Grid Tied W/generator Without Batteries

[NorthwestPV]

We have a customer that wants a grid tied inverter connected to their generator but does not want to purchace batteries. I know the Outback and the Sunny Island configuration both require batteries. Do any of you know of a solution?

 

[texie]

We have a customer that wants a grid tied inverter connected to their generator but does not want to purchace batteries. I know the Outback and the Sunny Island configuration both require batteries. Do any of you know of a solution?

Just design it so that the inverter is isolated upon grid source failure.

 

[jaggedben]

If that is truly what the customer wants - "grid tied inverter connected to their generator but [no] batteries" - then you will have to find a way to explain to them why it is fundamentally not possible. It is not a question of finding a 'solution', because there is none.

If the customer simply has a generator and wants to install solar and not have them conflict, take a look at the thread we've just had: http://forums.mikeholt.com/showthread.php?t=148284

 

[ggunn]

We have a customer that wants a grid tied inverter connected to their generator but does not want to purchace batteries. I know the Outback and the Sunny Island configuration both require batteries. Do any of you know of a solution?

So that in the case of a grid failure the PV keeps running? Not a good idea, generally. If the loads are demanding less than the PV contribution, the PV will backfeed the generator with unpredictable but most likely bad results for the generator. There are ways around this but in the grand scheme of things it is unlikely that keeping the PV running during a grid outage will contribute enough to the bottom line to justify the cost and effort of protecting the generator. The safe and simple way is to place the PV interconnect on the grid side of the transfer switch.

 

[K8MHZ]

What is the POCO's take on this?

If they won't allow the configuration (entirely possible that they won't) why bother trying to figure out how to do it?

 

[jaggedben]

What is the POCO's take on this?

It hardly matters what the POCO's take is.

First, the way he worded it, it's something someone could try to do off-grid where there was no POCO at all. (It would fail, but they could try it.) People ask about this all the time over on the Enphase forum, and all the time some of us are explaining that it will not work.

Second, it would still fail and possibly cause damage or fire, even if the POCO okayed it.

 

[NorthwestPV]

What they're looking for is an inverter similar to the Outback Radian, that has the two inputs and an output with a built in transfer that doesn't require batteries.

 

[Besoeker]

If that is truly what the customer wants - "grid tied inverter connected to their generator but [no] batteries" - then you will have to find a way to explain to them why it is fundamentally not possible. It is not a question of finding a 'solution', because there is none.

It's not my field and not something I would do but I can't see any fundamental reason why it couldn't be done.
Isn't it just a system with three generating sources in parallel?

One of the paper mills where we have drives installed gets about 400kW from a water turbine that used to drive DC generators. It's an old mill and it used to use DC for driving mill motors. With the advent of better power semiconductors (SCRs), variable speed DC drives were introduced and the need for DC vanished. But the turbine and genny were still there and the mill owner saw that a source of energy was being lost/wasted. So we built a fairly big inverter (1600A - the DC generator voltage was fairly low) that converted his DC to AC that he could then feed into the mill supply. The mill also had a power station and three steam turbines driving alternators also feeding into their supply. The three alternators plus the inverter could supply about half the mill load and power was also imported from the utility company.

So the system had three generating sources all operating in parallel. An inverter, local generators, and the grid. It can be done.

A little chronology for anyone interested.
The inverter was installed around 1975 and the steam turbines pre-dated that by probably 20 years at a guess.
The mill race is 1895. I don't think the "new" water turbine dates from then but it is at least 70 years old. And is the only part still in operation.
The DC generator and inverter have made way for a three phase cage (asynchronous) motor driven at above synchronous speed.

 

[ggunn]

What they're looking for is an inverter similar to the Outback Radian, that has the two inputs and an output with a built in transfer that doesn't require batteries.

The short (and mostly accurate) answer is that you cannot run AC loads from PV without batteries. On PV dedicated forums solar newbies are continually upset that their grid tied PV systems will stop working when the grid goes down and they are always floating schemes to "fool" their system into thinking that the grid is there with tiny batteries and such. It won't work.The demand curves from the loads and the generation curves from the PV do not match up; you need a battery bank to deposit into and withdraw from in order to service loads on demand.

 

[jaggedben]

So the system had three generating sources all operating in parallel. An inverter, local generators, and the grid. It can be done.

Yes, you can add small generators to a grid without batteries. What you cannot do is connect a PV inverter designed only for grid-tied applications to a small ICE generator when there is no grid.

In your story, surely there must have been a provision to automatically shut down your big inverter if the grid went out, right? Either that, or your fancy inverter was configured to follow the AC load if the grid was not present.

A grid-tie PV inverter, by the usual definition, is not configured to follow the AC load. And that is why the idea, as verbalized by Northwest PV, is not possible.

 

[jaggedben]

What they're looking for is an inverter similar to the Outback Radian, that has the two inputs and an output with a built in transfer that doesn't require batteries.

The Radian does require batteries. There is no similar animal that is an inverter that does not require batteries. As for the transfer switch, yes, you need the transfer switch to separate the grid-tie inverter (and the grid) from the generator when the grid goes down.

ggunn summed up the situation pretty well, except I would amend his statement "you cannot run AC loads from PV without batteries or the grid".

 

[Besoeker]

Yes, you can add small generators to a grid without batteries.

Not sure what you class as small. The inverter was 400kW and did about what PV systems do - it took available DC and converted it to fixed frequency AC. The steam turbine driven alternators in aggregate were 6MW. Mill consumption was around 11MW to 14MW

In your story, surely there must have been a provision to automatically shut down your big inverter if the grid went out, right? Either that, or your fancy inverter was configured to follow the AC load if the grid was not present.

At the sort of powers involved, there is fairly sophisticated protection. You wouldn't want to try to feed the whole of UKplc with just 6.4MW. That said, the plant could have worked fine in the absence of the grid if the total load was reduced to within the on-site generating capacity and the inverter would have worked just fine.

A grid-tie PV inverter, by the usual definition, is not configured to follow the AC load. And that is why the idea, as verbalized by Northwest PV, is not possible.

I took it to mean when he/she stated grid tied he/she meant grid tied i.e. tied to the grid in which case my original comments stand.

 

[jaggedben]

Not sure what you class as small.

How about small enough to never possibly provide more power than the load on the grid?

the plant could have worked fine in the absence of the grid if the total load was reduced to within the on-site generating capacity and the inverter would have worked just fine.

And if the load was less than the generating capacity? Aye, there's the rub. I'm sure you had some sort of protection to prevent the inverter from putting out it's full capacity when that capacity had nowhere to go. Now with your mill maybe it typically had a constant large load so this would rarely be a problem. But with a typical residence with a net-metered solar system the load will almost always be lower than the solar system capacity. So there is no point to allowing the solar to ever connect to the generator because it will a) cause the solar to shut down, or b) cause both to shut down, or c) cause damage to the solar or the generator or the electronics in the house experiencing the voltage surges. That is, unless you have an inverter with batteries so you can store the extra energy.

I took it to mean when he/she stated grid tied he/she meant grid tied i.e. tied to the grid in which case my original comments stand.

Except that makes no sense because why would a residential homeowner run a fuel generator when the grid was present. Clearly the original question was actually about running the generator and the solar together when the grid was not present.

Basically, the points you are making, while technically correct, cannot be applied to residential solar in any practical way. The practical and economic approach is energy storage.

 

[Besoeker]

How about small enough to never possibly provide more power than the load on the grid?

I don't think any inverter could support the load on the grid. Even in little UK, the generation capacity into the grid is over 50,000 MW.

And if the load was less than the generating capacity? Aye, there's the rub. I'm sure you had some sort of protection to prevent the inverter from putting out it's full capacity when that capacity had nowhere to go.

It isn't really a rub. The generating system doesn't have to put out its full capacity.
For the alternators there were AVRs that alter the excitation so their 6MW capacity didn't have to be fully utilised. This usually how these things, the alternators, even at utility level work. Their output has to match demand even if their capacity exceeds that demand.
Similarly with the inverter powered from from the water turbine. The turbine has a governor to regulate speed and thus the generator output. In this particular case the governor is mechanical, remember that it is very old, and it closes turbine inlet vanes. The excess water tumbles over the weir. Pretty to watch.
And the grid supplied what was required of it.

Now with your mill maybe it typically had a constant large load so this would rarely be a problem.

It wasn't.
But a couple of points.
That paper mill, like many others, routinely had what they call a summer shut. All the production machinery is shut down for annual maintenance. There used to be a joke, a wry view maybe, that we shut for a couple of weeks to fix everything then spend the next two months sorting out the new problems....
Levity aside, the required power was vastly reduced. Process steam wasn't required. The water turbine carried on as usual, still grid connected.

But with a typical residence with a net-metered solar system the load will almost always be lower than the solar system capacity.

You don't have to run the solar system at rated capacity.

 

[jaggedben]

You don't have to run the solar system at rated capacity.

As you said, this isn't your field. If it was, you would know that there are no residential-size grid-tie-only inverters on the market that are capable of limiting their capacity. They are all configured as current sources that output whatever current is available at the grid voltage they find. (I know of one exception, but it isn't really an exception because the feature only works in conjunction with a battery-connected off-grid capable inverter.)

Sometimes on solar forums people start suggesting Rube Goldberg methods for determining the required load and shutting off grid-tie inverters if the the load exceeds the solar capacity. I have never heard an account of someone actually successfully doing this. It's simpler, more reliable, and for most people worth the tradeoff of time vs. money, to simply use one of the various off-grid inverter solutions that utilize batteries.

(I faintly recall one well-known manufacturer announcing that they had a multi-source control system product in development, but I don't recall that it was ever released as a commercial product. And it probably included batteries.)

It isn't really a rub. The generating system doesn't have to put out its full capacity.

It's a rub if your generating system is solar, because the energy isn't always available when you want it. This is why generally all off-grid solar systems utilize batteries.

BTW, your description of the mill hyrdo-turbine still hasn't explained how the generator knows what to limit the output to in the case where the grid goes out.

 

[Besoeker]

BTW, your description of the mill hyrdo-turbine still hasn't explained how the generator knows what to limit the output to in the case where the grid goes out.

Not sure which bit of this you didn't get.

Similarly with the inverter powered from from the water turbine. The turbine has a governor to regulate speed and thus the generator output

 

[jaggedben]

Not sure which bit of this you didn't get.

What controls the governer? You said it was old and mechanical. Something must control it.

 

[Besoeker]

What controls the governer? You said it was old and mechanical. Something must control it.

The governor is the controller. It regulates speed.
That's why it's called a governor.

 

[jaggedben]

Does the governor limit the output to the required load in the case where the grid is out?

 

[BillK-AZ]

Note

Rotating AC generators have two control loops:

1. A governor to control rotating speed and hence the frequency,
2. A field control that controls the voltage.

The result is an output at a fixed frequency and voltage. The accuracy of the frequency and voltage depends on the sophistication of these control loops.

If a generator is to operate in parallel, these control loops are used to synchronize the phase and voltage before connecting. Damage and voltage swings can result if this is not done just right. With parallel generators, a little backfeed is normal at light loads as the load changes and the control loops respond. The control loops must be designed for parallel operation, or they can basically fight each other and affect operation. Most of the instructions for smaller AC generators I have seen prohibit parallel operation because the control loops are not suitable for parallel operation.