It is an interesting question. Which fault occurs most? Would like to have the stats for that one. However, rather than choosing supply type based on "which is less dangerous", what about focusing on how the various supply types might be improved. For instance some European countries use a TT system where the fault loop impedence is low enough to disconnect the OCPD. Why not consider that option?. RCD failing sorted!!
Why have we not explored an open PEN sensor to operate over the complete electrical installation in the event of an open PEN FAULT?
Yes I would very much like to know how often PEN faults occur, not sure how to obtain that information although
I like your thinking of a good TT system with a very low impedance electrode
South Africa apparently use an RCD BUT before the point where the CPC for the installation is joined to the PEN, so if there is a PEN fault and current is returning via the CPC by a person or extraneous conductive part there will be an imbalance between the line and PEN and disconnection will occur. And in theory a line - earth fault when the PEN is intact should not trip the pen protection RCD as any current returning on the CPC will also go through the PEN, although if there is a line - earth fault but not via the CPC it would trip the whole installation.
Anyway doing this in the UK is prohibited.
(although stop me if I'm wrong but I think some EV chargers are using current clamps around the supply cable in this way, and effectively doing the same thing but only disconnecting the EV)
If you have access to all 3 phases, you can wire 3 resistors in a Y configuration between the phases to give you a virtual star point, then if the voltage is monitored between this virtual star point and the PEN and it goes above 50v then an MCB shunt could disconnect all line conductors and earth within the installation eliminating the risk (this is how the 3 phase version of the matt:e o-pen works I think). obviously, this can't be implemented for domestic at individual houses as most are only single phase but if devices like this were installed in the distribution network it could reduce PEN fault risks, but it would cost a fair bit and I would imagine the reduced copper cost is why we have PME in the first place so doubt the DNO would like to spend more. Also important to note that you must have all 3 phases for the virtual start point to work, but its easy enough to use a voltage sensing relay to disconnect should you lose just one phase etc
I have been thinking of another way to monitor for a PEN fault; would be to have a device performing a Ze or even a loop test between L - N very quickly (every 50ms or so) and if the impedance went up it would indicate a PEN fault, even if the PEN fault is on a balanced 3 phase system the device would then be measuring the impedance through the connected loads on the other phases and therefore there would normally be a higher impedance.
