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

[woods170]

I wonder if B1050 is in the scrap heap.  Musk said he'd try to use it on an internal mission. This seemed like the best candidate use.  As any unexpected failure would only make the test more realistic.

Any unexpected failure would only make the test useless. It's timed to abort at a certain point and if it didn't, they wouldn't be proving what they need to prove.

A-003 disagrees.

https://en.wikipedia.org/wiki/A-003
https://twitter.com/jonostrower/status/1070739666444288000

[Comga]

I wonder if B1050 is in the scrap heap.  Musk said he'd try to use it on an internal mission. This seemed like the best candidate use.  As any unexpected failure would only make the test more realistic.

Any unexpected failure would only make the test useless. It's timed to abort at a certain point and if it didn't, they wouldn't be proving what they need to prove.

A-003 disagrees.

https://en.wikipedia.org/wiki/A-003
https://twitter.com/jonostrower/status/1070739666444288000

Isn't it clearly somewhere in the middle?
An abort triggered by an unplanned booster failure wouldn't be "useless".  It would be realistic and a demonstration of the automatic triggering, like on Little Joe.
An abort before Max-Q wouldn't test at that extreme of the envelope, so some extrapolation would still be necessary.

But it's all moot as Musk has said that first stage B1048.3+ will be used for the IFA, not the dried-out B1050, IIRC. 

[mme]

I wonder if B1050 is in the scrap heap.  Musk said he'd try to use it on an internal mission. This seemed like the best candidate use.  As any unexpected failure would only make the test more realistic.

Any unexpected failure would only make the test useless. It's timed to abort at a certain point and if it didn't, they wouldn't be proving what they need to prove.

A-003 disagrees.

...

I feel like people give A-003 too much of a pass. Apollo had several inflight abort tests and they got "lucky" this didn't happen on the Max-Q test or they would have had to re-attempt the test. Instead they could say this tested an abort during a spin and they don't really need to test an abort at 120,000 ft.

If this happened to the only F9 inflight abort test I doubt it would be hailed as a success. Maybe they'd spin it as a success, but it would not tested what the test is suppose to test.

[woods170]

I wonder if B1050 is in the scrap heap.  Musk said he'd try to use it on an internal mission. This seemed like the best candidate use.  As any unexpected failure would only make the test more realistic.

Any unexpected failure would only make the test useless. It's timed to abort at a certain point and if it didn't, they wouldn't be proving what they need to prove.

A-003 disagrees.

...

I feel like people give A-003 too much of a pass. Apollo had several inflight abort tests and they got "lucky" this didn't happen on the Max-Q test or they would have had to re-attempt the test. Instead they could say this tested an abort during a spin and they don't really need to test an abort at 120,000 ft.

If this happened to the only F9 inflight abort test I doubt it would be hailed as a success. Maybe they'd spin it as a success, but it would not tested what the test is suppose to test.

A-003 was far from useless. Prior to the LAS test flights the abort profile had been modeled extensively (for 1960's standards that is) by NASA and its contractors.
The unexpected, but successful, outcome of A-003 provided data-points that turned out to very nicely fit the abort models of the time. As such, A-003 helped in validating the abort models for Apollo, despite the unexpected outcome.

Flash forward to 2019: the abort profile for Crew Dragon has been extensively modeled by SpaceX and NASA.
An A-003 scenario on the inflight-abort test flight of Crew Dragon would miss the datapoints for max-Q, but would add data-points for whatever part of the profile the abort actually took place. So, even an A-003 scenario would still help to validate, or invalidate, the modeling of the Crew Dragon abort profiles. In both cases it would be anything but useless.

[Johnnyhinbos]

Considering that the in-flight abort is not a NASA requirement but instead is driven by SpaceX, any off parameter anomaly that triggers the abort system will provide valuable data that otherwise wouldn’t have been gathered. Granted not the targeted data, but data nonetheless.

For context - regardless of when / how the abort gets triggered, it will provide much more data / insights than Boeing’s in-flight abort, because Boeing isn’t doing one.

[Nomadd]

 I meant useless in that they'd have to do it again to get the results they needed. Maybe they'd get a pass if it was close enough.

[whitelancer64]

Apollo A-003 managed to achieve many of its objectives as an in-flight abort test in spite of the booster failure, not because of it. It was supposed to have been a high altitude abort test, and it didn't go high enough to hit all the criteria it was supposed to.

The subsequent test, A-004, was a successful high altitude abort test, and it picked up the remainder of the objectives of the previous flight.

If the Falcon 9 booster breaks up in flight, then it won't meet all of the desired test criteria (i.e. the capsule won't experience the dynamic pressure it was supposed to). It would mean, most likely, another in-flight abort test, at great cost and a big schedule delay for SpaceX.

[abaddon]

If the Falcon 9 booster breaks up in flight, then it won't meet all of the desired test criteria (i.e. the capsule won't experience the dynamic pressure it was supposed to). It would mean, most likely, another in-flight abort test, at great cost and a big schedule delay for SpaceX.

Disagree.  It would mean, more likely, that SpaceX would apply the findings to their simulation and just simulate the max-q abort, as Boeing is doing (without any flight data of the stack with the capsule at all).  Depending on how close it got to max-q, if it blows up early in the flight it would generate little valuable data in that regard.

Of course, it would also mean that SpaceX experienced an unexpected Falcon 9 failure that would need to be root-cause analyzed and set the progress of the program way back, frankly more than conducting another in-flight abort would likely result in on its own.  Little Joe failing on its own had no implication on the Apollo stack.

[whitelancer64]

If the Falcon 9 booster breaks up in flight, then it won't meet all of the desired test criteria (i.e. the capsule won't experience the dynamic pressure it was supposed to). It would mean, most likely, another in-flight abort test, at great cost and a big schedule delay for SpaceX.

Disagree.  It would mean, more likely, that SpaceX would apply the findings to their simulation and just simulate the max-q abort, as Boeing is doing (without any flight data of the stack with the capsule at all).  Depending on how close it got to max-q, if it blows up early in the flight it would generate little valuable data in that regard.

Of course, it would also mean that SpaceX experienced an unexpected Falcon 9 failure that would need to be root-cause analyzed and set the progress of the program way back, frankly more than conducting another in-flight abort would likely result in on its own.  Little Joe failing on its own had no implication on the Apollo stack.

Maybe, but as you say it would depend on how close to the desired test criteria the rocket was at failure.

As for Boeing, I will note that they will have their crew-less test flight data to include in their calculations. Presumably that will be part of their crew flight test and certification reviews.

[Norm38]

http://www.parabolicarc.com/2018/11/28/how-spacex-conduct-inflight-abort-test-crew-dragon/

I read through this article on the abort test, and I don't understand how this test is a worst case abort scenario.

The launch scenario where an abort is initiated during the ascent trajectory at the maximum dynamic pressure (known as max Q) is a design driver for the launch abort system. It dictates the highest thrust and minimum relative acceleration required between Falcon 9 and the aborting Dragon.

The Falcon 9 would be configured to shut down and terminate thrust, targeting the abort test shutdown condition (simulating a loss of thrust scenario).  Dragon would then autonomously detect and issue an abort command, which would initiate the nominal startup sequence of Dragon’s SuperDraco engine system. Concurrently, Falcon 9 would receive a command from Dragon to terminate thrust on the nine first stage Merlin 1D (M1D) engines.

If the S1 engines are shut down prior to abort, relative acceleration isn't minimum.  At MaxQ, S1 thrusting acceleration is not yet maximum.  And with MaxQ having maximum drag, the F9 will have maximum deceleration following separation.

IMHO, the worst case test is either early / near MaxQ in the flight where a guidance failure violently turns Dragon against the air stream before it can abort, OR near the end of S1 thrust where an abort is needed but S1 engines do not shut down, and Dragon has to outrun S1 at max acceleration and minimum drag.

Yes at MaxQ Dragon acceleration will have to fight against max drag, but with the factors I mentioned, why exactly is MaxQ the worst case scenario?

[abaddon]

http://www.parabolicarc.com/2018/11/28/how-spacex-conduct-inflight-abort-test-crew-dragon/

I read through this article on the abort test, and I don't understand how this test is a worst case abort scenario.

The launch scenario where an abort is initiated during the ascent trajectory at the maximum dynamic pressure (known as max Q) is a design driver for the launch abort system. It dictates the highest thrust and minimum relative acceleration required between Falcon 9 and the aborting Dragon.

The Falcon 9 would be configured to shut down and terminate thrust, targeting the abort test shutdown condition (simulating a loss of thrust scenario).  Dragon would then autonomously detect and issue an abort command, which would initiate the nominal startup sequence of Dragon’s SuperDraco engine system. Concurrently, Falcon 9 would receive a command from Dragon to terminate thrust on the nine first stage Merlin 1D (M1D) engines.

If the S1 engines are shut down prior to abort, relative acceleration isn't minimum.  At MaxQ, S1 thrusting acceleration is not yet maximum.  And with MaxQ having maximum drag, the F9 will have maximum deceleration following separation.

IMHO, the worst case test is either early / near MaxQ in the flight where a guidance failure violently turns Dragon against the air stream before it can abort, OR near the end of S1 thrust where an abort is needed but S1 engines do not shut down, and Dragon has to outrun S1 at max acceleration and minimum drag.

Yes at MaxQ Dragon acceleration will have to fight against max drag, but with the factors I mentioned, why exactly is MaxQ the worst case scenario?

Max drag is the most difficult in-flight regime in which to abort, your opinion not withstanding.  As far as shutting down the F9 engines or not, it doesn't really matter, because telemetry will tell them if it would be able to outpace a still thrusting stage.  As the throttle-down is conducted concurrent (not "prior to" as you wrote) with the abort, and throttle-down is a much slower process than the abort motors firing, the initial conditions of the abort will be essentially the same anyway.

[Norm38]

Relying on an interpretation of telemetry doesn't seem to satisfy the test.  They could extrapolate lots of things, including acceleration at MaxQ.  But they want real test data.
In the real world, engines may not shut down.  You're saying the shutdown is meaningless, so why would the engines be shut down before the Dragon is proven to be safely away?  The F9 is scrap the instant separation occurs.  It isn't for the sake of F9 to shut down the engines, so why do it until after say T+5 seconds?

[Norm38]

Max drag is the most difficult in-flight regime in which to abort, your opinion not withstanding.

Okay, but why exactly?  MaxQ is max drag.  After separation the streamlined Dragon will have less aero drag than the uncapped F9.  Wouldn't that increase the acceleration difference between the two?  And if any debris was thrown out by F9 disintegrating, that debris will also have max drag and be much less likely to catch up to Dragon.

I just want to understand better exactly why MaxQ is worst case when the forces seem to be working against it.

[mme]

http://www.parabolicarc.com/2018/11/28/how-spacex-conduct-inflight-abort-test-crew-dragon/

I read through this article on the abort test, and I don't understand how this test is a worst case abort scenario.

The launch scenario where an abort is initiated during the ascent trajectory at the maximum dynamic pressure (known as max Q) is a design driver for the launch abort system. It dictates the highest thrust and minimum relative acceleration required between Falcon 9 and the aborting Dragon.

The Falcon 9 would be configured to shut down and terminate thrust, targeting the abort test shutdown condition (simulating a loss of thrust scenario).  Dragon would then autonomously detect and issue an abort command, which would initiate the nominal startup sequence of Dragon’s SuperDraco engine system. Concurrently, Falcon 9 would receive a command from Dragon to terminate thrust on the nine first stage Merlin 1D (M1D) engines.

If the S1 engines are shut down prior to abort, relative acceleration isn't minimum.  At MaxQ, S1 thrusting acceleration is not yet maximum.  And with MaxQ having maximum drag, the F9 will have maximum deceleration following separation.

IMHO, the worst case test is either early / near MaxQ in the flight where a guidance failure violently turns Dragon against the air stream before it can abort, OR near the end of S1 thrust where an abort is needed but S1 engines do not shut down, and Dragon has to outrun S1 at max acceleration and minimum drag.

Yes at MaxQ Dragon acceleration will have to fight against max drag, but with the factors I mentioned, why exactly is MaxQ the worst case scenario?

The abort sequence shuts down the engines. If the engines keep running after the abort, the test found a serious bug/design flaw or you are using custom software and/or hardware. I think testing that the abort shuts down the engines is more important since they can tell via telemetry if it could have out run the booster. (I see you don't believe telemetry is good enough. How can it not be good enough? Of all the things to calculate that has to be one of the most trivial.)

Forcing the rocket into an attitude that aerodynamically destroys it does seem like a worst scenario as far as I can tell. I'm sure they debated it, I'd love to know their reasoning.

I am curious how they trigger the abort but will probably never know that either.

[abaddon]

Forcing the rocket into an attitude that aerodynamically destroys it does seem like a worst scenario as far as I can tell. I'm sure they debated it, I'd love to know their reasoning.

At some point, these imagined "worst case" scenarios get ridiculous.  Should the Proton failure where it tries to fly upside-down be considered the "worst case scenario?"  Aborting from a rocket pointing straight down at a few thousand feet would be worse than anything mentioned above, and yet suggesting a test be conducted this way is clearly ridiculous.

Of course, the clear and obvious answer is the abort system will detect a deviation from the planned trajectory and fire the abort sequence far before these abnormal aerodynamic forces will be relevant.

[matthewkantar]

Forcing the rocket into an attitude that aerodynamically destroys it does seem like a worst scenario as far as I can tell. I'm sure they debated it, I'd love to know their reasoning.

I am curious how they trigger the abort but will probably never know that either.

If the rocket get more than a degree or two off nominal and the abort sequence does not fire, that is a failure in itself. There should never be a reason to get away from a vehicle so far off course that it is structurally compromised.

[Norm38]

I didn't say that telemetry and extrapolation isn't good enough.  But I saw comments here that said that if an abort didn't occur at MaxQ, the test wouldn't be satisfied and they'd have to retest.  So they can extrapolate some things but not others.  Okay, I don't design rockets, I don't have to sweat the details.  Carry on.

[abaddon]

I didn't say that telemetry and extrapolation isn't good enough.  But I saw comments here that said that if an abort didn't occur at MaxQ, the test wouldn't be satisfied and they'd have to retest.  So they can extrapolate some things but not others.  Okay, I don't design rockets, I don't have to sweat the details.  Carry on.

I'm going to quote myself from up above, responding to that exactly point:

It would mean, more likely, that SpaceX would apply the findings to their simulation and just simulate the max-q abort, as Boeing is doing (without any flight data of the stack with the capsule at all).  Depending on how close it got to max-q, if it blows up early in the flight it would generate little valuable data in that regard.

Different people are expressing different opinions, YMMV.

[Alexphysics]

Max drag is the most difficult in-flight regime in which to abort, your opinion not withstanding.

Okay, but why exactly?  MaxQ is max drag.  After separation the streamlined Dragon will have less aero drag than the uncapped F9.  Wouldn't that increase the acceleration difference between the two?  And if any debris was thrown out by F9 disintegrating, that debris will also have max drag and be much less likely to catch up to Dragon.

I just want to understand better exactly why MaxQ is worst case when the forces seem to be working against it.

The drag goes against the move of the Dragon. Dragon imparts a force with its SuperDracos in one direction and the drag goes in the opposite one so the total force (which in turn determines the clearance between Falcon 9 and the Dragon) will be the force of the superdracos going in one direction minus the drag force that goes in the other direction over the mass of the Dragon. The thing is that from all possible phases of flight, the total acceleration is at its minimum when Dragon is at Max Q because the drag is the maximum one. And the thing about shutting down the engines is that in an in-flight abort, whenever they can if an abort is commanded the engines will shutdown. For example, imagine a CRS-7-like scenario, the first stage engines kept going even after the failiure. In a Crew Dragon flight the engines would shutdown and at basically the same time the SuperDracos would be activated to pull the capsule away from the rocket. This is what the in-flight abort test wants to emulate.

[mme]

Forcing the rocket into an attitude that aerodynamically destroys it does seem like a worst scenario as far as I can tell. I'm sure they debated it, I'd love to know their reasoning.

At some point, these imagined "worst case" scenarios get ridiculous.  Should the Proton failure where it tries to fly upside-down be considered the "worst case scenario?"  Aborting from a rocket pointing straight down at a few thousand feet would be worse than anything mentioned above, and yet suggesting a test be conducted this way is clearly ridiculous.

Of course, the clear and obvious answer is the abort system will detect a deviation from the planned trajectory and fire the abort sequence far before these abnormal aerodynamic forces will be relevant.

I agree, I'm sure SpaceX designed this test to provide them with the most relevant information. During the Mercury and Apollo programs they tried a bunch of different scenarios. SpaceX/NASA can't afford that kind of extravagance now and I am just curious what was debated and what the trade offs were.

I don't expect to ever find out since that will all be proprietary information.