Help in verifying battery / panel sizing

[merelda]

Hi everyone,

I'm new to this forum so apologies if I'm doing anything wrong...

I have a load of 150kWh/day (worst case), normally at 4-8kW but can go up to 16kW.
Since it's a normal building (220-240Vdc) then it's equivalent to about 680Ah.
I need to design a standalone PV system to power the load described above.
My daily radiation is about 5-6 kWh/m2/day and dalylight hours about 10-13 hours (Johannesburg).
And I made my autonomous day to be 7 days...

I run into a lot of problems when I'm trying to size the battery (according to IEEE std 1013-2007),

First of all, I'm not sure what the nominal system voltage is?
I'd imagine it being the inverter (with inbuilt MPPT) input voltage, but when I try to look up an inverter with 16kW 240Vac output, the input voltage range is 200-800V?
How do I get the input voltage so high if I don't connect the panels in series?
But if I do connect the panels in series then I need to connect the batteries in series?
With maximum depth of discharge at 75%, my unadjusted capacity is 7*680/.75=6347Ah
With 24V, 680Ah battery I need 90 of them???

That's a lot of information... but basically I'm just very confused with the nominal system voltage, because from all the examples I see they use like 24V, 48V, etc...
Is it because the load in the example has much lower power ratings?
I think I have something fundamental calculated wrong...

This is the inverter datasheet:
http://www.power-one.com/sites/power....5-outd_au.pdf


Thanks in advance!

Kind Regards,
Merelda

[vtmaps]

I am a bit confused about what you are trying to do. The inverter you mentioned is for a grid-tie system, but I think you are trying to build an off-grid system with a battery.

150 kwh per day with 7 days autonomy is huge and will require a professional engineer to design, if it can be done at all.

--vtMaps

[Cariboocoot]

Welcome to the forum.

I think you are getting the two main types of inverter confused: off-grid and grid-tie. An off-grid inverter has a nominal system Voltage which is based on the battery bank it uses. These will be 12, 24, or 48. Grid-tie inverters don't really have a "nominal system Voltage"; they take input from a PV array and convert it directly to AC without any batteries involved.

Then just to make things more confusing there is the "hybrid" type: grid-tie inverters which do use batteries and do have nominal system Voltage.

The inverter you link to is a standard grid-tie unit. It does not have batteries. As such it doesn't have a nominal system Voltage rating, nor will it provide back-up power in the even of grid outage. This type inverter is meant to feed the grid/loads directly and only needs an array capable of a Vmp in the input Voltage range and sufficient Watts to produce full output.

If you really need 150kW hours per day stand alone you need a massive battery bank. On 48 Volts that is over 3,125 Amp hours meaning the battery bank would have to be larger than 6,250 Amp hours (50% maximum discharge). Don't forget that the inverter itself will need power to run and there is loss in converting from DC to AC. This is definitely forklift battery territory.

For that much battery you would be looking at around 30kW of array. That would actually be multiple arrays and multiple charge controllers.

You would also need to know the biggest load at any given time to size the inverter or inverters.

At this point I kind of hope you slipped a decimal point somewhere and meant 15 kW hours per day, not 150.

[merelda]

Hi,

Thanks for the quick reply and clearly up the grid-tie and standalone inverters :)

My load actually varies between 60 - 150 kWh/day, depending on the season and the holidays, etc.
But I thought for preliminary design I'll just use the worst case.

From the initial calculations, I need 120 of the 250W panels (30kW, taking up 12000 square meters)...
And I need 64 of the 40V, 680Ah batteries in parallel.
I know that the maximum power at any point in time is 16kW, so the inverter should be 48Vdc/240Vac @ 16kW output right?
And yes, I do need multiple charger/controller/strings of arrays..

So, from my very limited experience, I don't think this is feasible at all?

I'm just trying to study the techno-economic feasibility of a standalone PV by sizing the system (as accurately as possible)..
Since I'm not trying to implement the system (yet), would it be more feasible if I maybe look into a grid-tie system rather?

Thanks for the input!

Regards,
Merelda

[Cariboocoot]

Let's put it like this:
Is it possible? Yes.
Is it feasible? Not really.
Is it practical? No.

You would need two Outback Radian inverters to handle that 16kW load ($10,000 right there). You would need that huge battery bank ($54,000 worth of forklift batteries). You would need that gigantic 30kW array (about $40,000) and about ten controllers to connect it with ($5,000). Not to mention wiring, circuit protection, mounting hardware ... You're looking at $150,000 investment to produce electricity that's worth what?

Around here this would be absurd because the utility charges ten cents per kW hour.

With a straight grid-tie system you are limited by how much (if any) you are allowed to back-feed and the fact that it does not supply power when the grid goes down (no batteries).

[stephendv]

So, from my very limited experience, I don't think this is feasible at all?

Hi Merelda,

Without knowing why you're considering this system it'll be difficult to tell whether it's feasible or not. Is the client an ecologist/environmentalist who values clean energy over coal? Is this a factory that wants to cut their electricity bill? or is this a standalone off-grid installation where there it's difficult/costly to get the grid to?
Each of those scenarios would have a different definition of "feasible" :)

There are many ways to design this system too, it need not be solar + battery only. You could do: solar + battery + diesel (with optional hot water cogeneration). And/or solar + battery + diesel + grid. Or gas or petrol generators instead of diesel. Depending on what the goals are you can play with these different variables, if you have a reliable backup source of power then you can undersize the batteries, for example install just 1 day of reserve battery power and rely on the backup generator or grid to take over if that isn't enough.

Entire islands and rural towns are powered using a combination of solar, battery, wind and diesel generation and 16kW of instantaneous demand is really not that high. SMA's multicluster off-grid inverters range from 20kW to 300kW capacity, providing 16kW can easily be done using 3 of their 6kW Sunny Islands for example.

For such a large installation you could also consider SMA's AC coupling idea which doesn't use charge controllers, but uses grid tied inverters + battery based inverters instead. So instead of 8 x 80A charge controllers for 30kW you can use 2 x 15000TL Sunny Boy inverters. You end up spending more on the inverters than on the charge controllers, but then you save on cabling, fuses and time during instalation. You can get an overview of their AC connection concept here: http://www.sma.de/en/solutions/off-g.../overview.html and here: http://www.sma.de/en/solutions/off-g...e-regions.html

The more information you give us about this install, the more ideas you'll get about how to skin this cat :)

[merelda]

Hi Stephen,

Thanks for the suggestions.

I'm asked to simply assess the economic feasibility of replacing the current grid with renewable sources (with diesel generators if needed).
After site screening, it is concluded that solar is the only available source.
The thing is, at Johannesburg, electricity supply is ridiculously unreliable. We get load shedding All the time, and are subjected to a substantial amount tariff increases each year.
The main reason of this investigation is to assess the possibility of going off-grid.

The solar panels will be mounted on the roof of the building, and there are inverters that connects 2 * 700W panels (for testing) to the db board already.

I think I might go with having more back up generators instead of battery as you suggested, although I'm not really certain of what the advantages are at the moment...
The problem on hand now is on how to determine the loss of load probability..

The AC coupling is certainly interesting, but from what I understand from the article you wrote for your blog, the cost of going AC is substantially higher.
I will probably have to work out all of the above before I can decide which one is the best option for my scenario.

Thanks so much for the input, much appreciated!