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

[gongora]

Does Boeing have to perform an in-flight abort before flying a manned mission?  If not (I haven't heard they have to) why not? 

Boeing is not doing an in-flight abort test.  They are doing computer simulations instead.  That is the certification path they and NASA agreed on.  SpaceX chose a different certification path and agreed to that with NASA.  Both certification routes are acceptable.  This has been discussed many times on the forum.

[Athelstane]

Does Boeing have to perform an in-flight abort before flying a manned mission?  If not (I haven't heard they have to) why not? 

Boeing is not doing an in-flight abort test.  They are doing computer simulations instead.  That is the certification path they and NASA agreed on.  SpaceX chose a different certification path and agreed to that with NASA.  Both certification routes are acceptable.  This has been discussed many times on the forum.

Yes, you can question whether NASA made a prudent decision on structuring the contracts such that an IFA was only one way of meeting certification requirements, but we can hardly blame Boeing for choosing an option (simulations) allowed by the contract.

Obviously, this remains confusing to a lot of people, even now.

If the parachutes really are a problem, that's what drop tests are for.

[sferrin]

Does Boeing have to perform an in-flight abort before flying a manned mission?  If not (I haven't heard they have to) why not? 

Boeing is not doing an in-flight abort test.  They are doing computer simulations instead.  That is the certification path they and NASA agreed on.  SpaceX chose a different certification path and agreed to that with NASA.  Both certification routes are acceptable.  This has been discussed many times on the forum.

Thanks.  I did not see previous discussions.  I mostly just lurk here.

[whitelancer64]

Does Boeing have to perform an in-flight abort before flying a manned mission?  If not (I haven't heard they have to) why not? 

Boeing is not doing an in-flight abort test.  They are doing computer simulations instead.  That is the certification path they and NASA agreed on.  SpaceX chose a different certification path and agreed to that with NASA.  Both certification routes are acceptable.  This has been discussed many times on the forum.

Thanks.  I did not see previous discussions.  I mostly just lurk here.

Specifically, NASA's requirements for the commercial crew program said that abort system testing must be done, along with modelling (by both NASA and the provider) to show that the testing demonstrates that the capsule can safely abort at any time in the launch sequence. The type of testing was not specified, that was left up to the provider.

When Boeing put in a bid for the commercial crew program, it included a pad abort test. SpaceX's bid had a pad abort test and an in-flight abort test. NASA accepted both bids.

[Athelstane]

Does Boeing have to perform an in-flight abort before flying a manned mission?  If not (I haven't heard they have to) why not? 

Boeing is not doing an in-flight abort test.  They are doing computer simulations instead.  That is the certification path they and NASA agreed on.  SpaceX chose a different certification path and agreed to that with NASA.  Both certification routes are acceptable.  This has been discussed many times on the forum.

Thanks.  I did not see previous discussions.  I mostly just lurk here.

Specifically, NASA's requirements for the commercial crew program said that abort system testing must be done, along with modelling (by both NASA and the provider) to show that the testing demonstrates that the capsule can safely abort at any time in the launch sequence. The type of testing was not specified, that was left up to the provider.

When Boeing put in a bid for the commercial crew program, it included a pad abort test. SpaceX's bid had a pad abort test and an in-flight abort test. NASA accepted both bids.

Exactly.

I tend to agree with the argument that NASA ought to have required an in-flight abort test, because there's nothing like actual testing to evaluate a system. But this CCiCap was trying to build in flexibility for vendors, and they knew *these* vendors had different strengths and weaknesses. Flying an in-flight abort was cheap for SpaceX, because it knew it would have a stack of recovered Block 5 Falcon 9's, meaning their costs would be just fuel, nominal refurbishment, and any incremental ground staff labour costs. Whereas Boeing had more experience in modeling vehicle performance.

Not surprising that each vendor chose the offered option that best suited their strengths.

[ncb1397]

Does Boeing have to perform an in-flight abort before flying a manned mission?  If not (I haven't heard they have to) why not? 

Boeing is not doing an in-flight abort test.  They are doing computer simulations instead.  That is the certification path they and NASA agreed on.  SpaceX chose a different certification path and agreed to that with NASA.  Both certification routes are acceptable.  This has been discussed many times on the forum.

Thanks.  I did not see previous discussions.  I mostly just lurk here.

Specifically, NASA's requirements for the commercial crew program said that abort system testing must be done, along with modelling (by both NASA and the provider) to show that the testing demonstrates that the capsule can safely abort at any time in the launch sequence. The type of testing was not specified, that was left up to the provider.

When Boeing put in a bid for the commercial crew program, it included a pad abort test. SpaceX's bid had a pad abort test and an in-flight abort test. NASA accepted both bids.

Exactly.

I tend to agree with the argument that NASA ought to have required an in-flight abort test, because there's nothing like actual testing to evaluate a system. But this CCiCap was trying to build in flexibility for vendors, and they knew *these* vendors had different strengths and weaknesses. Flying an in-flight abort was cheap for SpaceX, because it knew it would have a stack of recovered Block 5 Falcon 9's, meaning their costs would be just fuel, nominal refurbishment, and any incremental ground staff labour costs. Whereas Boeing had more experience in modeling vehicle performance.

Not surprising that each vendor chose the offered option that best suited their strengths.

The launch vehicle wouldn't have made much of a difference, it would have been the difference between 4.2 billion and 4.3 billion or 2%. Anyways, the Atlas V has a very lengthy record with zero problems that would require an escape either on the pad or in flight. When these awards were given out, Falcon 9 was pretty new (it flew a grand total of 12 times compared to 48 times for Atlas V) and went on later to have issues both in flight and on the pad. To compensate for that question mark on Falcon 9 reliability, SpaceX proposed more abort testing as mitigation. Mitigation that was probably needed.

[Athelstane]

My only point on cost was that it's clearly much cheaper for SpaceX to do an IFA than it is for Boeing. Which is not to say that cost was the *only* consideration for each company in framing their respective bids.

[tyrred]

Does Boeing have to perform an in-flight abort before flying a manned mission?  If not (I haven't heard they have to) why not? 

Boeing is not doing an in-flight abort test.  They are doing computer simulations instead.  That is the certification path they and NASA agreed on.  SpaceX chose a different certification path and agreed to that with NASA.  Both certification routes are acceptable.  This has been discussed many times on the forum.

Thanks.  I did not see previous discussions.  I mostly just lurk here.

Specifically, NASA's requirements for the commercial crew program said that abort system testing must be done, along with modelling (by both NASA and the provider) to show that the testing demonstrates that the capsule can safely abort at any time in the launch sequence. The type of testing was not specified, that was left up to the provider.

When Boeing put in a bid for the commercial crew program, it included a pad abort test. SpaceX's bid had a pad abort test and an in-flight abort test. NASA accepted both bids.

Exactly.

I tend to agree with the argument that NASA ought to have required an in-flight abort test, because there's nothing like actual testing to evaluate a system. But this CCiCap was trying to build in flexibility for vendors, and they knew *these* vendors had different strengths and weaknesses. Flying an in-flight abort was cheap for SpaceX, because it knew it would have a stack of recovered Block 5 Falcon 9's, meaning their costs would be just fuel, nominal refurbishment, and any incremental ground staff labour costs. Whereas Boeing had more experience in modeling vehicle performance.

Not surprising that each vendor chose the offered option that best suited their strengths.

The launch vehicle wouldn't have made much of a difference, it would have been the difference between 4.2 billion and 4.3 billion or 2%. Anyways, the Atlas V has a very lengthy record with zero problems that would require an escape either on the pad or in flight. When these awards were given out, Falcon 9 was pretty new (it flew a grand total of 12 times compared to 48 times for Atlas V) and went on later to have issues both in flight and on the pad. To compensate for that question mark on Falcon 9 reliability, SpaceX proposed more abort testing as mitigation. Mitigation that was probably needed.

Please provide a citation for your numbers, difference of launch vehicle 4.2 billion vs 4.3 billion. How are you measuring the difference of Atlas V vs Falcon 9?

[rpapo]

Please provide a citation for your numbers, difference of launch vehicle 4.2 billion vs 4.3 billion. How are you measuring the difference of Atlas V vs Falcon 9?

I think the difference in cost is a little greater, though not by that much.  Consider: An Atlas V is between $100M and $150M for the base version, depending on whether you want to look at it as cost versus price.  Then there is the cost of setting up and carrying out the mission.  I could imagine that as anywhere from $50M to $100M.  Finally, there is the time factor.  Atlas Vs are expendible, and each is made for one mission.  You don't just happen to have an inventory of them ready to use whenever you want to.

On the other side of the equation is the cost of the simulations.  They aren't cheap, but I would find it hard to believe if they ran more than $20-30M dollars for the whole program.  Especially if the base simulation software was already available, as I believe it would be for the Atlas V.

My two cents worth.

[Alexphysics]

They don't need an Atlas V to perform an in flight abort test and this is getting off topic for this thread which is about SpaceX's in flight abort test. There are other commercial crew threads in other parts of the forum.

[envy887]

Does Boeing have to perform an in-flight abort before flying a manned mission?  If not (I haven't heard they have to) why not? 

Boeing is not doing an in-flight abort test.  They are doing computer simulations instead.  That is the certification path they and NASA agreed on.  SpaceX chose a different certification path and agreed to that with NASA.  Both certification routes are acceptable.  This has been discussed many times on the forum.

Thanks.  I did not see previous discussions.  I mostly just lurk here.

Specifically, NASA's requirements for the commercial crew program said that abort system testing must be done, along with modelling (by both NASA and the provider) to show that the testing demonstrates that the capsule can safely abort at any time in the launch sequence. The type of testing was not specified, that was left up to the provider.

When Boeing put in a bid for the commercial crew program, it included a pad abort test. SpaceX's bid had a pad abort test and an in-flight abort test. NASA accepted both bids.

Exactly.

I tend to agree with the argument that NASA ought to have required an in-flight abort test, because there's nothing like actual testing to evaluate a system. But this CCiCap was trying to build in flexibility for vendors, and they knew *these* vendors had different strengths and weaknesses. Flying an in-flight abort was cheap for SpaceX, because it knew it would have a stack of recovered Block 5 Falcon 9's, meaning their costs would be just fuel, nominal refurbishment, and any incremental ground staff labour costs. Whereas Boeing had more experience in modeling vehicle performance.

Not surprising that each vendor chose the offered option that best suited their strengths.

The launch vehicle wouldn't have made much of a difference, it would have been the difference between 4.2 billion and 4.3 billion or 2%. Anyways, the Atlas V has a very lengthy record with zero problems that would require an escape either on the pad or in flight. When these awards were given out, Falcon 9 was pretty new (it flew a grand total of 12 times compared to 48 times for Atlas V) and went on later to have issues both in flight and on the pad. To compensate for that question mark on Falcon 9 reliability, SpaceX proposed more abort testing as mitigation. Mitigation that was probably needed.

That may have been true at the time, but Falcon 9 now has 46 consecutive full successes, plus 3 Falcon Heavy full successes with significant F9 heritage. There is no longer any significant difference in demonstrated or expected reliability between the CC rides.

I'd still like to see both providers do in-flight tests.

[theonlyspace]

Any update on a firm date for Spacea x In Flight Abort Test?

[edzieba]

Just to cap things off on the contracting: NASA did not 'offer options' on abort system testing plans. Bidders proposed test plans and NASA accepted them.

[marsbase]

I'd still like to see both providers do in-flight tests.

The Crew Capsule Draco and Super Draco systems have been tested hundreds (if not thousands) of times and worked reliably until they didn't.  If SpaceX had relied on simulations instead of going for an in-flight abort and the pre-testing leading up to that, the titanium combustion issue would not have been simulated because it was not suspected.  At some point in the future, that would have led to the death of crew members.  It's fine to use real-world tests to verify simulations.  It is not ok to use only simulations to replace actual tests because you can't simulate what you don't understand.

[woods170]

When these awards were given out, Falcon 9 was pretty new (it flew a grand total of 12 times compared to 48 times for Atlas V) and went on later to have issues both in flight and on the pad. To compensate for that question mark on Falcon 9 reliability, SpaceX proposed more abort testing as mitigation. Mitigation that was probably needed.

You've got your history screwed up so I'll enlighten you:

A: NASA never required any of the CCP contractors to perform flight tests to "prove" their abort systems. Flight testing was completely optional and could be added to the CCiCAP and CCtCAP contracts as VOLUNTARY milestones.

B. Flight history (or lack thereof) of the launcher had ZERO influence on the level of required abort flight testing exactly because no actual abort flight testing was required.

C. NASA only required the CCP contractors to "prove" their abort systems theoretically. Extensive computer modeling, coupled with extensive ground-based testing was deemed sufficient by NASA to certify the abort systems.

D. The number of flights by the time the contracts are awarded has ZERO influence on the contract award. Even the number of flights at the time of the proposals being due is of little-to-no value. It is all about HOW the contractor proposes to convince NASA that it will be safe to fly on the contractors vehicle(s).

[Senex]

When these awards were given out, Falcon 9 was pretty new (it flew a grand total of 12 times compared to 48 times for Atlas V) and went on later to have issues both in flight and on the pad. To compensate for that question mark on Falcon 9 reliability, SpaceX proposed more abort testing as mitigation. Mitigation that was probably needed.

You've got your history screwed up so I'll enlighten you:

A: NASA never required any of the CCP contractors to perform flight tests to "prove" their abort systems. Flight testing was completely optional and could be added to the CCiCAP and CCtCAP contracts as VOLUNTARY milestones.

B. Flight history (or lack thereof) of the launcher had ZERO influence on the level of required abort flight testing exactly because no actual abort flight testing was required.

C. NASA only required the CCP contractors to "prove" their abort systems theoretically. Extensive computer modeling, coupled with extensive ground-based testing was deemed sufficient by NASA to certify the abort systems.

D. The number of flights by the time the contracts are awarded has ZERO influence on the contract award. Even the number of flights at the time of the proposals being due is of little-to-no value. It is all about HOW the contractor proposes to convince NASA that it will be safe to fly on the contractors vehicle(s).

Thank you, once again, for providing facts to get the conversation back on track.  When I get to the Netherlands, I owe you a beer. 

So if we can take issue with anyone for a lack of testing to discover the "unknown unknowns," it's with NASA for not requiring testing.  We can see the same institutional attitude with the very minimal testing proposed for SLS. 

Fast, iterative testing is still the gold standard.  Testing cannot be replaced by simulations, as stated above by "Marsbase," it should only be supported by simulations.

[Star One]

When these awards were given out, Falcon 9 was pretty new (it flew a grand total of 12 times compared to 48 times for Atlas V) and went on later to have issues both in flight and on the pad. To compensate for that question mark on Falcon 9 reliability, SpaceX proposed more abort testing as mitigation. Mitigation that was probably needed.

You've got your history screwed up so I'll enlighten you:

A: NASA never required any of the CCP contractors to perform flight tests to "prove" their abort systems. Flight testing was completely optional and could be added to the CCiCAP and CCtCAP contracts as VOLUNTARY milestones.

B. Flight history (or lack thereof) of the launcher had ZERO influence on the level of required abort flight testing exactly because no actual abort flight testing was required.

C. NASA only required the CCP contractors to "prove" their abort systems theoretically. Extensive computer modeling, coupled with extensive ground-based testing was deemed sufficient by NASA to certify the abort systems.

D. The number of flights by the time the contracts are awarded has ZERO influence on the contract award. Even the number of flights at the time of the proposals being due is of little-to-no value. It is all about HOW the contractor proposes to convince NASA that it will be safe to fly on the contractors vehicle(s).

Thank you, once again, for providing facts to get the conversation back on track.  When I get to the Netherlands, I owe you a beer. 

So if we can take issue with anyone for a lack of testing to discover the "unknown unknowns," it's with NASA for not requiring testing.  We can see the same institutional attitude with the very minimal testing proposed for SLS. 

Fast, iterative testing is still the gold standard.  Testing cannot be replaced by simulations, as stated above by "Marsbase," it should only be supported by simulations.

Yet we’ve had people in the past suggest SLS is being tested too much. Seems naysayers seemingly like yourself want it both ways as far as the SLS test regime is concerned. I know SLS is really off topic for this thread but your the one that opened the door on that here.

[Athelstane]

Just to cap things off on the contracting: NASA did not 'offer options' on abort system testing plans. Bidders proposed test plans and NASA accepted them.

I think that's a better way of putting it, though one way or another ( I do not have CCiCap text in front of me) it was communicated to the bidders that NASA was happy to let them propose how they would qualify abort systems, subject to some minimal requirements, and the bidders submitted their plans. SpaceX proposed to actually do an in-flight abort, which of course was much more affordable and easy for them and meshed with their hardware-rich testing philosophy. NASA was hardly going to to say no to that.

To be honest, though, SpaceX at the time felt it was behind the 8 Ball more than, say, Boeing would be, and felt it had to go the extra mile in certain respects. Making its bid as rock bottom low as it could was one way in which they did this (probably to the point where they've actually lost money on Crew Dragon, in aggregate). Offering an IFA would be one way to drive up NASA management's comfort level, even as it also accorded with their strengths and philosophy anyway.

[ncb1397]

D. The number of flights by the time the contracts are awarded has ZERO influence on the contract award. Even the number of flights at the time of the proposals being due is of little-to-no value. It is all about HOW the contractor proposes to convince NASA that it will be safe to fly on the contractors vehicle(s).

The Source selection document says that Boeing's proposed launcher has "demonstrated reliability" while SpaceX's launch vehicle has "flight heritage". It then questions Boeing's proposal for an alternative launch vehicle as that would negate the advantages of the Atlas V baseline.

This is what it had to say about SNC's proposal:

I also agreed with the significant strength for using a launch vehicle that has demonstrated reliability and well-known failure modes and operating environments, which facilities an integrated abort system for a crew system and results in a safer launch vehicle.

https://www.nasa.gov/sites/default/files/files/CCtCap-Source-Selection-Statement-508(3).pdf

[ncb1397]

Does Boeing have to perform an in-flight abort before flying a manned mission?  If not (I haven't heard they have to) why not? 

Boeing is not doing an in-flight abort test.  They are doing computer simulations instead.  That is the certification path they and NASA agreed on.  SpaceX chose a different certification path and agreed to that with NASA.  Both certification routes are acceptable.  This has been discussed many times on the forum.

Thanks.  I did not see previous discussions.  I mostly just lurk here.

Specifically, NASA's requirements for the commercial crew program said that abort system testing must be done, along with modelling (by both NASA and the provider) to show that the testing demonstrates that the capsule can safely abort at any time in the launch sequence. The type of testing was not specified, that was left up to the provider.

When Boeing put in a bid for the commercial crew program, it included a pad abort test. SpaceX's bid had a pad abort test and an in-flight abort test. NASA accepted both bids.

Exactly.

I tend to agree with the argument that NASA ought to have required an in-flight abort test, because there's nothing like actual testing to evaluate a system. But this CCiCap was trying to build in flexibility for vendors, and they knew *these* vendors had different strengths and weaknesses. Flying an in-flight abort was cheap for SpaceX, because it knew it would have a stack of recovered Block 5 Falcon 9's, meaning their costs would be just fuel, nominal refurbishment, and any incremental ground staff labour costs. Whereas Boeing had more experience in modeling vehicle performance.

Not surprising that each vendor chose the offered option that best suited their strengths.

The launch vehicle wouldn't have made much of a difference, it would have been the difference between 4.2 billion and 4.3 billion or 2%. Anyways, the Atlas V has a very lengthy record with zero problems that would require an escape either on the pad or in flight. When these awards were given out, Falcon 9 was pretty new (it flew a grand total of 12 times compared to 48 times for Atlas V) and went on later to have issues both in flight and on the pad. To compensate for that question mark on Falcon 9 reliability, SpaceX proposed more abort testing as mitigation. Mitigation that was probably needed.

That may have been true at the time, but Falcon 9 now has 46 consecutive full successes, plus 3 Falcon Heavy full successes with significant F9 heritage. There is no longer any significant difference in demonstrated or expected reliability between the CC rides.

I'd still like to see both providers do in-flight tests.

Don't know how you came to that conclusion. Both Atlas V and Falcon 9 have had roughly the same number of flights (80 vs 77). Atlas V has had no problems that would have required an abort and Falcon 9 has had two. 46 consecutive successes isn't a significant deviation from the ~1/40 overall statistical rate.

To try to illustrate my point, Boeing/ULA didn't have to lean on the abort system like this...

https://twitter.com/elonmusk/status/771479910778966016?lang=en