The Boeing 737-NG has been in the news recently regarding cracks discovered in the "pickle fork" assembly which transfers loads from the wings to the body of the aircraft. The "pickle fork" was supposed to last the lifetime of the air frame (90,000 cycles), but cracks were discovered in airplanes with less than 30,000 cycles. Could the cracks be due to an "over G" situation where turbulence or a hard landing could have caused the airplane to exceed its load limit, or is this likely due to a design flaw?
Depends on the crack characteristics, number of striations (individual propagation events) etc., whether the cracks are from a small number of near-yield overloads or many small loads from regular cycles.
This sort of thing happens quite a lot. You are designing parts at all times to be "just strong enough" for all the various requirements and goals, since every pound over that is considered ballast. Computer finite element analysis makes it possible to fine tune the structure during design and reduce the fudge factor needed in the slide rule days.
Endurance fatigue testing of the part is used to validate the finite element analysis. A 3:1 safety or scatter factor is applied in the fatigue endurance test. To be certified for 90,000 cycles, the part had to last 270,000 cycles crack free in the fatigue rig.
As a back stop to all that, you have a structural inspection program to cater to the possibility that the scatter is bigger than expected or the analysis was flawed and wasn't revealed by the original endurance test. Everything has variability, generally on a pareto distribution (bell curve).
If you only endurance test one or two articles, you can't really be sure where those two parts would fall on the bell curve. You assume it's near the middle, and the inspection program covers your butt if you are wrong. Which is the case here probably. You could test a large number of forks instead of a couple to reduce the scatter factor risk, but $.
You'd need to see the lab reports for the Pickle Forks (I've read many). There may have been microscopic flaws in manufacture (pits or nicks or voids in a forging) that created crack initiation points. It may be a batch issue of a defect like that on x number of parts. Or the original analysis was off, and the fatigue testing got units that represented the right side of the bell curve (low probability) and gave an optimistic result, so the part was too skinny for the job, from a fatigue perspective, from the get go.
The latest findings of cracks in less than 20000 cycles (Lion Air, 2 planes), and 13% of cracked pickle forks in South Korean fleet are beginning to look really bad. This is a design flaw, no doubt about it.
As for how serious of a design flaw, I cannot judge. If the cracks would impose an immediate risk to flight safety, I bet the officials around the world would have grounded the whole fleet by now. Lets wait and see.