Re: Plug in solar panels - announced

The inverter follows mains frequency. When mains frequency goes away, the inverter stops. If there are two, one may go before the other but they are both following the mains. If there's nothing to follow they shut off within ms.

I doubt they're going to be rewiring the output stages.

If you want that, Proper Inverters (and batteries) have an emergency mains output that can run things while the mains is off.

You monitor the mains voltage, which is something the inverter has to do every few ms anyway. eg if you put a resistor in series with the live and measure the voltage waveform both sides of it, you can tell which direction current is flowing and whether it's in phase or if the mains voltage has gone away. To push power out they have to synthesise an inverter side voltage which is in-phase but higher than the mains-side voltage.

It is possible to spoof solar inverters by using an off-grid inverter (eg from a car battery) as a source of timing, but that's not relevant here. If you connected one of those to the mains I suspect it would either shut down or fry itself (its own timing reference being out of sync with the mains waveform).

Theo

That makes no sense, you can't "follow mains frequency". Yes, the inertia of the mains would be used to keep the inverter synchronised but there must be some sort of system within the inverter to keep its frequency around 50Hz. Thus, several inverters might have enough of their own inertia to 'see' the mains from themselves.

You underestimate people's daring (and stupidity)! :-)

But I surmised what happens if there's another inverter out there, providing the 'mains' voltage. That's where I came in.

The inverters don't have their own 50Hz timing generator. So either they see

They could just buy a different inverter that does that instead.

Theo

That's my plan. I have multiple locations which can take a few panels each, so it's better to have distributed microinverters rather than one central inverter.

I'll probably run dedicated circuits and/or FCUs rather than 'plug in', though. That also means I wouldn't be limited to 800W microinverters. Because of Germany/etc 'balcony solar' it's already expanded the microinverter market - eg before Enphase would only sell you a single 250W microinverter for £200, now you can buy a Hoymiles 4-MPPT-input 1800W for the same money.

I'll likely get a sparky in to do any notifiable work (eg add new circuits in the CU and check my work) but any sparky can do that, it doesn't need to be an MCS accredited firm. It's also only a small amount of work so won't cost much.

Theo

G98 defines RoCoF limits (rate of change of frequency) that inverters must monitor.

I bet they do!

The more usable route is a UPS whose battery is charged by solar for when the mains fails,

So you need in essence two inverters - one standalone and one that works when the mains is there

In fact the most usable solution is the solar panels plus a UPS operating independently

But as with all things 'renewable', the cost and complexity soon spirals out of control

but then you never hear of anything spiralling IN to control :-\

That is because controlled systems dont have spirals

Wouldn't they phase lock to the mains signal they see at the plug, then adjust their phasing to "push" current into the grid ?

On rotary prime movers, you can lead or lag the grid, and central control can tell you which is desired at the moment, to "regulate" the grid.

The inverter on the other hand, is not under central control. That means it has to have a policy "at the plug" as to what to do. Phase locking is essential to injecting AC into the grid in a controlled fashion. If the inverter and the grid are both running at exactly 50.03Hz, and as well, the sine waves exactly match (0 degree phase, plesiochronous 50.03Hz), then no power is transferred. If the inverter then adjusts its sine wave so it is slightly out of phase, this causes a net power transfer to the grid.

And if the grid has a step load applied to it, the frequency can change from 50.03 to 49.97Hz, and the inverter has to track and leak out the frequency change, so that the power being transferred does not change measurably. it has to track the frequency, in order to maintain the phasing in an intelligent way. The situation is plesiochronous, because the grid never stays at 50.03Hz for very long. By the time I type this, it could be 50.027 Hz, and unless I respond quickly to that, my power deliver is going to "wobble" or "oscillate", instead of be a steady 2kW transfer. I'm trying to maintain a constant phase offset, yielding some certain amount of power transfer.

if the inverter is only for on-grid usage as a plugin device, it doesn't need a circuit to make it operate at 50.0000Hz. As its operation is only to be "grid-relative". "Whatever the grid does, I do that too." The inverter still has to know what the frequency is, as the inverter should have its own policy regarding the determination of "grid stability", and the inverter should disconnect if the grid is 45Hz or 55Hz. Similarly the inverter may have to intervene if the plug power limit is being exceeded, and it is shoving too much power into the grid. Excessive output could overheat the inverter. Since a number of edge conditions exist, this is excellent material for microprocessor control.

Paul

So-called "hybrid" solar inverters have battery charging included, and often two AC outputs, one for the grid-tied connection to the home, another that powers devices purely from the battery - these remain powered after the grid goes down for as long as it has solar/battery.

Yes. Match frequency and voltage, then advance the phase until the required amount of power is delivered. Not easy.

WTF is plesiochronous?

So essentially two inverters then...and a battery = an inverter and a UPS

In S Africa where politics has reduced electricity supply to about 15/24 these setups are common now.

Solar panels, and a battery and a system that detects no electricity, disconnects the grid and feeds the house instead. It's expensive, but it can deliver enough power for computers and TV and a fridge. Cooking? use the braai!

I cannot recall whether the solar panels charge the battery or not, or whether the mains when it is there, does that itself

I just looked it up. It gave me a headache.

Indeed. Having looked it up, it appears to be the wrong term to use in this case anyway

Irrespective of its exact value, the grid has but one frequency and all generators are *synchronous* to it, varying only in slight *phase* difference, depending on whether they are powering it, or being powered by it

which is not what 'plesiochronous' appears to mean.

As I aid in 70 years of reading studying and implementing electrical and electronics machines I have never ever heard the term used before now

Quasi-synchronous would have been what I would use

Yes, I wasn't intending to say otherwise, just saying that such a device has a name ...

Fairy Nuff :-)

Probably, yes, but to phase lock it needs to 'guess' the frequency first so must have some sort of timing.

With spinning plant, you can input power by 'spinning faster', ie leading on phase. But for a software-controlled inverter you can also input power by leading on voltage. If the grid is 240V 50Hz phase=phi, if you output 241V

The main thing is this isn't rotating plant, it's a DC input powering an inverter H-bridge controlled via a software-defined control loop. Which means you can do a lot of things in software that spinning plant doesn't do.

Theo