New Solar PV System - Connection to Consumer Unit

[John D v2.0]

Interesting! So what I want to know is how the anti islanding knows it's an island, if anyone knows for sure.
So far we've had a few ideas but nothing definitive

 

[SimonH2]

I believe my last post said that. Specifically it's a combination of voltage and ROC of voltage and frequency, I believe mainly ROC of frequency. But it's for the provider of the system to determine the best combination of factors to meet the requirements.
There may be other generators trying to control both the voltage and frequency, possibly to a fairly good degree - but they will not be able to compete with the frequency stability of the whole of the UK grid. Voltage alone, and even ROC of voltage, won't do it given how voltage can change rapidly with changes in loading.
You can have a read yourself, see section 10.3 starting on page 54 of of the Energy Networks Association document.

 

[ericmark]

Although AC travels in both directions with the incomer at bottom the contact (3) isolate most of the device, so that is the preferred method, as to RCD it states in quite a lot of information for example Wylex this must be double pole.

411.3.2.2
The maximum disconnection time shown in Table 41.1 applies to circuits not exceeding 32A. 0.4 s TN System & 0.2 s TT Systems.
551.4.1
Fault protection shall be provided for of each source of supply or combination of sources of supply.
(Also, refer to chapter 55, regulation group 551- Low Voltage Generating Sets)
551.5.1
Over current protection should be located as near as practical to the generator terminals (where required).
712.537.2.1.1
To allow maintenance of the PV Inverter, means of isolating the PV inverter from the DC side and the AC side shall be provided.
(Also, refer to chapter 53, regulation group 537- Isolation and Switching)
712.434.1
The PV supply cable (AC side) shall be protected against fault current by an overcurrent protective device installed at the connection to the AC mains.
Also, refer to Part 7, all regulations in Section 712 - Solar Photovoltaic (PV) Power Supply Systems)
712.411.3.2.1.1
The PV Supply cable (on the AC side) shall be connected to the supply side of the protective device for automatic disconnection of circuits supplying current - using equipment.
314.1
Every installation shall be divided into circuits as necessary to : (ii) facilitate safe inspection testing & maintenance (vi) prevent the indirect energising of a circuit intended to be isolated.
314.4
In an installation comprising more than one final circuit, each final circuit shall be connected to a separate way in a distribution board.
The wiring of each final circuit shall be electrically separate from that of every other final circuit, so as to prevent the indirect energising of a final circuit intended to be isolated.
522.6.101 522.6.102 & 522.6.103
Installers must consider the need for Additional Protection by 30mA RCD in accordance with regulations (Previously 522.6.6 & 522.6.8) that relate to concealed cables in walls
& partitions.

The Best practice guide refers to problem when RCD not double pole.

I looked for disconnection times, it seems the PV array has to disconnect if the voltage goes too high or too low or the frequency goes too high or too low, it is considered if the mains supply is lost either the voltage or frequency will go out of the limits and so fail, there are times quoted as to how fast it must close down if the limits are exceeded, however the problem is until those limits are exceeded there is not time to disconnect. So in real terms no one can state with a power failure how long it will take for the PV supply to exceed the limits, it would depend on load and how much sun etc. So it may be within 0.1 seconds when tested once, but will it be within 0.1 seconds with all tests if test infinity time?

Since no one can answer that question, the PV needs to supply the consumer unit with a double pole RCD. Here we are down to makes, some manufacturers make single width RBCO's with double pole switching, others need two modular spaces to fit such a device.

So if you want to cross every t and dot every i then likely you need a lot more than if you do a risk assessment and work out what is reasonable. Voltage tripping is a real problem with a PV install, we saw the voltage officially drop from 240 to 230 volt, but in real terms it stayed at 240 volt, until solar panels arrived that is. At 240 volt it was all too easy to trip out the panels due to over voltage, so we have seen areas drop to 230 volt simply to cater for solar panels.

I am quite sure a RCD tripping would quickly cause a solar panel grid tie inverter to quickly shut down, but with solar panels on half the houses in a street, how quickly a loss of supply would cause them to close down is a subject for debate. Each house will likely have slightly different limits set, so how long before they all trip, and if that time will be long enough for some panels to auto re-energise is a question that could be debated until the cows come home.

 

[JohnW2]

I believe my last post said that. Specifically it's a combination of voltage and ROC of voltage and frequency, I believe mainly ROC of frequency. ....

That sounds very probable. However, as I have said, there has to be some way of monitoring characteristics of the grid supply (voltage, frequency, ROC of one/both etc.) whilst the inverter output is connected - and that presumably requires 'something' (other than a conductor of negligible impedance) to be between the inverter output and the grid supply. Can anyone confirm that (or otherwise) and indicate the nature of that 'something'?

Kind Regards, John

 

[JohnW2]

.... we saw the voltage officially drop from 240 to 230 volt, but in real terms it stayed at 240 volt, until solar panels arrived that is. At 240 volt it was all too easy to trip out the panels due to over voltage, so we have seen areas drop to 230 volt simply to cater for solar panels.

I have PV installations all around me, but my supply voltage rarely drops appreciably below 245V (and virtually never below 240V), so I wonder how those installations fare?

Kind Regards, John

 

[magicmushroom666]

voltage and current transformers can be used to detect the rate of change for this protection. The protection relay will detect the change when the grid connection is removed and the full load gets transfered instantly to the PV inverter.

That sounds very probable. However, as I have said, there has to be some way of monitoring characteristics of the grid supply (voltage, frequency, ROC of one/both etc.) whilst the inverter output is connected - and that presumably requires 'something' (other than a conductor of negligible impedance) to be between the inverter output and the grid supply. Can anyone confirm that (or otherwise) and indicate the nature of that 'something'?

Kind Regards, John

 

[SimonH2]

However, as I have said, there has to be some way of monitoring characteristics of the grid supply (voltage, frequency, ROC of one/both etc.) whilst the inverter output is connected - and that presumably requires 'something' (other than a conductor of negligible impedance) to be between the inverter output and the grid supply.

Why ?
The inverter is solidly connected to the mains, and hence to the local distribution network, and hence to the grid. While there may be some voltage changes due to currents in the network - but there is no way the frequency at the power terminals of the inverter will be different to that of the grid.

 

[bernardgreen]

but there is no way the frequency at the power terminals of the inverter will be different to that of the grid.

It is not just frequency, the phase of the invertor has to match the phase of the grid as well.

 

[John D v2.0]

There's definitely some gap in my understanding that's escaping others. Maybe it's a bit basic!
So to pick an extremely naive example, imagine pair of DC alkaline batteries at 1.5v.
They are each powering a small load and have a cross connection to parallel them.
The batteries include anti islanding protection, so if the cross connection were lost, the anti islanding would shut down the battery to prevent it service the small load.
The issue is I don't see what would change at the broken link to alert the Anti islanding that it would disconnect.

Extend that to ac and you have two inverters and an ac load, does that add something that would help?

 

[Adam_151]

John,

Bit of an explanation here: https://pure.qub.ac.uk/portal/files/868297/03-2006-JA-PS-1-020[1].pdf I have to admit I can only take a away a fleeting overview of how it actually works from that, the maths is a bit beyond me, I never understood calculus of anything but the most basic functions and lost interest in A level maths at that point... and as for Fourier transofrmations, forget it.

On site all you need to know is that the magic black box that the panel builder has built into the panel alongside a big contactor does it all, it then gets commisioned by a chap who knows what the DNO want to characteristics to be and has a very expensive instrument to test them, but likewise, probably doesn't understand a great deal of what is going on either!

 

[SimonH2]

It is not just frequency, the phase of the invertor has to match the phase of the grid as well.

Which it will do. The inverter will sync itself when starting up and cannot do anything but remain in sync thereafter. So while connected to the grid, it's output frequency (and phase) will accurately track the grid - so no rapid changes in frequency, and hence why RoC of frequency can be used to detect islanding since a small island is highly unlikely to contain the sort and quantity of spinning machinery to keep the frequency anything like as stable as the grid.

... imagine pair of DC alkaline batteries at 1.5v.
...
The issue is I don't see what would change at the broken link to alert the Anti islanding that it would disconnect.

Assuming the loads are accurately matched to the batteries - nothing ! But that's DC - you can't use RoC of frequency on a DC circuit.

OK, lets take a bit of a step back - I assume many of us are familiar with using a portable generator ? And know how the engine tends to change speed with load ? So you've the genny running on light load and switch on the kettle - the genny tries to keep up, the load on the engine increases, the frequency and voltage drop considerably until the governor can get the engine back up to speed.

Now suppose we've fitted suitable G59 protection, and we sync the genny to the mains and parallel run - something that's quite common with larger standby generators as it allows you to sell it's capacity to the grid and get some payments, as well as being able to run it properly under load for periodic testing and turn over the fuel supply a bit. So it's running against the grid and exporting some power - and we switch on the kettle, what happens now ?
Because we're synced to the grid, the frequency isn't going to change - the grid is just so big relative to both our genny and the kettle that the change will have an undetectable effect. The voltage may change a little due to changes in currents in the network, but that'll be less than the sort of voltage changes that'll happen all the time.
But if the local network gets disconnected so our little genny is now powering a small island - that large and "rigid" grid is gone, so once again the frequency is unsteady as loads change, so the G59 relay can detect that and either shut down the genny or disconnect the incoming mains and allow the genny to keep powering the local site.

... it then gets commisioned by a chap who knows what the DNO want to characteristics to be and has a very expensive instrument to test them, but likewise, probably doesn't understand a great deal of what is going on either!

Yes, quite probably the chap won't understand the maths behind it, he'll just have a set of settings derived by someone who does (or obtained from a table created by someone who does).

 

[JohnD]

it is also highly improbable that the voltage would remain remotely stable.

An "island" might by chance have about the same generation and usage, but the moment someone turns on a kettle, or the sun goes behind a cloud, it will change. In my district the voltage varies between about 240 and about 245 as the sun goes in and out, and that's with the grid connected.

 

[SimonH2]

Bit of an explanation here: https://pure.qub.ac.uk/portal/files/868297/03-2006-JA-PS-1-020[1].pdf I have to admit I can only take a away a fleeting overview of how it actually works from that, the maths is a bit beyond me, ...

Well consider that it's written by a Ph.D. student and two lecturers on related matters and both having a Ph.D. But skimming through it, it wasn't as "opaque" as some papers I've come across

it is also highly improbable that the voltage would remain remotely stable.

An "island" might by chance have about the same generation and usage, but the moment someone turns on a kettle, or the sun goes behind a cloud, it will change. In my district the voltage varies between about 240 and about 245 as the sun goes in and out, and that's with the grid connected.

You've just demonstrated the problem with voltage or RoC of voltage - it varies wildly even when connected to the grid. I imagine that with something like PV inverters, they are programmed to output all the power they can get from the PV panels and if generation exceeds load then the voltage will rapidly increase to the upper trip level. But many generators may have AVRs (automatic voltage regulators) and if there's enough of those on the island then the voltage may remain within limits. Conversely, a PV inverter may have a very stable frequency, while a (eg) water turbine or engine driven genny won't have a stable frequency without a grid to stabilise it.
The problem is that the problem is undefinable in the general case simply because there are so many variables in terms of amount and type of embedded generation, size of the island, quantity and type of loads, etc etc. Hence all the research, and debates about "how much" embedded generation is safe before the system risks going unstable. This is alluded to in the research paper - if using (say) RoC of frequency, it's possible for a grid event (such as loss of a large power station) to cause some embedded generation to trip out, worsening the effects of the initial event, causing more to trip, and thus cascading as more and more trips offline.

 

[John D v2.0]

Excellent explanation, very easy to understand, thanks @SimonH2

 

[bernardgreen]

From years ago I recall generators being bought on line manually when the synchoscope indicated they were in phase with the supply already on the bus bars. O assume this is all done "automatically" with panels and invertors feeding back into the grid.

So much easier with DC supplies.