4 ms·

Is anyone here familiar with how flight testing is actually performed? One thing I wonder about after reading this is why, if they were able to recreate the en

by gnoway 13y ago

Is anyone here familiar with how flight testing is actually performed? One thing I wonder about after reading this is why, if they were able to recreate the entire flight profile in a simulator the next day and see the same result, would they not just test everything in the simulator first?

InclinedPlane 13y ago
You can't test everything. There are so many independent variables, to try to test every possible scenario is a combinatorial nightmare. More so when you factor in transient events like engine unstarts and so forth.
skj 13y ago
Because testing every sequence would take exponential time relative to the length of the sequence. For the post-accident simulation, they say "ok since we know we are in this modification of the SR-71, in this phase of the flight, and this accident happens, and this backup fails... ok yeah it looks like that's pretty bad.
- gnoway 13y ago
  I misunderstood the article I guess. I thought they knew specifically what they were testing, i.e. 'we have altered the airplane this way and we will now do these things.' After the failure, they were able to essentially reconfigure the simulator to match the test parameters, rerun the test and observe the same failure in simulation.
  - skj 13y ago
    The failure was not inevitable. Two things happened to go wrong at the same terrible time.
GICodeWarrior 13y ago
I am sure they ran some sort of simulation first. However, "the right engine inlet's automatic control system malfunctioned". That probably wasn't part of their simulation. While it sounds like they could have erred on the side of caution and cancelled the remaining tests after that occurred, they must have been confident in their manual control abilities. Even if they had simulated with the inlet(s) under manual control, real life airflow and reaction times are going to vary.
robryk 13y ago
I'd imagine that they knew the exact timing of the failure from the telemetry from the plane.

by gnoway 13y ago

InclinedPlane 13y ago

You can't test everything. There are so many independent variables, to try to test every possible scenario is a combinatorial nightmare. More so when you factor in transient events like engine unstarts and so forth.

skj 13y ago

Because testing every sequence would take exponential time relative to the length of the sequence. For the post-accident simulation, they say "ok since we know we are in this modification of the SR-71, in this phase of the flight, and this accident happens, and this backup fails... ok yeah it looks like that's pretty bad.

gnoway 13y ago
I misunderstood the article I guess. I thought they knew specifically what they were testing, i.e. 'we have altered the airplane this way and we will now do these things.' After the failure, they were able to essentially reconfigure the simulator to match the test parameters, rerun the test and observe the same failure in simulation.
- skj 13y ago
  The failure was not inevitable. Two things happened to go wrong at the same terrible time.

GICodeWarrior 13y ago

I am sure they ran some sort of simulation first. However, "the right engine inlet's automatic control system malfunctioned". That probably wasn't part of their simulation. While it sounds like they could have erred on the side of caution and cancelled the remaining tests after that occurred, they must have been confident in their manual control abilities. Even if they had simulated with the inlet(s) under manual control, real life airflow and reaction times are going to vary.

robryk 13y ago

I'd imagine that they knew the exact timing of the failure from the telemetry from the plane.