Average Life of Power Supply ?
I'm not sure how PSU manufacturers determine MTBF. But here's how HD manufacturers do it. Perhaps there's a similarity:
A HD manufacturer will set up a large number of HDs and let them run continuously. The number of failures is noted, as well as the total run time of the HDs. If, say, 1000 HDs run for three months, and there are 4 failures within that period, then the computed MTBF would be
1000*24*90/4 = 540,000 hours.
Note that this number doesn't mean that a SINGLE HD will typically run for 540,000 hours before failing. We'd be able to judge better the likelihood if failures if HD manufacturers told us the actual test results from which they computed MTBF.
Anyway, a high MTDB means that the odds that a given HD will survive to its "useful life" are pretty good. "Useful life" is another term that can be found in HD technical specs. For example, I think Seagate rates the useful life of it's ATA HDs at five years (coincidentally, also the length of Seagate's warranty). Beyond the useful life, the odds of failure increase.
PSU manufacturers also probably have a "useful life" spec of some sort.