Phase failure relays (pfr)

It may be. Usually a single phase is lost if there are fuses,  rather than breakers, as breakers are ganged - while one fuse may fail before the other two, leaving motors unsure which way to rotate,  which is where the protection comes in.

 Therefore the detection needs to go anywhere after the last place that a single phase isolation  like that  may credibly occur, and before the load that actually cares if it does not get all three phases at once.  Nearer the load makes it easier to maintain, and keeps the fault levels that have to be interrupted to a lower level, as well as covering for single phase faults at all points further upstream.

The standby supply may make the fault tree harder to follow, but as I'm sure that will change over on all 3 phases together, the end result is the same as if there was one supply,

Mike.

It may be. Usually a single phase is lost if there are fuses,  rather than breakers, as breakers are ganged - while one fuse may fail before the other two, leaving motors unsure which way to rotate,  which is where the protection comes in.

 Therefore the detection needs to go anywhere after the last place that a single phase isolation  like that  may credibly occur, and before the load that actually cares if it does not get all three phases at once.  Nearer the load makes it easier to maintain, and keeps the fault levels that have to be interrupted to a lower level, as well as covering for single phase faults at all points further upstream.

The standby supply may make the fault tree harder to follow, but as I'm sure that will change over on all 3 phases together, the end result is the same as if there was one supply,

Mike.