-
#1400
by
Coastal Ron
on 10 Mar, 2020 01:25
-
So much for NASA oversight.
that poses the question of what NASA may have missed with SpaceX then as well. Perhaps Dragon should be grounded until NASA is "fixed."
That question has already been answered... Boeing was supposedly better.
Which was an unproven assertion, based on the assumption that because somewhere in the Boeing corporate history they had once had the knowledge to build human-rated spacecraft.
Certainly the un-crewed X-37 spacecraft has performed very well and survived a number of missions, and it comes from the same Boeing Defense, Space & Security division as the crewed Starliner does.
However despite all of that heritage the OFT flight had two software errors that occurred during the two-day flight that NASA has classified as a "high visibility close call." Plus they discovered a total of over
60 recommendations to NASA and Boeing for corrective actions that ranged from fixing these software issues to ferreting out others that may still exist in the spacecraft's flight code.Apparently there isn't any Boeing crew vehicle heritage left to use on the Starliner program, so they are kind of starting from scratch.
SpaceX has a lot of autonomous software heritage, not only for spacecraft but Elon Musk owns one of the top autonomous vehicle companies in the world (i.e. Tesla). And since Crew Dragon has Cargo Dragon heritage, and Cargo Dragon has performed without incident in 20 missions (it is in the middle of it's 20th), I'd say that should give NASA confidence that they don't need to rush a review or ground the SpaceX vehicle.
Though I think it would behoove NASA to go back and review what SpaceX has done - that makes sense in order for NASA to have enough information to validate what their processes should be in the future. The Starliner failure was as much of a "learning moment" for NASA as it was for Boeing, so reviewing how all contractors work makes sense.
-
#1401
by
ppb
on 10 Mar, 2020 02:38
-
I hope one of the IRTs 61 recommendations was to reconsider the decision to skip an inflight abort test. The prior justification for skipping it has pretty clearly been OBE.
Quam celerrime ad astra
-
#1402
by
woods170
on 10 Mar, 2020 09:54
-
I hope one of the IRTs 61 recommendations was to reconsider the decision to skip an inflight abort test. The prior justification for skipping it has pretty clearly been OBE.
Quam celerrime ad astra
Reconsidering about an IFA is not part of the 61 recommendations. Most are about procedures. Not about decisions made in the past.
-
#1403
by
Semmel
on 10 Mar, 2020 11:21
-
I hope one of the IRTs 61 recommendations was to reconsider the decision to skip an inflight abort test. The prior justification for skipping it has pretty clearly been OBE.
Quam celerrime ad astra
Reconsidering about an IFA is not part of the 61 recommendations. Most are about procedures. Not about decisions made in the past.
Did you see the list? Is there a reason why its not disclosed yet?
-
#1404
by
gaballard
on 10 Mar, 2020 17:49
-
I hope one of the IRTs 61 recommendations was to reconsider the decision to skip an inflight abort test. The prior justification for skipping it has pretty clearly been OBE.
Quam celerrime ad astra
Reconsidering about an IFA is not part of the 61 recommendations. Most are about procedures. Not about decisions made in the past.
Did you see the list? Is there a reason why its not disclosed yet?
I think someone is going to have to file a FOIA request for it.
-
#1405
by
woods170
on 11 Mar, 2020 09:17
-
I hope one of the IRTs 61 recommendations was to reconsider the decision to skip an inflight abort test. The prior justification for skipping it has pretty clearly been OBE.
Quam celerrime ad astra
Reconsidering about an IFA is not part of the 61 recommendations. Most are about procedures. Not about decisions made in the past.
Did you see the list? Is there a reason why its not disclosed yet?
No, I did not see the list. But I talked to a source who does have insight into the list. I had this specific IFA question as well, hence why I know that reconsidering the NO-IFA decision is not on the list.
-
#1406
by
edzieba
on 11 Mar, 2020 12:52
-
Can also be concluded through logic: what would an IFA be usefully testing?
The actual abort system performance has been proven during the pad abort test. Overall full system sequence simulation & testing is going to be performed now as a remedial action. The only item that has not either already been tested or is not going to be tested is the specific behaviour of the separation hardware under flight loads while part of a full stack (rather than load testing under isolation, as has been performed).
Unless the ongoing reviews uncover some new potential issue with that separation hardware that would not be satisfied with more isolated testing, then there is no more reason to perform an IFA now than there was before OFT.
-
#1407
by
JonathanD
on 11 Mar, 2020 17:04
-
Can also be concluded through logic: what would an IFA be usefully testing?
The actual abort system performance has been proven during the pad abort test. Overall full system sequence simulation & testing is going to be performed now as a remedial action. The only item that has not either already been tested or is not going to be tested is the specific behaviour of the separation hardware under flight loads while part of a full stack (rather than load testing under isolation, as has been performed).
Unless the ongoing reviews uncover some new potential issue with that separation hardware that would not be satisfied with more isolated testing, then there is no more reason to perform an IFA now than there was before OFT.
One thing I can think of is that Starliner is riding on a rocket that has strap-on solids. Flaming chunks of solid rocket fuel and parachutes don't go together very well, so it would be good to see that the integrated system performs sufficiently to get the capsule far enough away from the stage that it can safely land.
They mentioned the spacecraft had some thruster under-performance when going through the docking process (obviously it was not close to the station at that time). Presumably they tested those thrusters on the ground before hand. Point being is that flight conditions can produce very different results than ground testing.
At the end of the day these folks have to look at themselves in the mirror and decide how much an IFA test would be worth, knowing they have the lives of astronauts in their hands. I'm certainly not qualified to make that call, I just hope that they *are*.
-
#1408
by
abaddon
on 11 Mar, 2020 18:45
-
Can also be concluded through logic: what would an IFA be usefully testing?
The actual abort system performance has been proven during the pad abort test. Overall full system sequence simulation & testing is going to be performed now as a remedial action.
Overall full system simulation was always the Boeing proposal for in-flight abort verification. I was never a very big fan of that choice myself, and the recent reveals that Boeing simulation was gravely lacking has not helped that perception generally.
Nevertheless, it was Boeing's proposal and I can't imagine that anyone officially involved would suggest an IFA is now required. It does seem reasonable to requalify the IFA via simulation after the conclusion of the remediation actions identified as part of the independent review.
-
#1409
by
lrk
on 11 Mar, 2020 18:51
-
Can also be concluded through logic: what would an IFA be usefully testing?
The actual abort system performance has been proven during the pad abort test. Overall full system sequence simulation & testing is going to be performed now as a remedial action. The only item that has not either already been tested or is not going to be tested is the specific behaviour of the separation hardware under flight loads while part of a full stack (rather than load testing under isolation, as has been performed).
Unless the ongoing reviews uncover some new potential issue with that separation hardware that would not be satisfied with more isolated testing, then there is no more reason to perform an IFA now than there was before OFT.
I think the rationale being raised by some people on this site is that the in-flight abort behavior was being modeled by Boeing, and the results of the CFT reflected poorly on Boeing's ability to simulate things.
Not that I completely agree with that rationale (design modeling and Hardware-In-the-Loop software validation are very different things), but I am rather surprised that NASA was satisfied with them not doing an IFA in the first place. Supersonic separation and subsequent interaction with the abort engine plumes is already a really complex problem to model from an aerodynamics perspective, not to mention validating their control algorithms as the capsule is not aerodynamically stable and relies on the thrusters for control. Apparently they have done enough to satisfy NASA but there still is a non-zero chance that they missed an unexpected interaction that only shows up at supersonic speeds.
-
#1410
by
dglow
on 11 Mar, 2020 19:22
-
... and the recent reveals that Boeing simulation was gravely lacking has not helped that perception generally.
Which reveals do you refer to?
-
#1411
by
whitelancer64
on 11 Mar, 2020 19:45
-
Can also be concluded through logic: what would an IFA be usefully testing?
The actual abort system performance has been proven during the pad abort test. Overall full system sequence simulation & testing is going to be performed now as a remedial action. The only item that has not either already been tested or is not going to be tested is the specific behaviour of the separation hardware under flight loads while part of a full stack (rather than load testing under isolation, as has been performed).
Unless the ongoing reviews uncover some new potential issue with that separation hardware that would not be satisfied with more isolated testing, then there is no more reason to perform an IFA now than there was before OFT.
I think the rationale being raised by some people on this site is that the in-flight abort behavior was being modeled by Boeing, and the results of the CFT reflected poorly on Boeing's ability to simulate things.
Not that I completely agree with that rationale (design modeling and Hardware-In-the-Loop software validation are very different things), but I am rather surprised that NASA was satisfied with them not doing an IFA in the first place. Supersonic separation and subsequent interaction with the abort engine plumes is already a really complex problem to model from an aerodynamics perspective, not to mention validating their control algorithms as the capsule is not aerodynamically stable and relies on the thrusters for control. Apparently they have done enough to satisfy NASA but there still is a non-zero chance that they missed an unexpected interaction that only shows up at supersonic speeds.
NASA only required that testing be done to validate modeling that showed that the capsule could perform an abort at any time in flight. NASA did not specify what that testing should be, that was left up to the provider. So Boeing chose to put ground tests and a pad-abort test in their bid. SpaceX chose to put ground tests, a pad-abort test, and an in-flight abort test in their bid. NASA accepted both bids.
To be clear, both Boeing and SpaceX are doing modeling to show that the abort capability exists at other times in flight.
I personally feel that NASA should have required an in-flight abort test, but that's a decision very long past.
-
#1412
by
abaddon
on 11 Mar, 2020 20:04
-
... and the recent reveals that Boeing simulation was gravely lacking has not helped that perception generally.
Which reveals do you refer to?
I meant end-to-end testing (which involves simulation) but I must confess I don't know of anything specifically about Boeing simulation with regards to the IFA, so I have to withdraw that comment. The general perception that Boeing software practices in the Starliner program have been substandard still taints this aspect, in my opinion.
Presumably re-qualifying the simulated IFA after these issues are resolved is a prudent and hopefully not very expensive step to take.
-
#1413
by
erioladastra
on 12 Mar, 2020 14:15
-
Can also be concluded through logic: what would an IFA be usefully testing?
The actual abort system performance has been proven during the pad abort test. Overall full system sequence simulation & testing is going to be performed now as a remedial action. The only item that has not either already been tested or is not going to be tested is the specific behaviour of the separation hardware under flight loads while part of a full stack (rather than load testing under isolation, as has been performed).
Unless the ongoing reviews uncover some new potential issue with that separation hardware that would not be satisfied with more isolated testing, then there is no more reason to perform an IFA now than there was before OFT.
They mentioned the spacecraft had some thruster under-performance when going through the docking process (obviously it was not close to the station at that time). Presumably they tested those thrusters on the ground before hand. Point being is that flight conditions can produce very different results than ground testing.
This is confusing some of the information. Boeing conducted an "abort demonstration" (a test that if they were near the ISS - which they did not get near on OFT - the vehicle could perform an abort and get away safely). By design this test was done in configuration where all systems were on (to save power). Because of this, a thruster problem, which normally would be minor (and related to the issues getting into orbit), caused degraded performance until the ground team reactivated all the systems.
-
#1414
by
LastWyzard
on 12 Mar, 2020 21:22
-
I wonder where Boeing sits financially with this project. I know that NASA pays whenever pre-determined criteria are met. Presumably Boeing feels this project is profitable or they would try to squeeze more money from NASA or drop the whole thing entirely. Is it currently profitable or do they need to wait for a few missions before being in the black? With other projects in trouble (MAX, KC-45A) I would think they would purge everything that looks like a loss.
I work a couple of blocks from their headquarters. Walking up, knocking on the door and asking a few questions probably wouldn't get me anywhere. They do have a gift shop where you can purchase a nice Starliner model.
-Ron
-
#1415
by
xyv
on 13 Mar, 2020 02:21
-
... and the recent reveals that Boeing simulation was gravely lacking has not helped that perception generally.
Which reveals do you refer to?
I meant end-to-end testing (which involves simulation) but I must confess I don't know of anything specifically about Boeing simulation with regards to the IFA, so I have to withdraw that comment. The general perception that Boeing software practices in the Starliner program have been substandard still taints this aspect, in my opinion.
Presumably re-qualifying the simulated IFA after these issues are resolved is a prudent and hopefully not very expensive step to take.
OK it seems several posters are conflating simulation of aerodynamic events (IFA) with software development. The first is the domain of the analysts who use software and the second has to do with actual software development. I actually have good confidence in Boeing's ability in the former. They knew exactly what would happen with the Max under the loss of an AOA sensor...aerodynamically speaking.. they failed in their software design in anticipating what to do.
I doubt Boeings' analysis ability needs to be wrung out (I don't actually know and could be wrong here) but their software development process is clearly broken.
-
#1416
by
abaddon
on 13 Mar, 2020 02:25
-
I’m talking about simulating the abort, in other words simulating the behavior of the craft, running the software etc as the system executes an abort. I’m not talking about modeling.
-
#1417
by
xyv
on 13 Mar, 2020 02:36
-
Yea.. different problem. Obviously if they haven't done a full hardware in loop simulation of the abort sequence, that needs to be rigorously done. But an in flight abort is a heavy handed substitute for full coverage testing of the system across multiple scenarios.
-
#1418
by
john smith 19
on 13 Mar, 2020 09:52
-
Yea.. different problem. Obviously if they haven't done a full hardware in loop simulation of the abort sequence, that needs to be rigorously done. But an in flight abort is a heavy handed substitute for full coverage testing of the system across multiple scenarios.
True.
In principle a
very large number of ground tests could be done before you came
anywhere close to the cost of a full in flight abort test, even assuming each test loop had the (very expensive) flight hardware versions of components and not simulators standing in for them.
At one point in the telcom Boeing admit they did not have enough copies of the flight hardware (the propulsion controller was removed from this test and used on another test. Had there been two the mis-mapping would have been picked up on the ground). They were "hardware poor."*
An interesting question would be "Where did the thruster mapping in the simulator
come from?"
Was it in the documentation? If so it doesn't match the prop controller
Was it a misread of the documentation?
Did someone just make it up to get the SW working, figuring to fix to the correct values when they had a moment?
At bottom we're talking about one or more addresses in the memory map and
exactly which thruster would fire if a 1 is written to a certain bit in those words. In some sense it's not about
what they were but wheather they were documented and why didn't the simulator match the existing ("legacy") hardware?
And the questions people
should be asking (as per CMM5 practice) is how did this get through our build process, and how many
other examples are there?
Only then should they be thinking about a new IFA test as the final demonstration they have system for creating systems that work.
*
Despite pitching a 63% higher bid for the project.
-
#1419
by
Kansan52
on 13 Mar, 2020 15:23
-
OK, if we ran a poll asking "If there will be an OFT" (instead of 'should be') will the results be close to opposite of the current results?
Just musing.
The precedent that NASA may use is that memory says that the first two test flights of the Cargo program were combined to try to move the schedule to the right as much as possible.