Can I run a 100kW inverter off a 1MW busbar?

This is much bigger stuff than I'm used so, but as general principles...

MPPPT - generally that would have to operate independently for each group of panels that could produce output that varied compared with other groups - so if all your panels were identical and installed with the same orientation and lack of shading etc. I guess you could just have one for the lot. More likely you'd need something for each string before they're combined into a single bus (there are DC-DC MPPTs out there for small (battery charging) systems but I've no idea if there's anything for your size of system). Or I suppose you could go old school and just parallel everything up with a few blocking diodes and not worry about extracting maximum efficiency from it.

As for grid-forming (i.e. off (national) grid systems) - it all rather depends on how you intend to match supply and demand. Presuming you can't instantly alter demand to match supply, a battery driven inverter is certainly the most common option. But there are subtle variants out there. Some system can for instance have a conventional generator in the mix, and the PV inverter in effect reduces load on the generator (often the generator is the grid-forming element and the PV system behaves as if it's a conventional grid connected inverter). Likewise I've seen descriptions of systems where a small battery generated inverter does the grid forming, and can work with conventional PV inverters (if necessary tweaking the "grid" frequency to modulate PV inverter output) and/or calling on a conventional generator. I suppose if you had more controllable renewable sources available (e.g. hydro from stored water) you could add that into the mix too. Again these are typically much smaller systems (off grid homes/RVs etc. so may or may not scale). So never say never, but usually a battery is the most practical solution.

   - Andy.

This is way outside the normal rooftop solar scale.  But here are my thoughts.

It's not unusual to have en inverter that's under-sized compared with the kWp figure from the panels.  It means that on a bright sunny day, the inverter will attempt to MPPT, but eventually will hit the maximum power that the inverter itself can deliver.  Yours seems a rather extreme example of doing that.

If you're operating in a cloudy country like the UK, don't assume that 1MWp of panels can reliably deliver 100kW.  If you have a grid-forming (i.e. islanded mode) inverter, then the output could collapse suddenly if the demand exceeds what the panels are supplying.  That's a really good reason to add a battery.

Simon Barker said:

If you have a grid-forming (i.e. islanded mode) inverter,

Is this to be grid forming of the National Grid kind, which looks to mimic rotating machine generation, and HV transmission transient problems, or some other sort of limited grid? 

There's a lot of Catch 22 discussions go on about which parts are 'cause' and which are 'effects', and which are dominating and which are incidental. Historically inverters were incidental to the grid...

100kW inverter off a 1MW busbar

Are there other things tapping off from the busbar, other than the proposed inverter?

   - Andy.

I hadn't thought of that.  A single inverter can run in "islanded" mode.  But there's also the possibility of using an inverter to "black start" a local microgrid.

That's way outside my expertise.

Hi Sara,

Just to mention, this isn’t my main field of expertise, but here’s what I understand. The MPPT algorithm will still function with a single input, but without multiple strings there’s no balancing, so the process is simpler. If shading or mismatch occurs, you lose flexibility because you can’t isolate or optimise separate strings, so while it works, you miss out on the benefits of multiple MPPT channels in more complex arrays.

Regarding grid-forming inverters without a battery, these normally need a stable energy source to maintain voltage and frequency reference. Batteries are common because they provide fast response and stabilisation. Without a battery, it’s technically possible if the inverter can use the solar array as the source, but solar is variable and can’t always respond quickly to load changes. Most manufacturers recommend at least a small battery or another stabilising source, such as a generator, for proper grid-forming behaviour. In short, operating without a battery is risky unless the inverter is specifically designed for PV-only grid-forming and the load is very predictable.

If I'm wrong on any of this please point it out as I'm always learning

Can I run a 100kW inverter off a 1MW busbar?

I think this question means can you connect 1MWp PV array to a 100kVA inverter input.

This would give you a 10:1 DC:AC ratio. In C&I installations, a 1.2:1 is about average in the UK (subject to site-specifc considerations, which can cause considerable variation); in utility scale that ratio is nowadays often 1.6:1 or even edging towards 2:1 at the extreme. So this is a long way off piste.

Inverters normally operate with a limit from the AC output current, which, by and large, is where the inverter nominal power rating comes from. The MPPT algorithm operates by adjusting the DC bus voltage; normally this will be to maximise power (hence the MP of MPPT) but if it cannot due to excess power it will adjust the voltage along the IV curve, causing the array to operate less efficiently but within a range the inverter can convert. If it cannot get to an acceptable operting regime within the MPPT range, the inverter will shut down (going open circuit) due to overpower.

The same applies if you took a normal inverter configuration and limit the output through control parameters (e.g. G100)

Sara Helin said:

If there is a single input (not multiple strings) what happens?

As it's all on one DC bus into a single MPPT, it makes no difference how the downstream DC side is arranged. All the inverter "sees" (acts on) is one IV curve being the sum of the subarray.

Sara Helin said:

Can you operate a grid-forming inverter of this type without a battery?

Yes in principle, with ideal components.

However, with such an extreme oversizing, three extra considerations in practice are

• It is quite likely that the DC Isc will exceed the inverter's rated maximum, which is driven by busbar thermal ratings, switcgear capacity etc. This is an important safety consideration
• With such high oversizing you will probably invalidate the manufacturer's warranty. I have never seen a manufacturer cover such a high ratio
  • The MPPT will be forced to operate at a higher voltage more frequently than usual which could create greater stress on the components
  • The DC-DC conversion will require PWM switching a greater Isc than designed, and with a greater capacitance, which will again stress the components
• You will likely struggle with the cable terminations and/or overcurrent protection
• Due to the overpower shutdown, you will get far less energy out than you would expect. You'll probably get more by reducing the array size to something a bit more sensible

These considerations would also apply if there was a battery though!

This being the case it's probably a bad idea!

You have not said why you find yourself contemplating this - and I should say this sort  of power is well off the end of my experience too, but it seems a slightly odd thing to do.
It is fair to say that in the UK, a nominal 1kW panel rarely does generate the full 1kW output , so it is standard to have an inverter that looks slightly undersized relative to the panel array full power rating, especially if the array is not optimally aligned for midday sun.  However unless you are expecting to stack up five or more  such inverters, with a ten to one ratio,  the bulk of your power will be lost on sunny days and the MPPT circuit will spend a lot of time bumping on its upper limits, which is not really making best use of the panel array. You do not say how the bus voltage is derived - are there additional balancing inverters on sub groups of panels ?

As an aside to your question, but an area I know more about, large inverter type installations need to be thought out very carefully in terms of cable routing and if need be additional filtering  to avoid creating EMC interference problems. It is very easy to create accidental loop antennas, and the more amps in circulation, the more they radiate harmonics of the switching frequency/ frequencies and the more there are the more likely the site as a whole is likely to exceed emission limits. Enforcement is not usually very strong in the UK but that does not make it the responsible thing to do.
So, what  is the driver for considering this approach ? 

Mike