The strut that we believe failed was designed and material certified to handle 10,000 lbs of force, but failed at 2,000 lbs, a five-fold difference. [...] these struts have been used on all previous Falcon 9 flights and are certified to withstand well beyond the expected loads during flight [...]
Huh. So the strut was way over-engineered for the load and still failed, likely due to a hidden mechanical flaw. Would that make more sense to use a large number of smaller struts then, such that a failure of one or two would be less problematic? You can even test the design for it - make sure the rocket is still working when certain number of struts fail. The assembly cost would certainly rise, but the current design seems to be not sufficiently redundant.
They were using a large number of small struts, a few hundred per rocket. The problem with redundancy through quantity is that occasionally one of those many redundant parts will fail in an interesting way, like happened here, and cause that one in a million cascade. The trick is finding just the right amount of redundancy for the task at hand. Sometimes that's a few strong pieces, sometimes it's a lot of little pieces.
Overengineering by a factor of 5 seems decently conservative.
If you add triple-redundancy to all the mechanical elements you build a rocket that probably can't fly - which is certainly very safe, but maybe not so very useful.
I'd guess the problem has more to do with insufficient testing than insufficient design. It's likely the initial batch was easily up to spec, then there was little or no testing of further batches.
Easy mistake, IMO. If you think about the number of critical components, re-testing everything before every flight would be a huge and very expensive job.
I wonder if SpaceX are going to do less business with the strut supplier in future.
Huh. So the strut was way over-engineered for the load and still failed, likely due to a hidden mechanical flaw. Would that make more sense to use a large number of smaller struts then, such that a failure of one or two would be less problematic? You can even test the design for it - make sure the rocket is still working when certain number of struts fail. The assembly cost would certainly rise, but the current design seems to be not sufficiently redundant.