From reading the article it sounds as though he just tossed it after disassembly, but perhaps they’ve replaced parts similar to this before.
What I meant to say was “assuming they had been well characterize prior to *installation*”. It is likely that all the data they took prior to installation was “meets spec.” In addition, they would have to have created a “control” group assembly and have it subjected to the same periodic temperature regime. I doubt their is a whole lot of valuable engineering data to be gotten from the old assembly.