SpaceX Investigating the SpaceX Rocket Explosion of September 1, 2016

[liometopum]

Here is my hypothesis.

What we (think) we know:
1. The explosion appears centered on the connection between the rocket and the pipe. The explosion was external to the rocket.
2,. They were fueling the F9 at the time of the explosion.
So... is that pipe the place where the fuel is was being added?

Also observable:
3. The wind is blowing from right to left, as shown by the clouds of gas emitted by the F9.
4. Most rocket explosions look different, as they start slower. This started in an instant, less than 4/100 of a second, without any prior visible indication.

5. The first explosion frame shows fireballs far down the rocket, beneath the point of explosion..
6. Those fireballs are on the left side, in the vapor clouds.

7. The ignited gas, at these lowest levels of initial flash burn, have a story to tell.
7a..Looking at the first few explosion frames shows that these fireballs are separated, somewhat.
7b. The vapor burn is within the pre-existing clouds of vapor.
7c. The vapor burn areas follow the contour of the vapor clouds, That is, where the clouds are farthest to the left, so is the flash fire.
7d. There is no flashing or burning outside the vapor clouds.
7e. That means the volatile gases were already there, in the vapor; they were not pushed there from the point of the initial explosion.
7f. Even the vapor cloud directly to the left of the point of explosion flashes.
7g. The lowest areas of vapor clouds do not ignite. That probably helps us time when the fuel began leaking.

Hypothesis:
There was a leak in the 'fuel line to rocket' connection, Fuel was escaping for some time, prior to the explosion. Being heavier than air, the fuel vapors sank, and were carried to the left by the wind. The wind speed was variable, causing a contour of vapor to the left, and below, the point of emission of the fuel vapor (which was mixed with the other emission gases). Somewhere down lower, below the point of leakage, static electricity, or rubbing between parts, caused a spark, and that ignited the volatile gas which flash-burned outwards in all the directions it could. The flash went out to the furthest point in the vapor clouds containing fuel, and burned up to the point of leakage. There, at the leak, it ignited the fuel where it was leaking directly from the 'fuel line to rocket' connection and, due to high density of fuel there, burned hotter and faster and produced the explosion.

 

[.Scott]

Does anyone have any information about the distance between the camera and the rocket and how much time did the sound take to arrive ?

Just over 4000 meters. An audio delay of 12.1 seconds.

 

[Jonathan Scott]

2,. They were fueling the F9 at the time of the explosion.
So... is that pipe the place where the fuel is was being added?

They had completed fueling (with RP-1) earlier. They were loading LOX at the time (70% full). The pipes for loading LOX and RP-1 connect much lower down than the point of the explosion. As far as I know, there is no connection between the tower and the rocket at the point of explosion.

The rate at which the initial flame propagated seems too fast to be for a mixture of fuel and air. It suggests that LOX was involved, although at the left hand side and falling below, the darker areas suggest fuel burning less rapidly, with air.

 

[.Scott]

Why is everyone assuming the cause was in the rocket?

First, let me tell you why "everyone" is assuming that the cause was with the ground support equipment. It's because the giant X drawn by the lens flare falls right at the right edge of the rocket - near the GSE. But this hardly rules out the rocket, as described below.

In the first frame showing flames (frame #1), flames are propagating at least 8.5 feet in no more than 40msec - and that doesn't include the velocity component towards the camera. To achieve this, an ignited fuel/oxygen mix needs to be ejected at hundreds of feet per second. There is no source for this mix at the tower - certainly not in the quantity seen.

The direction of this burst is towards the camera - but since the X is falling to the right of the rocket center, it must be ejecting somewhat to the right.

But we know something more about that mix. For the most part, it does not appear to be "chunks" of anything - rather it is a spray. The primary indication of this is in subsequent frames, frames #2 to #4. What we see in those frame is that the original burst is arrested - presumably by traveling through air, and there is not a sea of fragments emitted from the spray. Although, of course, we do see larger, non-burning fragments.

So that initial (frame #1 and #2) burning mixture is fluid - entirely or almost entirely.

 

[.Scott]

BTW: If we assume that the problem was in the 2nd stage of the rocket, which at this point seems very likely to me, then it introduces this troubling question: Were struts really to blame in last years RUD?

 

[nikkkom]

Were struts really to blame in last years RUD?

I have a feeling you guys are going to rediscover some of the theories already discussed and Nasaspaceflight forum.

Yes, they thought about "maybe it's the same as last years RUD" too. They thought about "maybe it's a COPV failure". They talked about telemetry data (or lack thereof). They did some analysis of the video and visible flying debris. Etc etc etc.

You can save yourself a lot of time by just reading the thread there.

 

[.Scott]

You can save yourself a lot of time by just reading the thread there.

Actually, I am reading the posts in that Forum. And as I said, they are good - well, some are good. I have already posted a graphic from that forum into this thread.
But I don't know about "saving yourself a lot of time". I'm up to post #1000 there - I have a couple thousand more to go. As I find things I think are interesting, I will post them here.

Here is a video from post #991, but bear in mind that everything in the video is well-known to the SpaceX engineers:

 

[mheslep]

...
F9 doesn't have any cryogenic or volatile fuels onboard. It uses RP-1 (a refined grade of kerosene). The visible white venting should be cold oxygen (likely boil-off from LOX tank).

Payload used hydrazine.

 

[mheslep]

... Neither oxidiser nor fuel on its own will release chemical energy.

Though unlike the fuel, as the video you posted indicates, O2 will burn on contact with all kinds of materials including some coatings and metals given an ignition source.

 

[mheslep]

What I want to know is why there aren't super high res cameras covering the launch pad? All I've seen is grainy photos and video.

At 500 or 1000 fps, too many frames to store? Probably reserved for the launch alone, not fueling and test fires. Edit: The videos available fron the Apollo launches taken by cameras on the launch platform are started 5 s before launch and use pricey quartz mirrors.

 

[Greg Bernhardt]

At 500 or 1000 fps, too many frames to store? Probably reserved for the launch alone, not fueling and test fires.

This is Elon Musk we're talking about. Storage is cheap.

 

[Jonathan Scott]

I think that the initial flame could not have been that fast and intense without significant amounts of both fuel and LOX pre-mixed by a powerful ejection event. I think that if the helium pressure vessel (COPV) failed abruptly (perhaps due to stresses caused by buoyancy forces or thermal contraction) it could well have burst the LOX tank and at the same time partially burst the fuel tank. In such a case, I'm sure there would be many possible sources of ignition, so no separate explanation is required for that and this seems to require only a single point of failure.

 

[mfb]

It's water vapor - condensed out of the air by cryogenic liquids.

And what is it's role

It does not have a role - it just occurs due to the rocket operation, when air passes very cold surfaces the water vapor condenses and forms this mist.

Payload used hydrazine.

The satellite fell down later and made its own explosion.

 

[mheslep]

This is Elon Musk we're talking about. Storage is cheap.

The Falcon 9 and it's launch operations are cheap. Reduced cost is the most notable aspect of the Falcon 9 relative to other launch vehicles.

http://www.abelcine.com/store/Phantom-Flex-High-Speed-Digital-Camera/#tabs

Phantom camera line from $50,000 to $150,000 depending on specific model and features.

 

[.Scott]

I think that the initial flame could not have been that fast and intense without significant amounts of both fuel and LOX pre-mixed by a powerful ejection event. I think that if the helium pressure vessel (COPV) failed abruptly (perhaps due to stresses caused by buoyancy forces or thermal contraction) it could well have burst the LOX tank and at the same time partially burst the fuel tank. In such a case, I'm sure there would be many possible sources of ignition, so no separate explanation is required for that and this seems to require only a single point of failure.

As I make my way through 9/3 on the nasaspaceflight.com forum, there seems to be more and more support for this. Also, based on that discussion, there are not 2 separate tanks - there is a "common bulkhead" separating the LOX and fuel. So a COPV failure only require the rupture of a single wall.

There is also discussion of:
* COPV issues with both the 1st and 2nd stages in previous flights.
* Whether there was a preplanned hold event. It's important in determining how far along SpaceX was in filling the LOX tank. It may have been full during the anomaly, or it may have been less than half full.

 

[nikkkom]

I think that the initial flame could not have been that fast and intense without significant amounts of both fuel and LOX pre-mixed by a powerful ejection event. I think that if the helium pressure vessel (COPV) failed abruptly (perhaps due to stresses caused by buoyancy forces or thermal contraction) it could well have burst the LOX tank and at the same time partially burst the fuel tank. In such a case, I'm sure there would be many possible sources of ignition, so no separate explanation is required for that and this seems to require only a single point of failure.

COPVs in S2 are designed to withstand buoyancy forces during 1st stage burnout, which are some 4..5 times larger than on a stationary rocket. Not likely to fail because of that.

COPV failure should be visible in tank pressure telemetry. However, it is not known whether SpaceX in fact has that telemetry. It's also possible SpaceX is unwilling to disclose that it's a COPV failure again (I think this has low probability).

 

[mfb]

It's also possible SpaceX is unwilling to disclose that it's a COPV failure again (I think this has low probability).

"We don't understand what happened" is the absolute worst case for SpaceX and trust of the customers. If they know what happened they will share it.

 

[RonL]

As I mentioned in my first post about the light rays showing an X pattern that looked like a good identifier of the initial explosion, is it possible that a hammer like affect took place through the liquid and knocked out the base seams, allowing that large flood of liquid coming out of the base of the fireball ? To me it looked like thousands of gallons falling. I think a broken pump line would have looked more like a flame thrower.

 

[liometopum]

The explosion is not symmetrical.

Using the second frame of the explosion, I have the "x-marks-the-spot" overlapped with a circle drawn to show how far, approximately, the blast had proceeded. But notice that flame is present far below the blast circle, and that flame is in the rocket's vapors, and only in the those vapors. We don't see flame anywhere else, outside the blast circle, except downward and to the left.

That suggests that the rocket vapors had fuel in it prior to the explosion.

 

[Jonathan Scott]

We don't see flame anywhere else, outside the blast circle, except downward and to the left.

I think it's already been mentioned (perhaps in Scott Manley's summary) that those areas are probably just reflected light from the main flash.

 

[jim hardy]

those areas are probably just reflected light from the main flash.

I second that. Water droplets reflect and diffuse light. I once saw(from about fifty miles) a night shuttle takeoff go through a cloud. The whole cloud illuminated like a lamp globe.

 

[liometopum]

Ok. That is definitely a point!
In argument, there are no vapor clouds visible in that lower area marked by the oval; you can easily see between the rocket and tower, for example. That is a lot of lamp globe effect on what must be a sparse cloud, but certainly possible given the intensity of the light. I am getting partially convinced, as the arrow points to a reflection. That is a lot of reflection though, without shadow. In any case, here are the images for review.

 

[jim hardy]

That is a lot of reflection though, without shadow. In any case, here are the images for review.

Thanks !
Here's yours at 3X

Is that part of a vapor cloud ? I tend to suffer excess of imagination so won't assert either way.
Peculiar, if it is vapor it's on upwind side suggesting a spill of something cold enough to bring air below dewpoint, probably 80 degreesF in Florida this time of year. Oxygen isn't much denser than air, water vapor is lighter, but cold air falls and so does fog ... so your picture is er, could be consistent with a LOX leak.old jim

 

[jim hardy]

speaking of reflections

one frame only, the second one , 71.77 at RowVid

If an electronic artifact it's one several pixels big? (i set RowVid to1080 )
Reflection off lens internal surface ? Vapor puff ? Fragment exiting to rear ? Other mystery objects are black, this one appeared to reflect toward camera...(there's some hysterical UFO claims on Youtube's lunatic fringe sites)

old jim

 

[liometopum]

Thanks Jim. I am sold on the reflection argument, so trash one hypothesis, and time to make a new one.

 

[liometopum]

LOX.
Here are some videos of liquid oxygen fires, and they are striking!

 

[mheslep]

Ok. That is definitely a point!
In argument, there are no vapor clouds visible in that lower area marked by the oval...

Which means no *water* vapor, or no vapor that is visible. Also means no supercooled vapor of any kind, which would immediately form a condensed water cloud.

 

[liometopum]

Which means no *water* vapor, or no vapor that is visible. Also means no supercooled vapor of any kind, which would immediately form a condensed water cloud.

The areas above and below, to the left, have vapor clouds. This area of the rocket appears free of vapor clouds. I am not sure what to make of that. Is the column of light just light, or is it flame?

 

[liometopum]

Here is another observation from the video, http://rowvid.com/?v=_BgJEXQkjNQ
Explosion frame 1 (EF1) produces an X, that we have shown earlier, and is shown here again, as the image on the left. Explosion frame 2 (EF2), shown on the right, has a wider and brighter X, and it appears slightly higher. I marked out a pair of lines on EF2 showing the wider area it covers.

Interestingly, the center of the area marked by the intersection of the four lines on EF2 is higher up, and seems to be following the vertical pipe highlighted by the white oval.

 

[Dotini]

For clarification, was there conceivably an ignition source such as atmospheric electricity discharge, or could mere contact between leaking LOX and fuel in the vicinity of the loading ports spontaneously ignite?