Submain rcd fault protection

[bigspark17]

If a 2 core swa was run for a submain to an outbuilding with no extraneous parts in it, but the zdb reading was well over the maximum earth fault loop impedance. Could an rcd time delay be installed at supply end to cover the submain ie increasing its new maximum zs reading?
Hope thats clear ?

 

[Ian1981]

An rcd could be used for earth fault protection yes. As it’s a time delay then the disconnection time must be within 1 second.
Time delay rcd’s must disconnect between 500-130 mS anyway.

 

[Andy78]

I'd look at all other alternatives before resorting to RCDs for fault protection.

 

[pc1966]

Yes, it would give you the protection for cable faults to ground.

Ideally you ought to be checking also for the L-N fault case and if the cable adiabatic value is survivable, as the RCD won't help with that.

But is it simply a case of poor design (i.e. armour impedance too high over that length) or is the measured impedance (R2 & Ze) not adding up so a bad connection, etc, is possible?

 

[bigspark17]

An rcd could be used for earth fault protection yes. As it’s a time delay then the disconnection time must be within 1 second.
Time delay rcd’s must disconnect between 500-130 mS anyway.

what would the Max zs be if a 100ma time delay rcd were fitted @ origin to cover the sumain?

Yes, it would give you the protection for cable faults to ground.

Ideally you ought to be checking also for the L-N fault case and if the cable adiabatic value is survivable, as the RCD won't help with that.

But is it simply a case of poor design (i.e. armour impedance too high over that length) or is the measured impedance (R2 & Ze) not adding up so a bad connection, etc, is possible?

it is just poor design. Distance way to far to meet fuse disconnection times.

 

[pc1966]

what would the Max zs be if a 100ma time delay rcd were fitted @ origin to cover the sumain?

500 ohms.

it is just poor design. Distance way to far to meet fuse disconnection times.

OK. Sometimes you need another CPC connecting in parallel with the armour (connected at both ends!) to get an acceptable end Zs, more so for larger cables. Can be a 3rd core or an external run (if acceptable by the physical environment).

 

[Ian1981]

what would the Max zs be if a 100ma time delay rcd were fitted @ origin to cover the sumain?




it is just poor design. Distance way to far to meet fuse disconnection times.

The maximum Zs would be 500 ohms.
50/0.1

 

[davesparks]

As it’s a time delay then the disconnection time must be within 1 second.
Time delay rcd’s must disconnect between 500-130 mS anyway.

The maximum Zs would be 500 ohms.
50/0.1

That is only true for a TT system, the OP. hasn't specified the earthing system.

For a TN system regulation 411.4.4 applies so disconnection time would be 5s as it is a distribution circuit and max Zs would be 2185 ohms from 218.5/0.1

 

[bigspark17]

Correct @davesparks it is a TN system, i suspect the plan may have been to install a TT system as the outbuilding has upfront time delay main switch followed by full rcbo cu.

thanks for info brain wasnt working lastnight. So im left with 4 options to get this install to comply.
1- TT the outbuilding
2- time dealy rcd 100mA on the submain @ origin and possibly remove the time delay rcd in outbuilding and replace with standard 100A 2pole main switch.
3 change the Swa submain (not gunna happen)
4 run a parallel earth To bring disconnection times down.

 

[pc1966]

So im left with 4 options to get this install to comply.
1- TT the outbuilding

That is not going to protect the sub-main cable against a fault.

Already the building is safe internally as the RCBOs (and RCD incomer) will make sure all final circuits meet disconnection time no problem. The sub-main R2 is still going to be way lower than any earth rod's Ra value!

2- time dealy rcd 100mA on the submain @ origin and possibly remove the time delay rcd in outbuilding and replace with standard 100A 2pole main switch.

I suspect the original designer might have been aware of the high R2 and aimed at protected the installation that way, but not the feed cable, alas.

Moving the delay RCD to the cable origin would solve those problems, but puts resetting it in a more inconvenient place. However with an all-RCBO board the chances of the supply delay-RCD firing should be very slim as practically all faults should be cleared by the RCBOs.

3 change the Swa submain (not gunna happen)

No, that really is a last-resort option and only justified if the volt drop or whatever was unacceptable.

4 run a parallel earth To bring disconnection times down.

If the cable is ducted or on tray, etc, so you think it is feasible to run a CPC alongside then it might be the best overall option as the out building would retain access to reset the delay RCD and no need to find space for another enclosure or whatever at the source end.

If you have the cable length & type and have checked the measured R2 is consistent with it, then it ought to be simple to compute the new R2 and then Zs with a CPC of, say, half the phase conductor size run along and see if it then meets a 5s disconnection limit.

 

[pc1966]

Actually thinking about it a bit more: going TT would be a fix as the whole disconnection time requirement is for shock protection and with a TT earth rod a phase-armour short is not going to impact anyone (if the armour is inaccessible).

Of course with long disconnection times there is a risk of the cable frying on a short, but you could argue that a phase-armour short on SWA is pretty much part of an end-of-cable event anyway.

So that is option to revisit as it is even simpler to do, providing it is easy/safe to drive in a rod or two at the out building.

 

[Ian1981]

For a TN system regulation 411.4.4 applies so disconnection time would be 5s as it is a distribution circuit and max Zs would be 2185 ohms from 218.5/0.1
[/QUOTE]
@davesparks
Regulation 411.4.204 explains that table 41.5 maybe applied for maximum values of earth fault loop impedance where an rcd is used to satisfy the requirements of 411.3.2.2 , 411.3.2.3 and 411.3.2.4
Where your disconnection time maybe 5 seconds for a TN arrangement, the maximum values of earth fault loop impedance from table 41.5 maybe applied.

 

[pc1966]

Thinking about it again, the sub-main protection is something to discuss as most likely it has to be protected in case some poor sod puts a pick-axe through it or similar, so the disconnection times on a cable fault need to be met.

So sounds like it is back to either (a) move the delay-RCD to the cable feed, or (b) add a supplementary CPC so the end Zs allows disconnection within the regulation times.

 

[bigspark17]

Thanks @pc1966 For giving it so much thought. I feel that the submain does need to Be protected/meet Disconnection times. A parralel cpc is not an option. I agree that with the outbuiding fully rcbo covered that the time delay rcd on the sub really shouldnt be causing issues. I am going to leave the outbuildig as is (ie not move the time delay rcd) but just add another as origin to protect the swa submain.

where does the Initial figure 218.5 Come from to put into the 218.5/0.1? = 2185ohms. (Max disc time for TN)

 

[Ian1981]

Uo X Cmin
230 X 0.95