Originally Posted by
Ian W
A lot depends on how far down the probability tails you want to proceed. In almost all FMEAs you get to conditions occur only where several highly improbables occur concurrently in a highly improbable sequence. It gets very difficult to persuade system developers to redesign systems to cope with probabilities in the 10^-15 range.
OK Ian! Yes, you are correct. However, for simple folk, like me, we work in a simple world where the faults in aircraft electrical power systems work something like this.... Short circuit in a system=CB Trip. Overload of a Busbar=Isolation or,=Auto load shedding (normally alternate systems avail on another Bus). T/R fail=Isolation (also, normally alt sys avail). Gen Fail= Isolation, (normally single Gen fail accommodated with alt source select and, no services lost). Multi Gen fails, protected by isolation of Busbars to protect partial remains of system with working generators and avoid overload. Batt fail, can be complicated but, generally systems work as long as engines/Generators working. etc...
Therefore, rather than talking mathematics, it would be great to know what failure mitigation worked or, didn't work in this case?