non event.
Why this news is "made public", and why does this thread exist? Because the destructuve tests are far more news worthy than the non-destructive tests.
Is it normal practice to validate an FEA model based on empirical data? Yes, absolutely.
Is 3% a significant error margin? Not really, what is significant is what is done with the result afterwards.
Should we be more interested in details such as the crack propogation rate that was caused by the non destructive cyclic testing (that represents real world loading scenarios)? Yes, but this is dull so won't get the same tabloid attention.
What I would like to know is if the component(s) that failed in these tests were the same component(s) that first exceeded their elastic limit in the FEA modelling.