This is a failing of engineering reliable systems:
A part is tested to a point of failure, based on the quality parameters you specify. I'd put safety pretty high on a 115,000V line. Choose your sigma level based on your goals. I'd wager 97 years is out of the ballpark of "Yes, this might fail". And to save on cost, don't come back every year "OK, it's still OK." Replace it - it saves cost (and in this case, potentially lives).
So yes, I'd assign them blame. Be it an engineering part, an asphalt road, a power pylon.
There will be outliers, but this was an outlier of systematic negligence, with no (recorded) documentation prior to the year 2000.
Scheduled replacement saves costs in the long term, if bottom dollar is the aim.
Clearly it broke. Was fixed, sort of but not actually, and failed.
Observe, test, measure, analyse, mitigate (did I create a new acronym to add to the vast QC vocabulary, OTMAM?). Scheduled replacement! Mitigate problems before they cause knock-on effects that shutdown a production line or endanger lives and livelihoods.
A part is tested to a point of failure, based on the quality parameters you specify. I'd put safety pretty high on a 115,000V line. Choose your sigma level based on your goals. I'd wager 97 years is out of the ballpark of "Yes, this might fail". And to save on cost, don't come back every year "OK, it's still OK." Replace it - it saves cost (and in this case, potentially lives).
So yes, I'd assign them blame. Be it an engineering part, an asphalt road, a power pylon.
There will be outliers, but this was an outlier of systematic negligence, with no (recorded) documentation prior to the year 2000.
Scheduled replacement saves costs in the long term, if bottom dollar is the aim.