THE CAMPAIGN FOR REAL EARTHING

I certainly agree that a TN-S system with separate, not feebly-sized, PE conductor with multiple earthing is preferable to the existing 'PME' for shock safety, though not in initial cost. If allowed to work on feeling and technical quality I'd choose it for new build and renovation. I'm not confident how it would compare in cost/benefit to other measures that could be taken with electrical or other safety. Estimated figures for the extra cost would be interesting. Perhaps just saving the cost of concerns about car-charging could make it worth it in new builds.  Then the trouble is that different parties bear the two costs, so, as John notes, it's the regulator that needs to lay down the rules for utilities. With some regret I think it's unlikely to happen. If it did, I wonder how soon it would be before the environmental advantage (lower loss, "more potential for PV integration" etc) of bonding N and PE at various places would be discovered and implemented.  


The only safety advantage I see for TN-C-S is that multiply-earthed neutral helps to reduce equipment damage from overvoltage during a broken neutral (as was already mentioned in this thread).  If this damage could involve fire (which has been a discussion point at least in Scandinavia, and which I've heard claimed from Australia) then there's a safety aspect besides just cost of equipment damage. Some AFDDs (arc-fault detectors) include disconnection of their circuit at 270V, besides their AFDD and RCBO functions. On the down-side, this arguably makes dreams of gettings adequate broken-neutral safety with "fairly well balanced loads" harder to achieve, if the lightest-loaded phase will trip completely as soon as there's a moderate imbalance that pushes voltages just beyond their normal range.


'MEN' (AU/NZ): practically, it's PME by another name. Perhaps slightly different rules for the installation regarding whether an earth electrode is required at the installation. Correspondence with a retired Australian utility engineer has done nothing to reassure me that their implementation avoids any of the familiar troubles of a system in which a single break in a conductor that's not subject to regular inspection and testing can cause immediate danger due to normal loads. I wasn't aware of the 2018 accident (in an earlier link), but am not surprised. They've had much attention to broken neutrals, e.g. in Sydney and Tasmania.  Sydney plumbers are urged to take plenty of precautions. In Tasmania the utility is still offering what's basically a plug-in loop-tester with alarm, for permanent use by households. If the ground resistivity is really too high for TT to work nicely with modern RCDs, then: a) I agree with earlier commenters about the lack of shock danger, and b) it's clearly a waste of effort hoping that an earth electrode on the neutral will 'control' the potential if the neutral breaks on the supply side. 


Ufer (concrete-encased electrode): it's a person's name, not an acronym (cf. U.F.E.R. in an earlier comment).

SWER: yes, they seem to have tamed this adequately in some sparsely populated places. The linked presentation asserts successful use in India, though I've heard that in at least one region there it's applied with shared MV/LV earth electrode (certainly not in the AU/NZ prescription!) which has made it far from a nice system.

Brazil also permit what we would call a shared HV and LV earth on their SWER fed transformers.

It makes you wonder about lighting induced surges as well, as then the whole HV line is acting as the top wire antenna, and presumably brings a large part of the lightning pick up onto the LV side.

If earth current is expected as part of a legitimate path, then you cannot tell a fault from a real load without rather a lot more effort, involving DC bias and capacitors at the transformers, which in Brazil at least, is not done.

  Compare to our HV system, the 3 phases are earth free, so we can have an earth fault detector, rather in the manner of a big boy's RCD, and the power can be off more or less as the HV line hits the ground.

Brazil:  I didn't know that - thanks .. will ask someone I know from there. Perhaps they use TT for the LV side, or just have very reliable soil.  LV lines can be be hundreds of metres in some places, so even by themselves the lightning pickup isn't to be ignored. 

Yes, systems where earth-currents are expected in normal operation remove some important possibilities for fault detection. At medium-voltage (or high-voltage as UK DNOs and IEC would say - excuse me that I've become used to this 'MV' description) it's not just SWER systems with this problem, but also the system used in most of the US states and some places beyond, which is basically 'MEN' on the 13.8 kV system along with bonding to the LV neutrals. Then some people complain about things like induced currents in pipelines, partly due to currents in 13.8 kV lines not balancing. And also about corroded neutrals on 13.8 kV cables causing shock danger. Its main benefits, as at LV, are that simple overcurrent protection can provide selective protection of the many branches and transformers that are needed when using a lower low-voltage distribution in spite of heavy loads and big spaces between homes.  In contrast, in Sweden most MV lines (unless metal cased) have to detect and promptly disconnect for an earth fault below about 5 kohm, i.e. about 1 A (they're usually resonant-earthed, so noticing a fault isn't that hard, even if locating is).


Following from some earlier points, I agree that regardless of the continued use of 'PME' or a separate protective conductor, monitoring should (and eventually will) start to get some serious use.  There's no real technical barrier to it even now, and there are already smart meters in some countries that report strange voltages that could indicate a neutral problem. Much more can be done, including the use of multiple measurements together. But various issues of tradition, poor communication channels, no low-price products being pushed hard, not huge gains (serious accidents seldom, serious damages awarded even more seldom) limit the speed of uptake.

Be aware that Brazil more than most only has a nodding relationship with its own regulations, in the sense of not actually following them, and the regulations themselves are  a bit incomplete.

Despite rules to the contrary, much LV wiring is still first fault to danger. Any country that can standardise on the same type of  plug for 2 voltages and then not use it anyway  is a bit different anyway. I found it quite illuminating to see things like twist and tape joints in a shower and the attitude to earthing (as in who needs it).


More generally, and back to the UK,  I agree LV network fault detection could be a lot better if it was automated.


The HV side is becoming quite well networked already, with telemetry from new switches and transformers, automatic earth fault relays on overhead supplies,  but some of the LV side in the UK is still state of the ark. 


On the HV (either 33kV or 11kv phase to phase)  we do not have a neutral or a ground, so any earth leakage is a fault, so the earth can be an electrode and then we keep LV and HV separate if the earthing is not very low resistance, as the HV earth can see quite a rise of potential during an HV to ground fault.

(Used to be a simple one ohm limit for deciding combined or separate, now it is a risk assessment and a calculation, that still usually means that much above an ohm, HV and LV earth separates, or combined if much below)

"Ufer (concrete-encased electrode): it's a person's name, not an acronym (cf. U.F.E.R. in an earlier comment)."


So sorry Herbert G.

https://en.wikipedia.org/wiki/Ufer_ground


But oh dear......A disadvantage of Ufer grounds is that the moisture in the concrete can flash into steam during a lightning strike or similar high energy fault condition. This can crack the surrounding concrete and damage the building foundation.^[8]


Z.

After looking at my particular earthling arrangement I thought I would take a risk based evaluation walk through of the application of BS7671:2018 to it.  The 11kV 400 volt substation is about 150 metres away. It is cabled in the street by repurposed 1930 era DC cabling. Thus we have two line phases and a smaller neutral in PILC.  We have concentric single phase cable jointed to it for 4 metres to the 60 Amp Henley service cut-out.  If we measure the Line to Neutral Z and the Line to Earth Z we get 0.19 ohms in both cases with prospective fault currents of 1324 amps.  I think we can assume we have TN-C-S.  No label to designate PME.


Looking at the equipotential bonding zone we have steel gas service pipe and plastic incoming water pipe both becoming copper internally.  But critically both the gas and water conductive pipes become external and exposed, as does the metallic flue of the boiler, thus exporting the PME to the outside. The external exposed extraneous parts are intrinsic to the design of the building.


One of the consequences of ESQCR has caused the DNO to state that the risk of a broken PEN conductor is significant and PME must not be exported.  This is then mandated in BS7671.


If we risk asses TN-C-S it has some pluses and some minus features.

 Firstly the prospective fault current Line to Neutral and Line to Earth being high means that the over-current BS88 service fuse will protect the tails and bus bar of the all metal consumer unit. Also all the sub circuit over-current protective devices will clear line to earth faults within the allowed time without the need for additional protection, (even though it is mandated for other reasons.


A risk that is not widely known about by the general public is that if type 1 appliances taken outside the house, it then exports the PME.  We have the risk that a broken PEN conductor then allows the Neutral to true earth potential to rise to dangerous levels with no means of detecting of clearing the fault.  The risk is suggested to be around one occurrence per day with little evidence of fatalities or prosecutions yet. Also some concerns about the Neutral to true earth potential being high enough to cause it to be noticed under adverse conditions.  For example bare feet on wet ground.


Should the public be concerned or is this a tolerably small risk?


Then contrast with the TT system that is mandated to be used if the TN-C-S is not acceptable. We have Ze anything up to 200 ohms.  No protection on tails or Bus bar of the consumer unit. None of the sub circuits can clear a line to earth fault using over-current protection. Thus we have to have RCD either in sections or per circuit in the form of RCBO to clear line to earth faults.

We must ask how reliable is an RCD? Some manufacturer evidence suggests that failure rates are around 3 – 10 % after 10 years.  The smaller figure being obtained if the device is regularly tested.  Therefore to maintain safe operation we have to have the system regularly inspected and tested to find and clear the inevitable fails in the RCDs. How often does a line to earth fault occur on a type 1 appliance?  And how often is this coincidental giving rise to the hazard?  Whilst some homes are regularly inspected and tested I would guess the majority are not. So should we risk asses using what the regulation says should happen or risk asses based on what is actually happening in practice. If you want to use the regulation based approach then maybe it should be compulsory to have all installations routinely inspected and it be enforced.  Good luck with that approach come the next election.


The obvious choice is TN-S.  It does not have the drawbacks of the other systems.  But that is not a choice for the customer. I do think that BS 7671 should more forcibly go in this direction and eliminate the risks of the other systems. In a way BS 7671 is being caught between conflicting requirements of interested parties. There is only one obviously safe system, the others suffering significant but currently tolerable risks. Adding ever more complicated protection relays or convoluted constraining regulations is the wrong direction to go.


So going back to my assessment I think on balance that using the exported TN-C-S is a marginally less risk than going to TT despite the regulations saying otherwise.

This is then mandated in BS7671.

Which bit of BS 7671 do you have in mind?

 

We must ask how reliable is an RCD? Some manufacturer evidence suggests that failure rates are around 3 – 10 % after 10 years. The smaller figure being obtained if the device is regularly tested. Therefore to maintain safe operation we have to have the system regularly inspected and tested to find and clear the inevitable fails in the RCDs. How often does a line to earth fault occur on a type 1 appliance? And how often is this coincidental giving rise to the hazard? Whilst some homes are regularly inspected and tested I would guess the majority are not.

About 7% failure rate has been quoted elsewhere - so your numbers sound reasonable - although failure in this context can mean things other than just failing to trip - e.g. tripping slightly more slowly or at a slightly higher threshold than permitted - which might not be enough to make the difference between life an death in the end.


A common approach is to have two or more tiers of RCDs - e.g. a 100mA S-type incomer and 30mA individual circuit protection - thus if the 30mA fails the 100mA unit would still trip for ADS. Similarly caravan setups duplicate the 30mA RCDs - having them both on the pitch supply and as a caravan incomer. That sort of approach rapidly reduces the risk - statistics might suggest 7% is then reduced to 0.49% - although in practice the two RCDs being subject to similar conditions simultaneously might make it not quite as good as that.

 

The obvious choice is TN-S. It does not have the drawbacks of the other systems.

TN-S isn't entirely ideal. It still has the problem that any L-PE fault will raise the potential on the earthing system well above safe touch levels until the fault is cleared - which can sometimes take considerably longer than the 0.4s generally regarded as a safe maximum. Faults on larger or distribution circuits within an installation can take up to 5s to clear and faults on the DNO network even longer - and a raised voltage is then impose on just about everything in the system - including Class I appliances outdoors.


TN-S with an enlarged PE conductor so that the voltage about true earth is limited (say to 50V) might be an improvement - or indeed an arrangement with a neutral earthing resistor which can keep differences in Earth potentials even lower. If we were to go to the expense of scrapping PME, perhaps we could do better than TN-S?


   - Andy.