Originally Posted by
esa-aardvark
Dave Therhino
Shuttle - I think you mean per flight. Back when I worked the figure was 1/100 per flight.
I think the real concern is that Boeing programmed their flight software - MCAS
to rely on a potentially unreliable sensor. MCAS was not able (programmed) to detect
the failing sensor. Perhaps MCAS should have had a sanity check on reported AOA.
My previous comment should have indicated that the expectation of an event is in the numbers
that you (NASA or Boeing or FAA) should be calculating with.
I agree the number is also a good approximation of the per cycle risk because launch and re-entry were by far the most risky phases of flight. There were something like 150 missions (it's been a while since I added them up for a study), and two accidents, one on launch and one on re-entry. I specifically gave a ballpark per hour rate for launch and re-entry, for which I guessed 1 hour of exposure per cycle to those phases of flight. The risk in orbit is much lower, but still quite high theoretically compared to a commercial flight, though the record was good for those 150 or so flights.
I'm not sure I'm grasping everything you were trying to say with the rest of your comment, but my point was the AOA sensor behavior in the max fleet was not a huge outlier failure-rate wise, and it was the system architecture that caused the problem. I think we are probably saying the same thing.