Svenestron,
That is an extremely interesting contention!
How did you manage to establish that the event was 'extremely remote'; empirical knowledge would have been difficult to provide as it would have had to show that there had been no failure of any lines, fittings, valves, coolers, etc. over a period of 10,000,000 to 1,000,000,000 hours.
The very fact that, as squibb66 has pointed out, there were two 'loss of oil events' to the S92 type over a period of six months would indicate that making such judgements is fraught with danger.
I cannot imagine that the regulator would be sympathetic to such contentions in the future.
My interpretation of the intent of the rule would be
no more accurate than yours but I would consider it to be 'unless such failures
can be established to be extremely remote'. However, that was before the S92 accident; it might be difficult, since the accident, to contend that such failures are 'extremely remote'.
'Semantics' is not the way to address engineering issues:
Sematics: The individual meanings of words, as opposed to the overall meaning of a passage.
Jim