start the launch countdown process (enough for three launch attempts).

The process we used in Shuttle was to reduce the railcar pressure to about 300 psig and then pump it up to 6000 psig using Joy and Henderson compressors. Once the railcar was at 300 psig we would switch to a full one and start the process over.

These seem to be conflicting expert opinions.
Which way is it?
Is there a subtlety where both are true?

If you read the PDF posted above, it indicates the compressors are centrally located, not installed at each LC.  There were only 5 Joy compressors and 3 Henderson compressors to cover all pads at KSC and CCAFS.

[scubadown]

And Jim, are you saying that the He COPV's are filled passively, with just the high pressure in the storage tanks forcing it in?  That would seem to be very inefficient.  What are the relative pressures?

That is the way all gases are loaded on to launch vehicles.  There are no compressors on the launch pads.

In high pressure tanks (6k-10k psi) on railroad cars at the base of the pad.

http://ww1.hdnux.com/photos/51/45/04/10899724/5/920x920.jpg

The GHe railcars are more like 3500 psig, they may have been de-rated though as they are very old and I think no longer DOT rated. They were used for years on Shuttle before we switched to a LHe system. We would have to have around 25 railcars on the siding of the CCF to start the launch countdown process (enough for three launch attempts).

The process we used in Shuttle was to reduce the railcar pressure to about 300 psig and then pump it up to 6000 psig using Joy and Henderson compressors. Once the railcar was at 300 psig we would switch to a full one and start the process over.

These seem to be conflicting expert opinions.
Which way is it?
Is there a subtlety where both are true?

Well I don't know for sure, but the photos I've see of the railcars at the base of the SpaceX pads look exactly like the ones we used for Shuttle. They are DOT rated and have to be re-certified periodically or they can't travel over the rail lines. I think SpaceX uses them now for fixed storage but my guess is they had to de-rate them from DOT ratings to ASME Section VIII ratings to meet CCAFS safety requirements. 

[Comga]

There are no compressors on the launch pads.

The process we used in Shuttle was to reduce the railcar pressure to about 300 psig and then pump it up to 6000 psig using Joy and Henderson compressors. .

These seem to be conflicting expert opinions.
Which way is it?
Is there a subtlety where both are true?

If you read the PDF posted above, it indicates the compressors are centrally located, not installed at each LC.  There were only 5 Joy compressors and 3 Henderson compressors to cover all pads at KSC and CCAFS.

Ah!  The subtlety is that the compressors are remote to the launch pads, a site facility.
Does that high pressure He system extend beyond KSC and is it available at LC-40?
scubadown points to the tanks at LC-40 as evidence that it is not available.

[scubadown]

I do not know if SpaceX is tied into the the HP GHe facility (I think they are as it was a Titan pad). I believe Atlas and Delta are. The CCF is the nexus and keeps the VAB, LC39A, and LC39B storage batteries at a nominal 6K PSIG (I think it has been de-rated slightly). There is a high pressure pipeline running to the CCAFS which ties in their pads to the high pressure GHe system. During launch countdown at any of the tied in pads the CCF is manned and has backup equipment in case of power failure or compressor failure. The large storage batteries at the VAB and LC39 pads and interconnecting pipeline provides probably one of the largest high pressure stored volumes of GHe in the world.

[HMXHMX]

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.

I've had a suspicion for a while (more than a month) about the dissimilar metal joint between the aluminum liner and the outlet tube (which is stainless) that connects to the rest of the helium system. I have heard rumors about this joint being problematic, but aluminum-stainless joints are routinely used in vacuum service without a problem, so I haven't given the rumors much credit.  But if the welding process is not well-controlled, I imagine it could create issues...

I'm told this joint is inertially-welded to the aluminum blank before spin-forming the blank into the thin liner.  I was unclear about the details of the joint until today, when I watched the Cimarron video gospacex posted.  I captured a freeze frame from the video, below. 

I'm mainly putting this out there now in the interests of covering all the bases, since hopefully we should get some report on the investigation prior to the next F9 flight, whenever that takes place.

Edit: grammer

[.Scott]

Helium above 55 K has a negative Joule-Thomson coefficient.

That's interesting.  So, if we assume slight leakage/seepage from the Helium tank, then at 55.5K, this would have no thermal effect.  But at 54.5, it would act to further refrigerate the area around it.

LOX freezes at 54.36K.  I would guess that Solid Oxygen (SOX?) would be permeable to He.  So once this process started, it would not be self-correcting.  Depending on the physical properties of the materials involved, it could be self-aggravating.

----

On further reading of the Joule-Thomson wiki article:  If I read it correctly, a slight leakage/seepage from the Helium tank would actually have a warming effect at 55K.  And useful cooling does not seem to happen until either 51K or 40K?  I need to read some more.

[CyndyC]

Some more fun facts about helium, taken from comments on another website, when SpaceX had a problem with the helium system that delayed the Turkmen/Monaco Sats launch for a week in March of last year, not after the CRS-7 anomaly in June that year as I originally thought.

Inert gas my ass. It does what it wants to, when it wants to, hahaha.

Helium has a very small atomic cross section. It leaks where other gasses will not. Pain in the butt...

(mentioned earlier)

Helium is tricky though; it can leak through solid steel albeit slowly.

The main reason for use of helium in such systems is not that helium is inert (in such case one may use argon, which is chemically as inert as helium, but also much cheaper and easier to handle), or that it weighs less (its mass is negligible compared with mass of whole rocket, meanwhile hydrogen weighs even less than helium), but the fact that helium has the lowest boiling temperature among all known substances. Thus it will surely remain gaseous even at the lowest cryogenic temperatures used in space launch industry, thus enabling pressurization of the tanks.

Of course, there is no need to use helium for kerosene tank pressurization, but once there is a system with helium to pressurize oxygen tank, there is no sense to install a separate one with nitrogen or other gas for kerosene.

Bonus quote - here's some irony, after someone else had called the one week delay "major",

Major delay? Depends on you definition of major. I consider blowing up a rocket and having no alternative for customers for 6 months a major delay. I consider this a minor delay.