how do grid-tie inverters inject power back into the grid?

[hamster_nz]

I can see how generators do this, but if an inveter exactly exactly matches phase, frequency and voltage then no power will be transferred... Do they have to run a tiny bit ahead of the grid in phase, or at a higher voltage (whatever that means in AC)? Either of which makes power flow into it...

[fourtytwo42]

A grid tie is a current source, so it injects current in phase with the grid in a sine waveshape of the same frequency and phase. The amplitude of the current waveform determines the power.
Internally the voltage source from where the current is derived is higher than the grid in order to overcome internal losses.

[Psi]

Do they have to run a tiny bit ahead of the grid in phase, or at a higher voltage

Put simply, yes.

If it was 100% matched in voltage and phase no current would flow.

This is why it's so critical the electronics/code in grid-tie inverters works correctly.
One little error and it's a no holds barred fight to the death between the inverter and the national grid. (National grid wins)

[SeanB]

Fuse or circuit breaker though is the referee, though in most cases the clear winner will be the network, with the inverter being the smoking ruin left afterwards.

[Circlotron]

Whatever the ac voltage waveform is at any particular instant the inverter needs to pull up against that. It doesn’t have to succeed in pulling it higher, only try. As mentioned, if it is pushing a current into the mains that is proportional to the instantaneous voltage then all is good.

[T3sl4co1l]

The grid is a very low impedance.  Step 0: don't connect a voltage source to it!

Drive it with current.  You're making a class D audio amp, CCS output, driven from a phase locked sine wave oscillator.  Phase shift to inject reactive power.

This also answers the question: what to do when the grid goes down?  You have no voltage reference, so the voltage goes outside the expected range and protection kicks in.  Grid tie inverters are not grid island inverters!

You could make one that operates in voltage or current mode, of course, perhaps even alternately depending on condition, say if you're using a GTI with disconnect as household UPS.  Grid dies, disconnect opens, house maintained on voltage mode.  (When grid is detected again, PLL first resyncs, then disconnect closes.  Switching transient is limited by peak current output -- which is fixed, because you'd never actually get rid of current-mode operation, just set peak current output to maximum for best voltage regulation -- and then it flips back to grid-tie current mode.)

NB: I haven't actually seen an internal design, before.  The above is intuitively how I'd do it.

Tim

[Psi]

Fuse or circuit breaker though is the referee, though in most cases the clear winner will be the network, with the inverter being the smoking ruin left afterwards.

its like the few times in history when powerstations have been switch back onto the national grid 180deg out of phase.
There was one in either Russia or South Africa and the entire motor/generator assembly (were talking the size of a small house and mostly solid steel) exploded up into the air and landed many km away.

[tszaboo]

Yes, it has higher voltage than the grid voltage. Unlike some people describe the grid is not low impedance. Just with electronics terms. You have a feeder, this is usually a HV to LV transformer, on it's output you have 230V. You connect 10 houses to it. They are in parallel, but they have different length cable connecting to the grid. So house 1 will see 229V, house 2 228, house 3 227, house 10 is 220V. You are house 4 you have 226V before you turn on the inverter. After you turn it on, it is 228V. So current will flow to house 3 also, and the voltage for house 5,6,7 will increase (current flows that way also).

If all the homes have solar panel, and they dont have load at the moment, the voltage can increase above the grid's limit, and some of the inverters will turn off. And the voltage will be the lowest at the transformer.

Also, notice how it takes power to raise the grid's voltage. Power, which comes from the solar panels. So it cannot really go haywire and mess up the network, by inserting 1000V into it. It can mess up the harmonics though.

[Mjolinor]

It's worth remembering that the business case for distributed generation is mostly valid because distribution losses take about 17% of the electricity generated. Moving the generation to the customer location, as you do with small renewable generation, can cost 17% more but is still viable.

It would be interesting to see if the technical and non-technical losses have changed in the last ten years or so with the increase in local generators.

[Seekonk]

I'm going to be hooking a cheap GTI up to an off grid system.  There are power inverters that will feed power back to a battery.  If the battery is fully charged and can't take any more, the inverter may be damaged.  These inverters usually also have a way to send power to an external dump load.  There are many inverters that can't do either. Then the inverter over voltages and at a minimum the caps blow up. I'll have to build a fast dump for mine.  There is no shortage of cheap blown up GTI.  If looking for a payback over time, look to spend a little more.

[tszaboo]

It's worth remembering that the business case for distributed generation is mostly valid because distribution losses take about 17% of the electricity generated. Moving the generation to the customer location, as you do with small renewable generation, can cost 17% more but is still viable.

It would be interesting to see if the technical and non-technical losses have changed in the last ten years or so with the increase in local generators.

Since solar power accounts to only some 5% of the energy generation in the forward thinking part of the world, the benefit is maximum 5% of that 17%. AKA almost nothing.

[Mjolinor]

It's worth remembering that the business case for distributed generation is mostly valid because distribution losses take about 17% of the electricity generated. Moving the generation to the customer location, as you do with small renewable generation, can cost 17% more but is still viable.

It would be interesting to see if the technical and non-technical losses have changed in the last ten years or so with the increase in local generators.

Since solar power accounts to only some 5% of the energy generation in the forward thinking part of the world, the benefit is maximum 5% of that 17%. AKA almost nothing.

Hmm, well I certainly wouldn't object to that 5% being in my bank account.

It really is a huge amount of money. It is about what is lost in the UK through non-technical losses.