SpaceX F9 : Crew Dragon In-Flight Abort Test : Jan. 19, 2020 : Discussion

[Alexphysics]

The plan was not to trigger AFTS unless the vehicle went on a bad trajectory.  The environmental assessment said this.  It was said in no uncertain terms at the pre-launch press conference.
...

Yes, but by definition AFTS is autonomous.  Question is: Was it armed?

AFTS is armed in all Falcon 9 launches. Different thing is asking if it was activated which was clearly not. The AFTS destroys both the first and the second stage and the latter stayed in one piece and even still attached to the interstage and a chunk of the top of the booster's LOX tank

[sanman]

Dragon commanding the engine shutdown doesn't mean the abort wasn't triggered by loss of thrust. Engine shutdown and abort thrusters firing are two different things.

Correct; the abort was *not* triggered by loss of F9 thrust, but by Dragon based on other criteria, as addressed during the presser Q&A (see numerous posts above).

Presumably, the parameters which triggered abort firing are merely correlated with loss of booster thrust, since in real life you can't rely on failure scenario where booster shuts down explicitly and conveniently.

[Rondaz]

The trunk is on board GO Navigator.

https://twitter.com/Kyle_M_Photo/status/1219062943125508097

[joek]

AFTS is armed in all Falcon 9 launches. Different thing is asking if it was activated which was clearly not. The AFTS destroys both the first and the second stage and the latter stayed in one piece and even still attached to the interstage and a chunk of the top of the booster's LOX tank

Not sure what distinction you are making.  If it was armed then it is by definition "active".  If it did not activate--as in trigger the explosive-zip FTS on S1 and-or S2, as opposed to thrust termination--then something inhibited it.  If it was inhibited, then by definition it was not "active".  In any case, there are two distinct AFTS systems, one for S1 and another for S2.  If S1 AFTS system activates, then presumable S1 is still coupled to S2 (although not necessarily a give).  The S2 AFTS system must be able to operate independently of S1 as there are still conditions under which the S2 AFTS must operate independent of S1; obviously post S1-S2 separation is one of those conditions.

[joek]

Presumably, the parameters which triggered abort firing are merely correlated with loss of booster thrust, since in real life you can't rely on failure scenario where booster shuts down explicitly and conveniently.

Not necessarily.  For example, one of the engine's TVC's goes bonkers; or an engine has issues but it does not affect overall S1 performance.  Plenty of off-nominal conditions might trigger an abort while "booster thrust" remains nominal.  And plenty of off-nominal conditions that might be of concern, but do not warrant an abort.

This focus on "booster thrust" as the be-all end-all abort criteria is misplaced.

[racevedo88]

So, no, the other tests are not just single-chute unit tests.

As stated "single-chute unit tests or similar".  What is the difference between those and "system" tests?  No idea.  However, during presser Q&A the answer was quite clear: only two "system" level tests have been performed.  So obviously those multi-chute tests were not considered "system" level tests.

I don't think single chute and multi-chute tests are similar, in fact Elon Musk specifically made a clarification  to distinguish them:

https://twitter.com/elonmusk/status/1191475073758064640

You say "only" two system level tests have been performed, but we don't know what is the necessary number of system level tests either. Boeing only did 2 development tests, 5 qualification tests and 3 reliability tests in total. So 2 is not necessarily a small number in this context.

As a former tester for airborne delivered equipment for the army, this might clarify what it all means, SPACE x as per their twee in dec has competed 10 multichute test of the the mk3 parachute.  of which 2 of these are system test. What it means it that they have conducted 8 test of the mk3 system with 4 parachutes using what is called as a mass simulator, and two using a the system such as a boilerplate plate of full dragon capsule that replicates all the conditions of deployments.

[mme]

AFTS is armed in all Falcon 9 launches. Different thing is asking if it was activated which was clearly not. The AFTS destroys both the first and the second stage and the latter stayed in one piece and even still attached to the interstage and a chunk of the top of the booster's LOX tank

Not sure what distinction you are making.  If it was armed then it is by definition "active".  If it did not activate--as in trigger the explosive-zip FTS on S1 and-or S2, as opposed to thrust termination--then something inhibited it.  If it was inhibited, then by definition it was not "active".  In any case, there are two distinct AFTS systems, one for S1 and another for S2.  If S1 AFTS system activates, then presumable S1 is still coupled to S2 (although not necessarily a give).  The S2 AFTS system must be able to operate independently of S1 as there are still conditions under which the S2 AFTS must operate independent of S1; obviously post S1-S2 separation is one of those conditions.

I think the ATFS is all about the instantaneous impact point.  No need to trigger if it's well inside the hazard zone.

edit: activate -> trigger as per Alexphysics use of terminology because that's what I meant but may not be what I said.

[groknull]

But remember, NASA said that the analysis of the Dragon 2 parachute failure showed that the accepted modeling was inadequate despite decades of use and acceptance. Orion and Starliner used those models to do verification of their parachutes. Yet both were allowed to continue without new drop tests.
Did either of those programs have to update their models and recertifications?

How much do we know about this "Mark III" parachute design?  Very little.[1]  How much do we know about the differences between this design and previous, potentially more conventional and well-modeled systems?  Very little.  How much do we know whether the models of this new design accurately reflect what has been observed in testing?  Very little.

Yeah, there may be an overabundance of caution.  But would not be surprised if this new-improved Mark III system held some surprises which caused pause in some quarters.  It's a set of parachutes, right?  We've done this for years, right?  What possible new-and-potentially-radical improvements could SpaceX's Mark II have which might upset the cart and make our old models questionable?

I have no idea what's really in play.  But I think it's safe to say that SpaceX might not have been satisfied with the humdrum well-proven designs (and models) and as typical, gone their own route...and thus the need-desire on NASA's part to perform additional validation.


[1] Speaking for myself.  Have seen little-to-no details.  Anyone?

Publicly released info on the differences between Mark 3 and earlier designs is sparse.  From pictures I have seen, the individual parachutes are of fairly conventional ring-sail design.  The issues appear to have been related to interactions in a 4 parachute system.

Some Mark 3 details in the October 10, 2019 briefing.  Elon at 25:30 in, plus Jim Bridenstine's comments immediately after.

Elon: Materials, stitching, strength, managing stress concentration.
Jim: Learned a lot, especially asymmetry that can be shared to improve safety throughout the industry.
(Thanks to whitelancer64 for the link plus a thread for parachute info: Parachute Model used by NASA )

NASA has provided existing knowledge and models to the commercial providers to help them with their design.  It appears that the "we've done this for years" parachute knowledge was a little incomplete, especially in the area of multi-chute interaction, asymmetric deployment, and asymmetric component loads.

A post by PM3 contains a NASA blog entry related to the SpaceX Dragon 2 parachutes and parachute tests:
https://forum.nasaspaceflight.com/index.php?topic=46136.msg1993322#msg1993322
Notably: "...caused a closer examination of the current industry standard used to calculate the asymmetry factor".

A post by thirtyone (also linked by whitelancer64 in the Parachute Model used by NASA  thread) has some good observations and links:
https://forum.nasaspaceflight.com/index.php?topic=46136.msg1993863#msg1993863
That post, and the referenced NASA PDF are definitely worth a read.

In the forum threads there are additional details and history relating to the Dragon 2 parachutes, but it is spread out and a bit difficult to find.  The NASA request to change from three to four parachutes is there somewhere too.

So not a radical design based on some SpaceX whim, but rather real world nonlinear dynamics and chaotic fluid flow throwing some surprises for 4 parachute systems.

[TheRadicalModerate]

I don't understand how you could get RP-1 and LOX to mix efficiently enough to generate that kind of a deflagration.  Seems like you'd need both tanks to rupture at almost the same time and vent enough aerosol into an interior space to get that kind of energy.

In that context it has been suggested that - considering a shared bulkhead between RP1 and LOX tanks - rupture of one tank might cause the bulkhead to invert/rupture, with the effect that one tank losing pressure will automatically fail the other.
This will also create a vent, exactly where the tanks join, and cause the still-at-full-pressure tank to vent into the lower pressure venting tank. So the only way for the Lox to get out - at this point - is through the RP1

If you look at the 4k boom-footage frame by frame, you can see exactly that. a first, yellow flame, creating a fireball, that is then followed by a much more violent and larger fireball shortly after.

I can't tell which tank burst first, Lox or Rp1, but it definitely made a nice big BOOM!

Sounds plausible.

One of the things I noticed from the close-up of the D2 pulling away is that the S2 starts venting something as soon as the angle of attack gets big.   From other footage, it looks like the core is also venting before the explosion.

I agree with your analysis that there are two different ignition events.  Looks to me that the first one (the yellow one) is in the slipstream, likely from RP-1 and LOX mixing.  That might easily provide enough overpressure to drive LOX through the RP-1 bulkhead, which would likely provide a very nice fuel-air explosive as both tanks ruptured.

[TheRadicalModerate]

It does not make much sense to abort at less than maximum thrust since even the maximum acceleration is fairly benign (compared to other aborts). Remember that the 3.5 Gs is what is left after overcoming air resistance at max Q.

Assuming that "It does not make much sense to abort at less than maximum thrust..." refers to purposes of testing, and specifically this test... the SpaceX test was at "max drag" (their words)--a bit later than max Q as the noted during the test call-outs.

So did it abort at max F9 thrust?  Maybe for a short period, until the F9 engine shutdown took effect.  But separation between Dragon-F9 is more than what happens in that instant.  Suppose that the F9 engine shutdown failed and that F9 kept boosting?

To achieve sufficient Dragon-F9 distance in a timely manner means more than simply overcoming residual aerodynamic resistance.  It also means (in a worst case) that Dragon can sufficiently outrun F9 should F9 keep boosting.

We did not see that situation today, although it was touched upon during the presser--and the answer was that Dragon could still outrun F9 in that situation to achieve sufficient separation.  That would likely require more than 3.5G boost by Dragon, as it would also require overcoming F9's continuing boost (again, worst case).

The details as to how Dragon might make the decision to increase boost in that situation was not discussed during the presser.

I think the D2 is going pretty close to flat-out at abort.  8 SuperDracos is 8 * 73 kN * cos(10º) = 575 kN.  If the D2 weighs 12 t, then the no-drag acceleration could be as high as 4.9 g.  But you can probably throw in at least 100 kN of drag, with maybe another 50 kN of residual post-shutdown thrust, which would make it pretty close to 3.5 g.

Remember that there's residual Merlin thrust after they've been shut down.  I count 6 1080p frames (~180 ms) from when the exhaust plume starts to shorten (presumably because of the initiation of shutdown) and first motion of the D2 away from the S2.

Assuming that it takes 100 ms to pressurize the SDs, open the valves, and release the S2 latches, that sounds like abort is initiated within 80 ms of the shutdown indication.

That's a perfectly reasonable real-world scenario, because the first thing you'd want to do in any abort would be to shut down the engines if you could.  And if you can't shut them down, it's probably because there's an explosion between you and them, and the nature of F9 "thrust" is getting a bit murky:  Thrust from the engines?  From half the rocket being explosively propelled forward by the other half?

If that explosion had occurred with the D2 attached, I'm guessing that the shockwave would likely have crushed the D2 no matter what acceleration it had.  Some events aren't survivable.  But most events start with an indication that's at least a couple of seconds before things go completely pear-shaped.  So abort at shutdown, with some residual thrust, is probably pretty close to the worst survivable case.

[penguin44]

I do recall the commentary saying once it went boom, as a reminder our flight termination system is not active for this flight.

[FlattestEarth]

The trunk is on board GO Navigator.

That is pretty impressive even if it is not fully intact.  At what altitude was it jettisoned?

[Rocket Science]

So this is unrelated to the current discussion happening in chat but will recovery vehicles like GO Searcher still be deployed during every crew dragon launch in case of an abort or will they just be on standby in port

Welcome to the forum!

[Rocket Science]

I really appreciate and enjoy the NSF live feeds. Great work by Chris G., Das and Julia at the port welcoming home Dragon.

[Alexphysics]

AFTS is armed in all Falcon 9 launches. Different thing is asking if it was activated which was clearly not. The AFTS destroys both the first and the second stage and the latter stayed in one piece and even still attached to the interstage and a chunk of the top of the booster's LOX tank

Not sure what distinction you are making.  If it was armed then it is by definition "active".  If it did not activate--as in trigger the explosive-zip FTS on S1 and-or S2, as opposed to thrust termination--then something inhibited it.  If it was inhibited, then by definition it was not "active".  In any case, there are two distinct AFTS systems, one for S1 and another for S2.  If S1 AFTS system activates, then presumable S1 is still coupled to S2 (although not necessarily a give).  The S2 AFTS system must be able to operate independently of S1 as there are still conditions under which the S2 AFTS must operate independent of S1; obviously post S1-S2 separation is one of those conditions.

You're making a mountain from a grain of sand and confusing yourself over something rrrrrreally simple. AFTS was armed which means it could have been triggered at any point in flight, this was said by Benji Reed on the pre launch news conference (let's also watch these before commenting please, it helps on the discussions). The AFTS when triggered is triggered for the whole rocket, it is triggered separately only when the stages actually separate like when the first stage is landing and they say "AFTS has been safed" which means the AFTS is not active anymore and won't be able to be triggered by the booster anymore because it doesn't pose any danger to the people on the ground. If you understand the difference between "active AFTS" and "triggered AFTS" then this whoooole pages long discussion is as simple as "it was active, but it wasn't triggered". That's it.

[tyrred]

It does not make much sense to abort at less than maximum thrust since even the maximum acceleration is fairly benign (compared to other aborts). Remember that the 3.5 Gs is what is left after overcoming air resistance at max Q.

Assuming that "It does not make much sense to abort at less than maximum thrust..." refers to purposes of testing, and specifically this test... the SpaceX test was at "max drag" (their words)--a bit later than max Q as the noted during the test call-outs.

So did it abort at max F9 thrust?  Maybe for a short period, until the F9 engine shutdown took effect.  But separation between Dragon-F9 is more than what happens in that instant.  Suppose that the F9 engine shutdown failed and that F9 kept boosting?

To achieve sufficient Dragon-F9 distance in a timely manner means more than simply overcoming residual aerodynamic resistance.  It also means (in a worst case) that Dragon can sufficiently outrun F9 should F9 keep boosting.

We did not see that situation today, although it was touched upon during the presser--and the answer was that Dragon could still outrun F9 in that situation to achieve sufficient separation.  That would likely require more than 3.5G boost by Dragon, as it would also require overcoming F9's continuing boost (again, worst case).

The details as to how Dragon might make the decision to increase boost in that situation was not discussed during the presser.

I think the D2 is going pretty close to flat-out at abort.  8 SuperDracos is 8 * 73 kN * cos(10º) = 575 kN.  If the D2 weighs 12 t, then the no-drag acceleration could be as high as 4.9 g.  But you can probably throw in at least 100 kN of drag, with maybe another 50 kN of residual post-shutdown thrust, which would make it pretty close to 3.5 g.

Remember that there's residual Merlin thrust after they've been shut down.  I count 6 1080p frames (~180 ms) from when the exhaust plume starts to shorten (presumably because of the initiation of shutdown) and first motion of the D2 away from the S2.

Assuming that it takes 100 ms to pressurize the SDs, open the valves, and release the S2 latches, that sounds like abort is initiated within 80 ms of the shutdown indication.

That's a perfectly reasonable real-world scenario, because the first thing you'd want to do in any abort would be to shut down the engines if you could.  And if you can't shut them down, it's probably because there's an explosion between you and them, and the nature of F9 "thrust" is getting a bit murky:  Thrust from the engines?  From half the rocket being explosively propelled forward by the other half?

If that explosion had occurred with the D2 attached, I'm guessing that the shockwave would likely have crushed the D2 no matter what acceleration it had.  Some events aren't survivable.  But most events start with an indication that's at least a couple of seconds before things go completely pear-shaped.  So abort at shutdown, with some residual thrust, is probably pretty close to the worst survivable case.

As Elon said in the post-presser, it's more of a fireball than an explosion. Highly unlikely there would be a shockwave that could crush the CrewDragon.  Plus the base heat shield, which is designed to handle heat, as also stated by Elon in the presser.

And the reaction times to these types of abort scenario events are usually in the range of milliseconds, not a couple of seconds.

[codav]

As Elon said in the post-presser, it's more of a fireball than an explosion. Highly unlikely there would be a shockwave that could crush the CrewDragon.  Plus the base heat shield, which is designed to handle heat, as also stated by Elon in the presser.

And the reaction times to these types of abort scenario events are usually in the range of milliseconds, not a couple of seconds.

In addition to that, Elon stated on Twitter that Dragon even would've had escaped successfully with an AMOS-6 failure mode, where the second stage oxygen tank suddenly went up in flames, which is the topmost part of the rocket just below dragon:

https://twitter.com/elonmusk/status/771479910778966016

[clongton]

If that explosion had occurred with the D2 attached, I'm guessing that the shockwave would likely have crushed the D2 no matter what acceleration it had.  Some events aren't survivable.  But most events start with an indication that's at least a couple of seconds before things go completely pear-shaped.  So abort at shutdown, with some residual thrust, is probably pretty close to the worst survivable case.

This was specifically addressed at the news conference. Elon said that Dragon would easily have survived the explosion and brought any crew onboard safely home. Jim then injected that Dragon is designed to withstand an exploding rocket. One of them said Dragon is designed to survive and safely "fly out of the fireball", although they hope they never have to test that.

[Alexphysics]

Considering how intact the trunk is after falling from 40km high I have no doubts an explosion wouldn't pose an inmediate danger to the crew considering they have the trunk and the heatshield in between everything to shield them while at the same time it would be pulling away quickly with the help of the SuperDracos. Now if these were to fail then yeah, the crew is totally dead.

[jpo234]

Considering how intact the trunk is after falling from 40km high I have no doubts an explosion wouldn't pose an inmediate danger to the crew considering they have the trunk and the heatshield in between everything to shield them while at the same time it would be pulling away quickly with the help of the SuperDracos. Now if these were to fail then yeah, the crew is totally dead.

After CRS-7 Musk said that the Dragon would have survived, if it had deployed its chutes. That would imply that the SDs are not required in all abort scenarios.