SMA Sunny Island AC coupling

stephendv

I recently attended a training course in the SMA Sunny Island off-grid inverter/charger and overall was very impressed with it. One of the notable differences between it and other inverter/chargers is its ability to allow power generation sources such as photovoltaics and wind turbines, to be connected on the AC bus.

It was interesting to note that half of the attendees at the course were planning to use a DC coupled system with the Outback charge controller instead of AC side coupling. I wrote an article that explores both the advantages and disadvantages of AC side coupling using the Sunny Island and Sunny Boy with more traditional DC side coupling using the Outback MPPT charge controller.

To summarise briefly:

Advantages of AC coupling:
- Simpler and cheaper wiring
- Wider choice of PV panels
- Supports long distances between PV and batteries
- Allow flexibility in future expansion

Disadvantages:
- Higher initial cost (with some caveats)
- Less efficient for most off-grid applications such as off-grid homes

Full article here:
http://www.casanogaldelasbrujas.com/blog/2009/05/13/sma-sunny-island-ac-coupling/

BB.

Re: SMA Sunny Island AC coupling

Stephen,

I posted a reply to your other SMA post comparing the XW with the 5048u... But I have a specific question--from your article:

The Sunny Island changes the grid frequency to manage the available and required power - which happens to be the same way the commercial electric grid responds to surplus and deficit power: when too much power is available in the grid the frequency rises slightly, and when there’s too little it falls.

Generally, the grid frequency is kept in a very small window and while it does drift up and down a tiny bit--as far as I understand--it is not used to regulate power consumption of the grid.

So--without reading a ton of SMA's technical literature--how does the typical SunnyBoy Grid Tied Inverter interact with the SunnyIsland inverter system?

The GT inverter normally converts 100% of the available PV power into AC power (more or less by raising the voltage of the AC line). And the only time a GT inverter "regulates"--is when one or more the AC line parameters (voltage, frequency, Power Factor, etc.) falls out of range.

The problem occurs when the Grid Consumption is less than GT Inverter production (like in the middle of a sunny day, grid requirements are low, battery banks are charged) and the "standard" GT inverter has no mechanism for throttling back its AC output (other than faulting when AC Grid voltage/frequency is out of range).

So--how do the SunnyBoy GT inverters work with the 5048 Island? Are you telling me that they Island raises the frequency of the line (by 0.x Hz) which signals the SunnyBoy GT's to throttle back their output to less than available Solar Panel Vmp*Imp?

Or--is there some other mechanism that is used (there is a communications link available--required?--between the various SMA function blocks)... Would a com-link tell the other SMA's to throttle back (reduce "off-grid sales" which appears to be a parameter for GT inverters special ordered for Island functions)? Or would they just use "droop sharing"--i.e., as the shared AC voltage rises, they "sell less" to the grid (even the SunnyBoy GT inverters)?

-Bill

boB

Re: SMA Sunny Island AC coupling

BB. wrote: »

Stephen,
~~~~ ~~~~~~~
Generally, the grid frequency is kept in a very small window and while it does drift up and down a tiny bit--as far as I understand--it is not used to regulate power consumption of the grid.
~~~~ ~~~~ ~~~~~ ~~~~
So--without reading a ton of SMA's technical literature--how does the typical SunnyBoy Grid Tied Inverter interact with the SunnyIsland inverter system?

~~~~~~~~ ~~~~~~~ ~~~~~~ ~~~ ~~~

-Bill

From what I hear, the Sunny Boy and Sunny Island communicate the power needs to each other by nudging the grid frequency a wee bit back and forth. This sounds pretty neat and I guess it works OK?

You know, you can also AC couple a Sunny Boy (and other GT inverters) to a regular bi-directional AC <--> DC inverter like an OB FX or a Magnum MS. The thing about tying to those "grid substitutes", ("I just can't believe it's not grid ! "), are that you must do something to keep from overcharging the batteries and keeping the grid voltage from going out of the ieee-1547 window... Like, dump loads or something. This is becoming quite popular these days and works very well (in most cases.)


boB

BB.

Re: SMA Sunny Island AC coupling

boB wrote: »

You know, you can also AC couple a Sunny Boy (and other GT inverters) to a regular bi-directional AC <--> DC inverter like an OB FX or a Magnum MS. The thing about tying to those "grid substitutes", ("I just can't believe it's not grid ! "), are that you must do something to keep from overcharging the batteries and keeping the grid voltage from going out of the ieee-1547 window... Like, dump loads or something. This is becoming quite popular these days and works very well (in most cases.)

That is the part I am wondering about... Between batteries requiring 3 stage chargers (with three different voltage requirements), temperature compensation, and a possible redundant failure protection mode--How are people doing this?

-Bill

stephendv

Re: SMA Sunny Island AC coupling

BB. wrote: »

Stephen,

So--how do the SunnyBoy GT inverters work with the 5048 Island? Are you telling me that they Island raises the frequency of the line (by 0.x Hz) which signals the SunnyBoy GT's to throttle back their output to less than available Solar Panel Vmp*Imp?

Exactly. The range when the Sunny Boy starts throttling it's output is between 51 and 52 Hz. At 51Hz the sunny boy still produces 100% power, at 52Hz, it produces 0%. The Sunny Island shifts the frequency to accomodate the demand and supply. Saw it working in their lab, was quite impressive

Some clocks depend on the grid frequency for accuracy, since the Sunny Island keeps changing the frequency this would throw them off. So there's even a function in the SI to compensate for this change (on average).

Also, there is no physical difference between the Sunny Boy's used in off-grid systems with the sunny island and those used for Grid tie. They just need 1 parameter changed in their configuration to work in off-grid mode. All this does is disable the impedance test when testing whether the grid is active or not.

stephendv

Re: SMA Sunny Island AC coupling

BB. wrote: »

That is the part I am wondering about... Between batteries requiring 3 stage chargers (with three different voltage requirements), temperature compensation, and a possible redundant failure protection mode--How are people doing this?

At the risk of sounding too much like an SMA fan boy (is that a new product?), a feature I like about the AC coupling is that all the battery charging is controlled only by the Sunny Island using it's own SoC algorithm. Having multiple devices all applying their own charging regimens to sensitive batteries seems like a bad idea.

SMA also have a DC charging unit available, and it acts as a slave to the sunny island through a dedicated communication link so charging is still controlled by the SI. The DC charger is ridiculously expensive, so I don't pay it much attention.

boB

Re: SMA Sunny Island AC coupling

stephendv wrote: »

a feature I like about the AC coupling is that all the battery charging is controlled only by the Sunny Island using it's own SoC algorithm. Having multiple devices all applying their own charging regimens to sensitive batteries seems like a bad idea.

Yes, the SI is probably the best way to take care of the charging with the AC coupling with a Sunny Boy.

How would the SI work with other batteryless PV inverters though, like, the Fronius, which other ppl also use, (among others) ?

Another method (inverter as grid), works very well though, from the numerous examples I have seen (and tried myself). I think that the only problem is the price of the SI as compared to the other options. And, some people may already have the other inverter and not want to shell out the extra $$ for a SI and just deal with the charging issues.

Please keep us posted on your progress if there is anything new to report.


One way to do the 3 stage charging could be with something like a C-40 (diversion mode) that works on voltage and time to adjust the voltage setpoints. When the Sunny Boy (or whatever) is supplying most the energy, you would want the battery charging voltage to be just slightly above resting voltage, but not too high so you don't cook them and waste energy keeping the batteries at that voltage. I would think that the SI does this already, but I haven't looked into it that closely.

stephendv, how does the SI take care of the battery charging ?
boB

BB.

Re: SMA Sunny Island AC coupling

Stephen,

At this point--it would appear that SMA / SunnyIsland is the only "integrated" system out there--a good thing for them.

And the concept of back-driving a regular Off-Grid Inverter using a Grid Tied Inverter has been known for a while--and the SWWP Skystream originally offered that as their way of going off grid with the Skystream Wind Turbine (GT Inverter in turbine--using an Outback(?) inverter for "local grid", and using SkyStream's RF link to control the turbine's output...

I am not sure that this configuration ever shipped.

The nice thing with the "Open Sourced" GT + Off-Grid system--over the years it would be possible to 2nd source / upgrade / competitive bid for the optimal system--vs SMA's (appears to be) closed architecture (other than varying frequency--apparently a well thought out architecture).

However, at this point, I do not know of any off-the-shelf method to impliment the GT+OG inverter option--I can see ways of doing it (dump loads, using Digital Relay in the GT output to turn off when the batteries are full--etc.)--but (I have) more questions than answers at this point.

-Bill

stephendv

Re: SMA Sunny Island AC coupling

boB wrote: »

How would the SI work with other batteryless PV inverters though, like, the Fronius, which other ppl also use, (among others) ?

Firstly, I don't know if other inverters would even detect a mini-grid because of the low impedance. The GT sunny boy uses an impedance test in addition to voltage and frequency to determine if the grid is there or not - this additional test has to be turned off for SI mini-grids because the impedance is too low. As long as they don't check impedance then they should in theory be fine.

If other inverters could be configured to reduce their output based on grid frequency - or perhaps even shutdown completely at a certain frequency then they might work, but it would be a bit cludgy. E.g. you could reconfigure the sunny island so that the upper limit of the frequency shift was at 54Hz, then if you could configure the fronius to completely shutdown at that frequency, the system might work. But it would be a binary on - off as opposed to a gradual reduction of power as required.

boB wrote: »

Another method (inverter as grid), works very well though, from the numerous examples I have seen (and tried myself). I think that the only problem is the price of the SI as compared to the other options. And, some people may already have the other inverter and not want to shell out the extra $$ for a SI and just deal with the charging issues.

It was interesting that half the attendees at the SI course were planning to use it with an Outback charge controller primarily for reasons of cost... and it is a compelling argument, hope SMA consider adjusting the prices for their sunny boys when sold with a sunny island - or bring down the price of their own charge controller.
The price of the Sunny Island itself, in Europe anyway, is quite competitive with other 5kW inverters. The Victron Quatro and the Studer are both slightly more expensive than the Island.

boB wrote: »

stephendv, how does the SI take care of the battery charging ?

Not sure what you're asking... the charging is the same with any other inverter/charger. 3 stage charging with configurable battery level alarms to start a generator or perform some other action. Low battery level shutdown. If there is too much power available on the mini-grid (e.g. the frequency shift isn't working), then it will shut it down.
The SMA engineers are very proud of their SoC measurement
The SoC is determined by 3 methods:
The Ah going into the battery and coming out are counted.
The temperature of the battery is monitored
The voltage
Then the SI also learns to adjust it's SoC algorithm during the lifetime of the SI. Apparently it can take months of use for the SoC algorithm to refine itself.

stephendv

Re: SMA Sunny Island AC coupling

BB. wrote: »

The nice thing with the "Open Sourced" GT + Off-Grid system--over the years it would be possible to 2nd source / upgrade / competitive bid for the optimal system--vs SMA's (appears to be) closed architecture (other than varying frequency--apparently a well thought out architecture).

However, at this point, I do not know of any off-the-shelf method to impliment the GT+OG inverter option--I can see ways of doing it (dump loads, using Digital Relay in the GT output to turn off when the batteries are full--etc.)--but (I have) more questions than answers at this point.

SMA also produce a Smart Load product that is intended to be connected to the grid if you have unmanaged power generators connected to it. This is probably the official solution if you want to connect another inverter to the grid. The price is a problem, I think around 600 euros.

Towards the bottom of this page: http://www.windandsun.co.uk/Inverters/sunny_island_inverters.htm there are some "Distributed Intelligent Load Controllers" that appear to do the same thing. Around 60 euros each