CEFOSKEY
With due respect to you, I think you make my point quite well when you say
... seems to be under the belief that designers should be aware of failure modes they dont anticipate!
- that's just the point, they can't be, so how can they be reasonably sure that there predicitions that a critical failure will be less than 10^-8?
Yes I do consider fatigue calculations to at least have an element of guesswork - at a simple level because a certain flight profile (in terms of power settings, manoeuvring and gust loads, number of takeoffs/landings, weight, C of G etc) is assumed for the life of the component. But in fact the actual profile spectrum is at the discretion of the pilots operating the aircraft long after the designer has moved on to another project.
And at a more complicated level because its not possible to exactly model the entire aircraft and its environment. Approximations and simplifications are made (I am not a fatigue engineer but I would suspect that eg some kind of finite element analysis is used?). And as you say, defects and oversights can (I would say
will) occur.
Perhaps I should make my point clearer - I am not saying that design engineers are stupid, rather that the certification rules assume they can do something that in fact they cannot.
Yes the 92 hasn't killed anyone yet but it does seem to have had more than its fair share of serious events in its short life so far, at least 4 major transmission lubrication events, one or two engine failures (caused by the engine installation, not the engine itself), problems with tail rotor control and probably some others I can't immediately recall.
But I am drifting - really my point was to question the robustness of the certification procedure, and perhaps to wonder how an aircraft whose certification was based on certain failures being extremely remote, can continue to be considered airworthy when those failures are demonstrated to be much more frequent.
HC