A reported quote from Elon (no URL, yet):
We are close to figuring it out. It might have been formation of solid oxygen in the carbon over-wrap of one of the bottles in the upper stage tanks. If it was liquid it would have been squeezed out but under pressure it could have ignited with the carbon. This is the leading theory right now, but it is subject to confirmation. The other thing we discovered is that we can exactly replicate what happened on the launch pad if someone shoots the rocket. We don’t think that is likely this time around, but we are definitely going to have to take precautions against that in the future. We looked at who would want to blow up a SpaceX rocket. That turned out to be a long list. I think it is unlikely this time, but it is something we need to recognize as a real possibility in the future.” [Emphasis mine]
I think that the Arocket list had speculated about this early on.
Solid oxygen? How? That’s almost as cold as LH2. Do they really chill the oxygen that much?
I would find it easier to believe if the hypergolics they use to start up the second stage mixed up by mistake, or if the carbon composite tank had a crack and a leak somewhere.
There is only 10 kelvin difference between their subcooled LOX and it’s freezing point. With weird properties of helium mixed in at these temps.
The speculation on NSF is that cLOX and He flow rates plus He plumbing geometry conspired to create an accidental themoacoustic heat engine. This may have pumped energy away from the bottle, solidifying the LOX between the Al liner and overwrap. As the bottle filled with He the pressure increased between the two and boom.
I’m also wondering what would be cold enough to freeze O2 around the rocket.
USAF has considered it a real possibility for quite a long time now, hence the security at CCAFS.
They’ll also need to order an AFDB for everyone in the ops crew.
So this would be a direct consequence of slushing the LOX, then, wouldn’t it. That doesn’t sound like good news at all.
Perhaps, but perhaps also an avoidable one. They’d been doing it for a while.
Uh, how about coating the tanks with Teflon? Or something like that.
The mass added by the lining would eliminate the advantage of having the Oxygen at supercold temperatures.
I think he means coat the bottle in the tank, not the inside of the tank.
Do you have any idea how close they are to the LOX freezing point?
I dredged up some adiabatic expansion formulae and started looking at how various sized pressure transients would affect tank temperature. The relevant one seems to be:
T2/T1 = (p2/p1)^(γ/(γ-1))
Helium, being monatomic has a γ=1.66, so
T2/T1 = (p2/p1)^2.515
The O2 freezing point is 54.4°K. Suppose the LOX is chilled down to 55°K. Some example values for various pressure drops:
0.1% pressure drop = 0.25% temp drop = 54.9°K (still liquid)
0.5% pressure drop = 1.25% temp drop = 54.3°K (frozen)
1% pressure drop = 2.5% temp drop = 53.6°K (frozen)
5% pressure drop = 12.1% temp drop = 48.3°K (frozen)
10% pressure drop = 23.3% temp drop = 42.2°K (frozen)
Of course, transients are, well, transient, so how quickly a temperature dip gets transmitted to the outside of the COPV is going to be a big imponderable. But the whole purpose of a COPV is to chill the helium to the same temp as the LOX, so I’d think that they’d have pretty decent thermal conductivity.
Here’s the problem: Fairly minor design decisions in the He system (e.g. opening a valve to the tank pressurization and causing a decent-sized pressure drop in the COPVs), which may have occurred back in the pre-F9FT days, can cause pretty big adiabatic changes in the tank temperature. Those probably aren’t much of a problem if your LOX is at 85°K, but it’s a different story when you’re hugging the melt line.
In the long run, it’s an isothermal system, because the LOX will wind up warming the He back to its own temperature. But if the mechanism of that warming involves accidentally tossing some heat of fusion into the COPV, then that would be bad.
On the other hand, you could have a pure operational problem, like pressurizing the LOX tank after chill-down, which might be fixed simply by pressurizing (or partially pressurizing) the tank before chill-down. But then how do you top the lil’ sucker off as the LOX densifies? Or maybe you simply need a slower valve opening to keep the system closer to isothermal.
I’m clearly speculating wildly here, but you can see how there could be some pretty serious re-engineering required.
Following along on your suggestion of potential oxygen freezing, solid oxygen at the melting point has a density of 1.426 and slush oxygen has a density of 1.338, whereas liquid oxygen at the freezing point has a density of 1.23 g/cm^3.
Going by a rough O2 phase diagram I Googled (the Internet needs better O2 phase diagrams), it looks like the transition from solid to liquid oxygen has a slope of about 14.3 MPa per K. So to keep barely solid oxygen solid under a 1 K temperature increase requires about 14.3 MPA (2075 psi) of extra pressure.
So the adiabatic helium pressure/temperature fluctuations could cause solid oxygen to form in a confined space (say a pipe or small void in some fitting, and then you raised its temperature inside while there was still an oxygen freeze plug blocking the exit of the void, the internal pressure would rise dramatically and the required volumetric increase to relieve this pressure would be large (16%) as the inner frozen oxygen melted, possibly causing a rupture or structural failure.
So the next obvious question is whether there is any place in the system where such a trapped freeze/thaw could have occurred.
Hmm. I think I messed up in multiple ways on my comment.
First, I wound up using the reciprocal of the exponent in my temp vs. pressure drop computations, so the temp. drops less than the pressure, not more. (Imagine that! A math error! That never happens to me!)
Second, it looks like the solid form of oxygen is more dense, not less, than the liquid form, so if LOX had infiltrated the overwrap, it wouldn’t be the pressure increase of expansion that would provide the energy to make it ignite with the carbon fibers. It would have to be the LOX pressure itself (which could be from gas and/or hydrostatic pressure).
On top of those two errors, the accident occurred at about T-8 minutes, long before tank pressurization begins–or at least before pressurization with helium begins.
Finally, I found that the LOX is subcooled only to about 66°K, which is well above the melting point, so I’m completely confused.
Bottom line: never mind!
One thing they could do would be to put high-speed (or higher resolution) cameras near the launch site next time so you actually have more data to work with.
Stopping a shooter doesn’t seem very easy. AFAIK the Cape is fenced and the USAF does control entries and exits but it’s not like its impossible to sneak something through.
Never heard of a radar or IR sensor accurate enough to track a bullet. At best there are anti-ATGM systems in the latest generation of tanks which can detect and intercept a projectile but that’s for larger weapons.
Obviously erect bullet proof blast shields around the rocket.
Use some decoy rockets. It would just take a few extra pads and some fake fog to simulate boil off. Ron Howard could consult.
Unless they are considering strengthening the vehicle which will add more weight to it.
I believe this was from a reddit post earlier today. It has since been taken down.
As to that long list of folks who would take a shot at a SpaceX rocket, they should also think about who may not like the payload it was launching. I think that list is longer & more powerful than any industrial saboteur.
Now that the idea is out there and the stakes are so high it almost guarantees a bullet will happen in the future. They need a launch site that they can sweep of shooters.
If they get launch confidence high enough to allow an inland launch site, somewhere in Arizona, New Mexico, or West Texas comes to mind. A few square miles of flat open desert and nothing for a sniper to hide behind.
Launch complexes could be re-structured into silos however at the Cape water seepage due to a high water table would be an issue. Shielding the rocket in a launch cannister erector would be another possibility. Yes this has been and probably will be discussed in A-Rocket. Another possibility that mighr even be cheaper. Strengthening the erector strong-back such that it is capable of supporting the payload should the rocket explode beneath it. We saw that until the strong back weakened it actually held the payload aloft above the rocket for a period of time past the worst of the rocet explosion itself. Thus even if potshots at the rocket are successful the payload is still salvaged.
Even if the payload survived on the strongback, I’m pretty certain it would be a total loss anyways. Vibrations, jostling and just plain cooking would be enough question whether the satellite would actually function as intended.
I’d like to see hard data on that before coming to that conclusion. The fairing itself is an acoustical enclosure designed specifically to reduce vibrations in flight. Thermal issues could possibly cook the payload but you could mitigate with water cannons if they could be made survivable. However a problem remains in that in the current process the strong back is retracted at about T minus 4 minutes so that leaves the payload exposed in that interval. So just a strengthened strong-back without a change in launch procedure is sub-optimal.
Coating the outside of the LOX tanks *seems* good, however, whether they can find a coating whose thermal expansion/contraction equals that of the carbon fiber composite is the key question there.
As to the shooter problem, the best thing to do is not to armor the rocket, but to do two things:
First, “hang 40 meters on a side curtains” around the launch site that obscure the exact position of the rocket, on a movable launch pad (that has to be rebuilt anyway), and might be placed anywhere in that 40 meter stretch.
Second, put up as many audio and video monitors/direction finders as needed so that anyone simply trying to pump bullets into the rocket area will be found, seen, and dealt with by return fire before they have a significant chance of hitting the rocket in that time.
This will make *any* single shot very chancy for success, and multiple shots deadly for the shooter.
Titanium and carbon fiber composite have almost exactly the same coefficient of thermal expansion, which has been handy for top-end bicycle manufacturers who often use carbon fiber tubes mated to titanium joints.
And just add more microphones to the Cape. With more mics they could have triangulated on the source of that initial bang to establish whether there was a muzzle blast and supersonic projectile and if so where it came from.
ULA is obviously the top of my suspect list. Their plan is to infiltrate the first Mars flight and sabotage the mission, leaving the hundred passengers stranded on Mars where they’ll await a rescue from ULA’s far more experienced engineers with a vastly longer history in aerospace and far more proven vehicle designs. But I’ll bet their operative screws up and the first mission ends up wandering around the galaxy for years, hopping from planet to planet in a hopeless attempt to find their way back to Earth.
Is that Jupiter 2 or SS Minnow? 😉
I like this idea. Or instead of curtains, a walled enclosure with lightning arrestors atop and the movable launch platform inside. But in all likelihood the shooter / drone bomb problem is probably an operations issue and one so old that it has been scenario-ed to death already. I’m sure the USAF and SpaceX working together can address it.
I have been to a few shuttle launches. The closest I could get without actually being on the base is 7 miles from the pad.
So, if a shot was fired at the rocket, t would have had to have been fired from somewhere on the Cape itself. It would have been suicidal to do it with anything less than a sniper rifle (really, would you want to be within a quarter mile of a fully loaded rocket you *know* is going to blow up?).
Security is a routine launch operations issue, not an engineering or bleeding-edge tech issue. I keep coming back to Jim Bennett’s essay in The New Atlantis on the Space Guard:
http://www.thenewatlantis.com/publications/proposing-a-coast-guard-for-space
“Making sure nobody is shooting at the rocket” would definitely fall under the aegis of the Space Guard.
No, that is the responsibility of whoever controls the range (in this case the 45th Space Wing of the US Air Force). The Space Guard operates in space.
But like all organizations, it will suffer mission creep. Kind of like having the Air Force guarding bases with ground troops.
Eventually Space Guard would have an officer assigned to sea turtle hatchling population surveys at the Cape.
Under Bennett’s idea, the responsibility would be transferred from the Air Force to the Space Guard, along with all other Air Force operations that are space-related but not warfighting-related.
Range in general would be, not sure that would include base security.
My problem with SpaceX on this is the stupid article about them not getting immediate access to the ULA building’s roof. Besides the glint of light and the appearance of the grassy knoll; the only suggestion of sabotage via sniper is that SpaceX’s investigator wasn’t given immediate access to the ULA building. Then you read the story, and ULA insisted they get in touch with USAF investigators to gain access.
I’m my view, that is exactly what ULA should do. USAF has jurisdiction, and even if the building was built by ULA, it’s on a USAF base. It would be like some private eye showing up to your house, demanding access to your roof, and your demand that they go get a real magistrate with a search warrant being a suggestion of guilt. It’s a cheap mystery novel sidebar. It makes SpaceX, and now Musk himself, look like paranoid adolescents.
If I really wanted to blow up a rocket on the pad I expect I could do it from beyond 7 miles away. I hope you’ll understand my not saying how in a public forum.
What? Hasn’t Obama taught us that handing our plans to the enemy first is the patriotic thing to do?
Well, not with a sniper rifle then. Two miles is world record territory for that. In which case we’re not talking about a bullet, but rather malicious software or lasers or something like that.
It’s probably a bit over two miles if you use an anti-materiel rifle to hit a Falcon 9 size target (rather than a human sized target).
Ed, I wouldn’t be confident of that. How hard could it be to engineer a 50 cal rifle to fire tungsten/uranium rounds with increased powder charge, computer-aided targeting, and anchored to something massive (recoil being a big reason why large caliber sniper rifles are range restricted)? Maybe the barrel life would be a few shots, but you’re aiming at a huge, soft target. Seven miles would be difficult, but not out of reach of a variety of actors who might also have access to steerable bullets or railguns, to name a couple even more exotic technologies.
Alternately, install the rifle within 2 miles of the rocket and fire remotely.
I don’t think either happened, because either you end up with a rifle in the middle of a heavily searched zone or the sound of the gun on someone’s recording of the test firing (probably even from the ocean side). You don’t need a mat of sensors to determine a suspicious sound that arrives before the sound from the rocket could arrive.
Military grade Laser
Made by Boeing or Lockheed.
Google: “Boeing Laser” or “Lockheed Martin Laser”
Actually, it wouldn’t take much of a laser with liquid oxygen present. You just need to raise the auto ignition temperature of whatever material that oxygen is next to and…