New Solar PV System - Connection to Consumer Unit

[Geflugenmitt]

Hi,

I am having a 4kW solar PV system professionally fitted. I have a degree in electrical engineering which is sometimes a curse. Whilst I don't have much knowledge of household electrical practices and regulations, I have enough theory to question what the installers are doing.

I asked today how the feed from the inverter is patched into the house system, and they told me it will use a spare fuseway in the distribution board. It was only after he'd left that the following hit me:

1. Does he mean by this that the live wire from the inverter would go into the 'top' of an MCB, thus going through the MCB in reverse to the live bus? Surely you can't draw current the wrong way through an MCB? They are not bidirectional like wire fuses.
2. This would also bypass the RCD protection for the bus because the incoming feed would be connected to the protected (output) side of the RCD. If so this seems dangerous to me if not illegal.
3. My consumer unit has a split bus. It is wired so the downstairs ring shares a bus with the upstairs lighting, and vice versa. This is so a trip caused by (eg) a faulty hairdryer won't extinguish the lights in the same room where the fault happened. But, similarly to my question about the MCB, whichever bus you connected the inverter to, current would have to flow backwards through the RCD on that bus (and then forwards through the RCD on the opposite bus) in order to deliver power to both.

Like I said, if I didn't have the first clue I would just sit back and let the installers do as they see fit, but I know enough to be concerned and I want to fully understand how this setup should be done. I appreciate any help cutting through the confusion.

 

[JohnW2]

Does he mean by this that the live wire from the inverter would go into the 'top' of an MCB, thus going through the MCB in reverse to the live bus?

My understanding is that that is how it is commonly/usually done.

Surely you can't draw current the wrong way through an MCB? They are not bidirectional like wire fuses.

Your degree in electrical engineering ought to lead you to a different conclusion There is nothing 'non-bidirectional' about an MCB. The magnetic tripping results from the magnetic effect of current going through a coil, and the thermal tripping by electrical heating - in neither case does it it matter which side of the coil or heating component is connected to the source - all that matters is what current is flowing through the coil or component.

Kind Regards, John

 

[Iggifer]

1. Yes, you backfeed into the CU via an MCB - this both supplies power to the inverter and allows the inverter to feed into the grid.
2. It does not bypass RCD protection - assuming, like it is in your case, protected by an RCD. An RCD monitors current flow between the two poles, and detects an imbalance between the two. It matters not in which direction this current is flowing.
3. The flow back into the grid will be via the RCD into the main switch and back into the grid. It can feed both RCDs - which is what you want because it would be somewhat useless to only benefit from your PV on half of the installation.

Some useful points
- If you lose power you also lose PV, the inverter needs a 230 supply from the grid, once this drops out the inverter stops converting DC to AC - both because some level of AC is required for the inverter to run and secondly because it could potentially be dangerous to those working on the reason for the power outage.
- I prefer to install them (we don't do them anymore, the FIT has somewhat killed the market) on either an MCB - assuming an allowable installation method, or an RCBO as inverters can become leaky over time and cause RCD trips.
- Ensure they install an AC and a DC isolator. Some less scrupulous people like to just use the MCB as the AC isolation, which works but it's rough.
- A DC isolator as close to the strings entering the property is ideal - if the inverter is not in the loft and the isolator is local to the inverter, you'll still have live DC cables from the roof to isolator even with the DC isolator off.

 

[flameport]

a. MCBs are not directional
b. it's AC, so the 'direction' of current is irrelevant.

 

[Risteard]

It shouldn't share an RCD with the rest of the installation as otherwise if the RCD disconnects with an Earth fault the PV will continue to supply the installation until it shuts down for anti-islanding protection once it detects the RCD trip. This will cause excessive disconnection times.

 

[Geflugenmitt]

OK do I feel stupid. AC - Non-directional - Doh! It's just that it felt conceptually wrong for a feed to go 'into' the wrong side of an MCB or RCD. It was quite late when I was posting.

I still however feel that my point 2 is valid. If the RDC trips, I am led to understand that the inverter will stop. BUT how fast does that happen? Won't there be a lag before the power from the PV system cuts out?

I *think* this issue is what Risteard is referring to, although I'm confused about the wording "...shouldn't share an RCD with the rest of the installation..." - should that not be "...shouldn't be connected to the circuit side of the RCD..."? He is saying this might cause excessive disconnection times, which is the same point I'm making.

I am also interested to know how the inverter senses that the RCD has tripped. It cannot sense the voltage drop on its output because it is still supplying power. It would self-sustain!

Thanks for all the replies, really helpful.

 

[Iggifer]

I am also interested to know how the inverter senses that the RCD has tripped. It cannot sense the voltage drop on its output because it is still supplying power. It would self-sustain!

It wouldn't self sustain because it requires a small amount of AC from the grid to run just like any other appliance. That's how they're still powered at night.

Once the AC drops out so does the inverter

 

[JohnW2]

It wouldn't self sustain because it requires a small amount of AC from the grid to run just like any other appliance. That's how they're still powered at night. ... Once the AC drops out so does the inverter

Indeed. I can but presume that there must be 'a little something' (maybe simply a very small resistor) between inverter and the grid supply - otherwise, as Geflugenmitt has said, it would be unable to sense loss of grid power (or frequency/phase of grid power) and therefore would be 'self-sustaining'.

Kind Regards, John

 

[John D v2.0]

Maybe it can sense the output impedance of the supply perhaps using a high frequency sub wave?

 

[Lectrician]

It shouldn't share an RCD with the rest of the installation as otherwise if the RCD disconnects with an Earth fault the PV will continue to supply the installation until it shuts down for anti-islanding protection once it detects the RCD trip. This will cause excessive disconnection times.

EXACTLY this. We’re seeing this done far too often.

 

[JohnW2]

Maybe it can sense the output impedance of the supply perhaps using a high frequency sub wave?

I have no idea, but I think that something along the lines of what I suggested is more likely - don't forget that the inverter has got to be able to not only sense the loss of the grid supply, but also to get a handle on its voltage and frequency/phase - things which would presumably be impossible (or next-to-impossible) if the output of the inverter were connected directly to the grid supply by a copper conductor 'of negligible impedance'.

Kind Regards, John

 

[John D v2.0]

Maybe, but I'm sure it would realise very quickly if the supply and grid due to the large imaginary power flowing

 

[DetlefSchmitz]

As I understand it, if the frequency or voltage fall outside a certain tight range it is assumed that the grid power is lost, and with the inverter supplying a load higher than it can (i.e. the whole neigbourhood) it very rapidly (< 5s) falls outside the required parameters and switches off.

 

[JohnW2]

As I understand it, if the frequency or voltage fall outside a certain tight range it is assumed that the grid power is lost, and with the inverter supplying a load higher than it can (i.e. the whole neigbourhood) it very rapidly (< 5s) falls outside the required parameters and switches off.

I'm sure that's true, but the question is how it senses the voltage, frequency and presence of the grid supply.

As I've said, if the output of the inverter were simply connected directly to the grid supply via a copper conductor of negligible impedance, it would merely 'sense itself' in the absence of a grid supply - so, as I've said, I think there has to be 'a little something' (probably just a very small impedance) between the inverter something and the grid supply.

Kind Regards, John

 

[SimonH2]

It wouldn't self sustain because it requires a small amount of AC from the grid to run just like any other appliance.

NO, this is not correct.
In principle, it could self-sustain powering it's internals from the AC it's generating - there is really no difference between running from AC that's going out and running from AC that's coming in on the same wires. However they explicitly have anti-islanding protection to force a shutdown if not connected to a "solid" grid.
The concern is that under the right (or wrong depending on how you look at it) conditions, embedded generation could maintain the supply to an island of the distribution system - causing danger to public and supply workers. Bear in mind that 4kW PV systems are not the only embedded generation - at my last place we had a customer with an 89kW hydro plant at the bottom of the farmyard. While the nature of a PV system means that it's unlikely to be able to maintain an island at anything like a stable voltage, other forms of generation (such as a diesel generator) might make a half-decent job of it. And lets not forget that there could be several embedded generators and a few motors (spinning inertia) connected to the island.


It's not enough to sense just voltage and frequency - they have to sense rate of change of these. Look up G59 for details.
On small generation (up to 4kW) you don't need permission to connect it PROVIDED that the equipment (inverter in the case of PV) is certified to the right standard - no DIY arrangements Over that, you need permission from the DNO and part of that will be them witnessing commissioning tests of the anti-islanding (and other) protections.