Constant High Mains Voltages At Home

[Alti]

Don't you think that a failure of single RCD could put whole neighbourhood in danger with this 1 ?

I don't. How I see it working is that the supplier provides each customer a ~10 A or so rated, simple, robust, reliable differential current trip in the same assembly as they currently provide a fuse, and a second backup unit at the substation. (..) Of course it'll never happen, too much installed infrastructure exists, and nor do I claim to have "invented" it.

You are aware that current IEC regulations prohibit use of any switchgear on PE? I mean, zero. No switches, disconnectors, isolators, fuses. You would have to install sth that disconnects phases and neutral only (so you need 4-pole RCD). So the failure of this one RCD would have tripped whole area, finding nuisance tripping of this fault in a city would have been an experience very close to horror story.

[richard.cs]

You are aware that current IEC regulations prohibit use of any switchgear on PE? I mean, zero. No switches, disconnectors, isolators, fuses. You would have to install sth that disconnects phases and neutral only (so you need 4-pole RCD). So the failure of this one RCD would have tripped whole area, finding nuisance tripping of this fault in a city would have been an experience very close to horror story.

I wasn't expecting to disconnect PE (though note it is now common for EV chargers to disconnect PE in the event of lost TNC-S neutral, yey for ugly workarounds).

In the current system the failure-to-open of a single fuse at the customer* results in a the same sized area going off supply when the fuse at the distribution substation consequently opens. Yes, fuses are more reliable than RCDs, but that in part is because most RCDs contain contain finely balanced mechanics to allow tripping from just 30 mA and this makes them susceptible to friction, corrosion, etc. A 10 A RCD should have a reliability closer to a 10 A MCB than a 30 mA RCD.

*This happens fairly often, most commonly because the customer damages the cable prior to that fuse but also due to faulty, damaged or hopelessly-obsolete (not capable of breaking modern fault current levels) fusing arrangements at the customer incomer. A distribution substation fuse blows, 30-40 customers loose supply for a couple of hours.

[bsdphk]

[...]fusing arrangements at the customer incomer. [...]

I'm trying to understand what you're saying, and failing.

Can you tell me how UK differs from how we do it here in DK ?

At the 10kV/400V trafo, each distribution cable (4x400mm² typically) is fused with Nx100A.  Those almost only blow when a backhoe does something stupid.   The neutral is grounded at the trafo.  No PEN in the distribution cables.

Each installation is fused at the point where it connects to the distribution cable, typically in a small road-side cabinet or at the top of a mast.

For a house 3x25A is normal.  They are in peril when preparing for X-mas evening, but usually make it through.

From there it goes underground, or in a few straggling cases via overhead wire, (typically 4x4mm², but 4x16mm² for longer runs) to the house, through the meter, which is mounted on the outside of the building these days, then to the distribution panel where, it hits the HPFI first.   Each building has its own PEN electrode.

[Trader]

https://www.quora.com/Where-can-I-find-and-size-a-whole-house-voltage-stabilizer-We-live-in-a-developing-country-where-voltage-fluctuations-are-frequent-and-extreme-It-is-not-unusual-for-a-phase-to-disappear-altogether-220V-1-50Hz

[richard.cs]

[...]fusing arrangements at the customer incomer. [...]

I'm trying to understand what you're saying, and failing.

Can you tell me how UK differs from how we do it here in DK ?

Basically that UK houses typically have high-current single phase supplies. This is not really a very good idea, but the vast majority of domestic supplies, old and new, are single phase.

A typical underground distribution cable might be 240 mm², 3 core aluminium plus concentric PEN usually copper. Line conductors are fused at ~300 A. Each customer gets a single-phase feed which typically branches off the distribution cable somewhere under the street, this might be around 25 mm² and is usually concentric as well. This cable runs through the customer property, often under the driveway or front garden, and ends in a "service head" which looks something like this: https://talk.electricianforum.co.uk/uploads/monthly_2016_04/570411866dc11_Servicecablehead.png.194acc5547d6d78e3f6464db9ab1b692.png. The service head typically contains a 60, 63, 80 or 100 A cartridge fuse in L only. This fuse is to limit let-through energy into the customer installation, as well as to protect the branch cable and the meter, etc. If TNC-S earthing is used this is where PE and N separate. Typically 25 mm² conductors go from the service head to the meter, and then to the customer's equipment. There are variants on this, TNS and TT earthing are both fairly common, older cables may be paper insulated lead covered, old service heads containing open-wire fuses still exist but are now rare. Often the supply looks like (and originally was) TNS with a split-concentric cable or a 2 core plus sheath PILC cable but in fact some section of the street cable combines N and E making it non-obvious TNC-S. 

A typical L-E fault current in this setup is 2-10 kA. Usually a L-E fault in the customer installation trips MCBs in their consumer unit. More rarely such a fault, or a long duration overload, will open the 60-100 A fuse in the service head. More rarely still a fault in the consumer installation can open a substation fuse; if the substation fuse is already running hot due to heavy load, the service head fuse has been replaced with e.g. a bolt by a customer* who blew it cooking Christmas dinner, the service head has some ancient open-wire fuse that can't break the arc, or the customer simply damages the cable on their property prior to the service fuse (spade, nail, etc.).

*illegal, but it happens.

[Gyro]

[...]fusing arrangements at the customer incomer. [...]

I'm trying to understand what you're saying, and failing.

Can you tell me how UK differs from how we do it here in DK ?

At the 10kV/400V trafo, each distribution cable (4x400mm² typically) is fused with Nx100A.  Those almost only blow when a backhoe does something stupid.   The neutral is grounded at the trafo.  No PEN in the distribution cables.

Each installation is fused at the point where it connects to the distribution cable, typically in a small road-side cabinet or at the top of a mast.

For a house 3x25A is normal.  They are in peril when preparing for X-mas evening, but usually make it through.

From there it goes underground, or in a few straggling cases via overhead wire, (typically 4x4mm², but 4x16mm² for longer runs) to the house, through the meter, which is mounted on the outside of the building these days, then to the distribution panel where, it hits the HPFI first.   Each building has its own PEN electrode.

A more up to date and hopefully more representative / clearer TNC-S installation (situated in integral garage). Although the service head and incoming supply cable date back to the early 1980s.

The 3-phase street cable is paper insulated, so presumably the individual house branches are too. Each house is connected to one phase with PEN to the outer sheath. These  connections are buried underground and inaccessible. If you look at the cable boot on the feed cable entry into the service head, you can see that the cable is coaxial with the PEN conductor forming the outer sheath. You can also see a Yellow disc on the back board, signifying that the house is on the Yellow phase of the street cable (old colour coding).

As richard.cs says, the service head contains the electricity company cartridge fuse of up to 100A. A PFC of up to 10kA sounds too high for a domestic situation though, as it would exceed the 6kA maximum breaking capacity of the main switch and breakers (MCB / RCD / RCBO) in a normal domestic consumer unit.

Current standards require the 25mm² supply tails from the service head through the meter and to the consumer unit (everything after the meter is the customer's responsibility), with 16mm² Protective ground conductor, connected to the PEN conductor at the service head.

The optional additional customer-side isolator is fitted on request (normally free if you remember to ask at the time of a meter change) and allows safe isolation procedure for an Electrician working on, or replacing the Consumer unit without the energy supplier call-out to remove and replace / re-seal the service head fuse. If you look at the cables coming down into the Consumer unit, you can see a 10mm² earth bonding conductor (in this case daisy-chained) for the incoming Gas and Water service pipes.

[richard.cs]

A more up to date and hopefully more representative / clearer TNC-S installation (situated in integral garage). Although the service head and incoming supply cable date back to the early 1980s.

Yes that's a better photo 

The 3-phase street cable is paper insulated, so presumably the individual house branches are too. Each house is connected to one phase with PEN to the outer sheath. These  connections are buried underground and inaccessible. If you look at the cable boot on the feed cable entry into the service head, you can see that the cable is coaxial with the PEN conductor forming the outer sheath. You can also see a Yellow disc on the back board, signifying that the house is on the Yellow phase of the street cable (old colour coding).

House branches installed the same time as the lead street main would have been lead but your cable looks like PVC concentric. I don't think I have ever seen a boot like that on a lead cable. Note also the "Protective Multiple Earthed" label bottom right indicating that it's TNC-S.

As richard.cs says, the service head contains the electricity company cartridge fuse of up to 100A. A PFC of up to 10kA sounds too high for a domestic situation though, as it would exceed the 6kA maximum breaking capacity of the main switch and breakers (MCB / RCD / RCBO) in a normal domestic consumer unit.

You are allowed PFC > consumer unit rating because the cartridge fuse is considered to be energy-limiting. You are correct that 10 kA would be uncommon, unless you are in central London or are right next to the transformer. Really rural areas can have a PFC of just a few hundred amps, these can have problems with fault clearance times and voltage drop on load (and should never be TNC-S).

[Gyro]

Yes, the cable does have a PVC outer covering - as does our street cable, which shorts and fails with monotonous regularity. The house builder apparently over-flexed it whilst laying it in clay soil. When water gets in, the cable paper insulation breaks down (one phase to PEN) and takes out the substation fuse. Strangely these don't tend to be hard faults, they normally have to replace the fuse several times, or fit an automatic re-closer, until the cable develops a permanent short. Our pavements are more patch than original, with tell-tale 'sniff test' holes drilled at regular intervals. I wish I had kept the cut out section from the spoil heap when it failed outside our house one Christmas Eve. It had a beautiful solidified Aluminium 'stalactite' running down the side. Unfortunately it kept oozing pitch, so tossed it back.

[f4eru]

Lots of photos covering a range of construction types and voltages here:
https://www.spenergynetworks.co.uk/userfiles/file/scottishpower_cables_equipment_metal_theft.pdf

is it me, or is this document a "guide for thieves ?"

Basically that UK houses typically have high-current single phase supplies. This is not really a very good idea, but the vast majority of domestic supplies, old and new, are single phase.

Yep. Same scheme here in France. It simplifies wiring you home, no need to balance your home, but you need a beefier incoming cable, and you are stuck with single phase.

[bsdphk]

Basically that UK houses typically have high-current single phase supplies. This is not really a very good idea, but the vast majority of domestic supplies, old and new, are single phase.

Ohh, wow...

Now I understand why the UK plugs have built in fuses!

The detail which makes my hair stand on end is that the customers feed cable is only fused by the substation fuse.

This picture shows the road-side cabinet where my house is connected, they had to upgrade it to find space.

The main distribution cable in+out 4x150mm² go on the unfused blue clamps.

Usually the distribution is run in a sort of two-string zig-zag pattern along a residental road, this allows them to isolate one or two of these cabinets, while keeping power on the rest of the road.  The unconnected cable on the right is one such fused "diagonal" connection, probably 4x100mm².

The white plugins hold three fuses or six fuses, and supply one or two cables.

The "drop" cables to the nearby installations are 4x6mm²

Because our house is 200m from this cabinet, they ran a 4x75mm² to another cabinet about halfway, that holds our 25A "mast-fuse" and the cable from there is 4x16mm².

One benefit of this system, is that the fire-brigade can pull the fuses when a house is on fire.

[Monkeh]

Basically that UK houses typically have high-current single phase supplies. This is not really a very good idea, but the vast majority of domestic supplies, old and new, are single phase.

Ohh, wow...

Now I understand why the UK plugs have built in fuses!

That.. has absolutely nothing to do with it being a single phase supply.

The detail which makes my hair stand on end is that the customers feed cable is only fused by the substation fuse.

It's also protected by the fuse in the house. Mechanical damage is unlikely due to depth of burial.

[Gyro]

Basically that UK houses typically have high-current single phase supplies. This is not really a very good idea, but the vast majority of domestic supplies, old and new, are single phase.

Ohh, wow...

...

This picture shows the road-side cabinet where my house is connected, they had to upgrade it to find space.

The main distribution cable in+out 4x150mm² go on the unfused blue clamps.

Usually the distribution is run in a sort of two-string zig-zag pattern along a residental road, this allows them to isolate one or two of these cabinets, while keeping power on the rest of the road.  The unconnected cable on the right is one such fused "diagonal" connection, probably 4x100mm².

The white plugins hold three fuses or six fuses, and supply one or two cables.

The "drop" cables to the nearby installations are 4x6mm²

Because our house is 200m from this cabinet, they ran a 4x75mm² to another cabinet about halfway, that holds our 25A "mast-fuse" and the cable from there is 4x16mm².

One benefit of this system, is that the fire-brigade can pull the fuses when a house is on fire.

That's a neat system - assuming the cabinets are physically secure against vandals, vehicle strikes, Copper thieves etc!

The fire-brigade access is very useful.

Now I understand why the UK plugs have built in fuses!

As Monkeh says, irrelevant, too far down stream of the house breakers (it's there to protect the appliance cable), but I suspect that it wasn't a serious comment.

The detail which makes my hair stand on end is that the customers feed cable is only fused by the substation fuse.

Yes, I agree with you on that one, it is a concern. The house feed cable, of smaller conductor area than the street cable isn't appropriately fused. The cable, whilst buried for most of its length, in older properties does come up through the house foundations and is exposed to some extent before entering the fused service head. In these properties, this entry is often under the stairs - a primary escape route.

Taking your 'plug fuse' comment as an example, there would, in all other circumstances, under the IEE wiring regs, be a protective device of appropriate rating to protect the house feed cable.

Electricity distribution companies are unfortunately a rule unto themselves. Whilst current edition wiring regs require things like non-combustible consumer units to reduce fire risk, they are not required to follow them (they have their own codes of practice) and continue to install plastic cased service heads, isolators and meters new-builds (and upgrades - as my previous photo). They are, however, now brought up into a plastic cabinet built into the outer wall of the building. This houses the service head, meter and hopefully an isolator switch, with the connection to the customer's consumer unit inside the property via 25mm² double insulated tails. This should mean that any supply related fire should occur only on the outside of the building.

I don't believe that the supplier equipment is a significant cause of domestic fires (as badly installed (loose connections) older plastic consumer units can be. I think this is probably due to careful installation, but the small risk is there.

[Monkeh]

I don't believe that the supplier equipment is a significant cause of domestic fires (as badly installed (loose connections) older plastic consumer units can be. I think this is probably due to careful installation, but the small risk is there.

Meter installers yanking on the tails is a significant cause of fires in otherwise perfectly good CUs, for which they take no responsibility whatsoever..

[Gyro]

Meter installers yanking on the tails is a significant cause of fires in otherwise perfectly good CUs, for which they take no responsibility whatsoever..

Yes, I've heard of that one too!

Thick stranded cables with few strands suffer from  'strand shuffle' when flexed, which can loosen clamping screws. A standard 25mm² tail has 7 strands, while 'flexi-tails' have 25 strands.


EDIT: I know from recent experience that they don't open the consumer unit to check the clamping screws afterwards - a case of not their property, not their problem!

[bsdphk]

Now I understand why the UK plugs have built in fuses!

That.. has absolutely nothing to do with it being a single phase supply.

It was a comment on the "architecture" of the overall solution :-)

As for the road-side cabinets here in Denmark:

They seem remarkably resilient, everything in them is well insulated by plastic, and even when hit by a car, they mostly seem to keep working as intended. 

They seem to have a "frangible zone", they seem bend over around ground level with the actual box being pretty intact.

After a hit they get inspected, and if anything could possibly be exposed, they are wrapped in a solid orange plastic bag with warning signs.

Proper fixing happens some weeks or months later.

There isn't much of an opportunity for stealing copper in these cabinets, the cables go straight into the ground and the fat ones are usually alu.

(Denmark have no native metal ores, all metal is imported, this gave norwegian alu an early leg up in our cabling.)

[Siwastaja]

Large single-phase house supply is also common in the USA.

Three-phase power only has point in distribution, and running large motors off the line directly.

In homes, 3-phase power itself has had very little use and sees less and less use due to availability of variable frequency drives.  From the household perspective, 3-phase power is like three separate power lines, fused separately. The problem is the small fuse size. Here, 3x25A is the most typical household fuse size. Assuming I never want to drive classical off-the-line 3-phase motor loads, I would take 1x75A over it any day. Same nominal power, but a lot more actually usable power.

25A is such small current that 3x25A requires careful balancing acts for many households and even then it's quite common to accidentally blow the main 25A fuses.

Of course, from power company viewpoint, this is superb because the difficult task of balancing is forced down the throats of the households. Because 25A is so little, everyone has to do some balancing and large resistive loads have to be designed to load all phases equally. So now a home with a 10kW heater is in perfect balance, unlike a system where single-phase supply is used and one house has a 10kW heater and the neighbor does not.

[bsdphk]

25A is such small current that 3x25A requires careful balancing acts for many households and even then it's quite common to accidentally blow the main 25A fuses.

I think it is important for the discussion that we all understand how much this is "per-country" due to political and technological history.

Even between the nordic countries there are huge differences, for instance in EV uptake or how much resitive heating is used, it has been almost outlawed in DK since 1973, whereas I belive it is still pretty much the norm in NO and FI (not sure about SE).

I was surprised to discover that my swedish ground-circuit heat-pump came with three current transformers to be installed at the infeed, so that the compressor could avoid overloading the 3x25A.  Here in DK it seems to be regular practice to just put them on the heat-pumps own cable, but I insisted them mount them in the infeed, but so far they have never had any effect.

But their very existence support your argument, that this is a concern other places.

Having worked in computing reliability for decades, I'll take three separate fuses over one fuse any time, and speaking from experience, designing the installation in a new house to distribute the load over three phases is really not an issue  According to my monitoring, we peak out at 20A.

But yes, adding an electrical car would require 3x35A, but that is just a one-time fee here in DK, so I dont consider it a problem?

[gnuarm]

But yes, adding an electrical car would require 3x35A, but that is just a one-time fee here in DK, so I dont consider it a problem?

Just so I understand, you are saying that a typical home only has 25 amp service on a three phase circuit?  Also, for some reason, EV charging it typically done with a 35 amp, three phase circuit? 

I don't know if your country requires derating the current on a circuit like we do here in the US for continuous loads.  If not, that gives 15 kW at 240V.  That's a lot of charging rate and seems not necessary.  It will charge my car from 0 to 100% in under 7 hours.  A real world need would be much less.  It could easily charge overnight on a 20 amp circuit, which is 8 kW.

I know in the UK they use 9 amp tea kettles.  It would seem a 25 amp circuit for a whole house means limited usage of electricity for heating, either the home or water.  My hot water heater is on a 30 amp, 240 volt circuit actually drawing 18 amps and over 4 kW.  Instant on hot water heating must draw more current since my heater won't keep up with taking a shower.

[bsdphk]

But yes, adding an electrical car would require 3x35A, but that is just a one-time fee here in DK, so I dont consider it a problem?

Just so I understand, you are saying that a typical home only has 25 amp service on a three phase circuit?  Also, for some reason, EV charging it typically done with a 35 amp, three phase circuit?

Yes, the default "mast fuse" here in Denmark is 3x25A.

If I add EV charging on top of our current load pattern, it would require the "mast fuses" to be upgraded from 3x25A to 3x35A.

That is literally all there is to it: Change the fuses, pay the "investment fee" (as you now have bigger access to the grid) and you're done.

[bsdphk]

I know in the UK they use 9 amp tea kettles.  It would seem a 25 amp circuit for a whole house means limited usage of electricity for heating, either the home or water.  My hot water heater is on a 30 amp, 240 volt circuit actually drawing 18 amps and over 4 kW.  Instant on hot water heating must draw more current since my heater won't keep up with taking a shower.

Two things:  The fuses we are talking about, are slow fuses.  It will take 25A sustained forever, and brewing a cup of tea, even if it puts you at 34A for a few minutes, will not blow the fuse.

Your 240V/18A hot water heater would claim 18/(3*25) = 24% of the available continuous capacity, but it runs only every so often and the chances that it, the oven, the dishwasher, the washer, the dryer and the hair-dryer all being on a the same time is practically non-existent, and even then, it would only be for a few minutes.

The trouble is when you add a huge "base-load" like EV-charging:  That constant hour-long load steals the head-room, and means that the probability of intermittent loads may push the fuse to blow increases a lot.

[Ian.M]

I know in the UK they use 9 amp tea kettles.  It would seem a 25 amp circuit for a whole house means limited usage of electricity for heating, either the home or water.  My hot water heater is on a 30 amp, 240 volt circuit actually drawing 18 amps and over 4 kW.  Instant on hot water heating must draw more current since my heater won't keep up with taking a shower.

Err...  UK kettles can be up to 3KW, which is 12.5A @240V. 

If they use an IEC 60320 C15/C16 hot condition connector pair, then they should be under 10A due to the connector's limitations, but its increasingly rare to find such kettles.  Really old kettles used a round 13A connector - see: https://www.amazon.co.uk/Spares2go-Universal-Element-Connector-Grommet/dp/B00TFTKXNQ

Many older homes in the UK were on 60A single phase feeds, and didn't have enough current available for permanently installed high wattage electric heating/cooking appliances.  I would expect that such installations are now vanishingly rare, as in most cases they would have been upgraded to 100A feeds when the fuse board was replaced with a modern consumer unit.

[fcb]

I know in the UK they use 9 amp tea kettles.  It would seem a 25 amp circuit for a whole house means limited usage of electricity for heating, either the home or water.  My hot water heater is on a 30 amp, 240 volt circuit actually drawing 18 amps and over 4 kW.  Instant on hot water heating must draw more current since my heater won't keep up with taking a shower.

Two things:  The fuses we are talking about, are slow fuses.  It will take 25A sustained forever, and brewing a cup of tea, even if it puts you at 34A for a few minutes, will not blow the fuse.

Your 240V/18A hot water heater would claim 18/(3*25) = 24% of the available continuous capacity, but it runs only every so often and the chances that it, the oven, the dishwasher, the washer, the dryer and the hair-dryer all being on a the same time is practically non-existent, and even then, it would only be for a few minutes.

The trouble is when you add a huge "base-load" like EV-charging:  That constant hour-long load steals the head-room, and means that the probability of intermittent loads may push the fuse to blow increases a lot.

A number of EV chargers have options for incomer current transformers, this way they can back-down the charge current to limit the peak load. I haven't seen this done often in domestic UK installs.

[bsdphk]

A number of EV chargers have options for incomer current transformers, this way they can back-down the charge current to limit the peak load. I haven't seen this done often in domestic UK installs.

Yeah, same as my heat-pump.

But do you want to bet that if both the EV charger and the heat-pump monitor the infeed current, they're going to confuse each other mightily ?

[Siwastaja]

I think it is important for the discussion that we all understand how much this is "per-country" due to political and technological history.

Even between the nordic countries there are huge differences, for instance in EV uptake or how much resitive heating is used, it has been almost outlawed in DK since 1973, whereas I belive it is still pretty much the norm in NO and FI (not sure about SE).
...
But yes, adding an electrical car would require 3x35A, but that is just a one-time fee here in DK, so I dont consider it a problem?

Here, 3x35A is significantly higher monthly fee than 3x25A so most try to live with 3x25A (which usually is doable) unless they really need 3x35A. For example, here, 19.67 €/month vs. 33.96 €/month. It seems like waste if you only need it rarely, or can handle the situation by some careful balancing act instead.

And of course the 25A fuse can take 35A for quite some time (a few minutes 100% guaranteed; likely tens of minutes; possibly hours).

In any case, our eco-government is now paying a 2500€ subsidy if we dismantle working oil heating system and replace with direct electric heating; or 4000€ if replaced by air-to-water heat pump, which is what I did (haven't got my 4000€ yet, though). In any case, every air-to-water heat pump* revert to being simple direct electric heaters when the temperature drops to below around -20 to -25degC which is commonplace enough. This obviously coincides with when heating power is required the most; so now your 2.5kW input, 9kW output investment is temporarily just an expensive decoration which inputs 9kW to output 9kW. The grid here mostly can take it no problem, but production is a big question mark, because for a some years now, Finland has been buying electrical energy from all neighbors. Gone are the days of buying from Russia and selling to Sweden. This winter, Sweden had close calls with their own supply and demand but thankfully were still able to sell us.

*) except a very few very expensive models, but even those won't have COP much over say 1.5 at such extreme temperature differential


The change is big, most homes built before 1990's use oil heating. Keeping existing, working oil-based burners as support devices for the few coldest days, still only contributing a few % to the total CO2 emission, would totally make sense. Most of the population live in areas where there are typically less than 5-10 days per winter of such low temperatures that air-source heat pumps have to rest (or work at ridiculously low COP say <1.5), but you need to design the whole infrastructure to be able to get through those days, and now suddenly we will have tens of thousands more households running with direct electric heating (so some 10kW). Which is 3x16A so you have 9A left per phase for everything else.

[ocset]

This is a great thread.
ie, what is causing the overvoltages experienced by OP?
I wish we could summarise all the salient points of this post, distill them into a few lines.

From my reading, it looks to me that there is almost definetely some kind of problem with the neutral connection. Its got damage such that it goes disconnected or high Z every now and then....intermittently. Although possibly it may be compromised all the time, and just waits till the  3 phase loads get unbalanced before giving the overvoltage.

There is possibly corrosion into the cable somewhere along the underground connection, and it would be very expensive for the power company to find out exactly where this point is.

I am just thinking....maybe what is needed is to chop the mains input to the house, at the house gates......and build a switch mode battery charger, with ability to handle up to 415VAC.....then put a big battery there, and make it (the mains input bit that got chopped) charge the battery....then put an AC generator at the battery, and make that give 240VAC to the house.

I suspect this is going to be the cheapest way round this problem....and make the power company pay for it.

Failing this, put up little pylons from the nearest substation to the house...and get the power company to pay for it....a fresh mains connection.

I wouldnt mind betting that the power company would rather pay OP for his house, then bulldoze it!