Originally Posted by G-EMMA
IFMU, The AAIB report for G-BKCB looks robust enough, they didn't find fatigue, corrosion, or anything pre-existing as in my opinion it wasn't there to find. The finite element model they created is convincing,...
Originally Posted by SNS3Guppy
Metal remembers. Fatigue is cumulative (which is why it's called fatigue). Cracks grown, damaged areas are weakened...and mechanical things do not repair themselves. Problems don't go away. They just hide for a while.
I spent a part of my aerospace career in test. We would chug on metal parts until they busted, then send them out to the metlab. The metab would put the parts under a microscope, look at grain stucture, origins, etc. Under the microscope you could see what parts were ripped apart by fatigue, and then as the part got weaker static overload would take over and finish the job up quickly, assuming the initial fatigue damage was not detected with non-destructive test. We would take all the test data and figure out s-n curves, to help in predicting the life of parts. Miner's law, cumulative damage and all that.
What struck me about the G-BKCB report was the one liner:
Also, no evidence was found in the failed area of the wing of corrosion, fatigue cracking, repairs, material defects, or any other feature, which could have degraded the strength of the structure.
Maybe the investigator turned the pieces over to the metlab, they did the detailed analysis, electron microscope study, etc of the failed parts and the terse one liner came out of the results of that report. Or maybe not. Why would they do a detailed analysis of the metal, and not show it, then do a detailed FEM analysis, and present it in gory detail?
I have seen a lot of FE analysis. It's a great tool. But it is not always right. I've seen it both ways, where somebody comes along and does a FE analysis of a 30 year old part with a million fleet hours on it and says it only has a 5 hour life. I've also seen parts that were shown to have huge margin in analysis end up with a limited life and a redesign. Even if the FE model for G-BKCB was spot-on, it seems all that it did was to validate the static strength of the PA28R spar splice area. I bet that part was designed, static tested, and shown to have margin over the standard category g-loads a long time ago.
-- IFMU