Exact size of the Transformer for a Commercial solar project.

[inforaj]

Hello,
Could you please help me figure out the exact size of the transformer that is back-feeding to the utility? Is the procedure similar to a typical transformer, or is there any factor that needs to be considered?

Three Phase Transformer Example: V = 208, I = 175; Therefore: kVA = (208 x 175 x 1.732) / 1000 = 63.05kVA; this calculates to 63+ kVA, thus we round up to a standard Three Phase size 75kVA.

Any reference or links for the calculation will be highly appreciated.

 

[PWDickerson]

I don't design lots of systems with transformers, but there is nothing special about calculating the size of a transformer for a PV system. Your math looks right to me. The trick is ordering the right transformer, and I will probably screw that up, so I will other chime in on the specifics. You need to connect the utility conductors to the primary side of the transformer even though power will be flowing in the opposite direction (from the secondary side to the primary side), and this is because the utility is the source for in-rush current. You also need to specify the correct Y-delta configuration.

 

[ggunn]

Hello,
Could you please help me figure out the exact size of the transformer that is back-feeding to the utility? Is the procedure similar to a typical transformer, or is there any factor that needs to be considered?

Three Phase Transformer Example: V = 208, I = 175; Therefore: kVA = (208 x 175 x 1.732) / 1000 = 63.05kVA; this calculates to 63+ kVA, thus we round up to a standard Three Phase size 75kVA.

Any reference or links for the calculation will be highly appreciated.

Even simpler, select a transformer with a kVA greater than the AC kW rating of your PV system, i.e. the total kW rating of your inverter(s).

 

[inforaj]

Even simpler, select a transformer with a kVA greater than the AC kW rating of your PV system, i.e. the total kW rating of your inverter(s).

There was little difference between my calculation and the Utility calculation; I did not agree with the utility calculation.

PV SYSTEM TOTAL KW = 1200 KW (6*200kw = 1200kw)
INVERTER EFFICIENCY = 98.5% = 0.985
LOADING DEMAND FACTOR = 90% = 0.9
TRANSFORMER SIZE = 1200 * 0.985 * 0.9 = 1063.8 KW
TRANSFORMER POWER FACTOR = 0.9% = 0.9 ( CHECK THE TRANSFORMER DATASHEET)

TRANSFORMER POWER FACTOR = 1063.8 / TRANSFORMER PF = 1063.8 / 0.9 =1182 KVA = 1250KVA- standard size.

BUT,
UTILITY CALCULATED,
PF = 0.85,
LOADING DEMAND FACTOR = 90%

1200 * 0.85 * 0.9 =910 KW - 1000 KVA IS SUFFICIENTS. (utility said)

 

[jaggedben]

Okay I'm not an engineer so I won't speak to all of the issues but I know enough to make a couple comments.

First, I'm pretty sure inverter efficiency shouldn't be part of the calculation. That's a measure of DC to AC efficiency and you are interested in the AC output of the inverter(s), which can be the AC datasheet maximums no matter the efficiency. So inverter efficiency just can't be counted up to reduce the transformer loading in any way. (Perhaps it affects the loading factor, but your loading factor seems like an educated guess anyway.)

Second, I wouldn't start with an inverter kW rating if I'm supposed to be doing a calculation based on kVA and a set power factor. I would start with either the datasheet kVA of the inverter, if given, or else the continuous output current rating multiplied by the nominal voltage (and then multiplied by 1.73 for three phase, of course). Because this may be different than the inverter nameplate kW.

 

[Elect117]

Without really understanding the purpose of of the transformer, oversizing can be better but also means a higher impedance (kwh consumption) and higher fault currents (FLA/%Z).

A larger than needed transformer will leave head room for harmonics and reactive power.

Most transformers can be overloaded effectively (~150%) without negatively impacting their life span, For the application of solar, the transformer will have time to cool at night. The same is true with keeping it in a cooler spot (shade or inside).

Most utilities also approach transformer sizing / loading differently than the way the NEC does.

 

[electrofelon]

Is this a utility transformer being supplied and owned and maintained by the utility? If so why do you care what they come up with?

I agree with jaggedben, I don't agree on your use of inverter efficiency, nor do I agree with your use of transformer power factor. Not sure if that is exactly what the utility wrote, or your transcribed it wrong, but power factor should be divided not multiplied. Finally, not sure I agree with use of a loading demand Factor, considering today's high DC/AC ratios, you will get Max output for a large chunk of the day

 

[electrofelon]

Some additional comments and thoughts:

Regarding power factor: I just wanted to clarify that for the utility calculation, I am assuming the PF is for the inverters NOT the transformer. Yes GTI's have a PF very close to unity, but perhaps they are being conservative and/or are allowing for a change of GTI PF (which to my understanding is something utilities may like to do to adjust for grid conditions). I still say you wouldn't have a field for the power factor of the transformer that you are doing loading calculations for, perhaps someone can explain that to me if I am incorrect.

Regarding loading demand Factor: I'm going to walk back what I said a little as utilities certainly have a lot of experience sizing transformers and it could certainly be reasonable that they conclude a 10% overload even for say five or six hours would be fine. If I was sizing a customer-owned transformer, I would probably be more cautious and see if I can find some temperature/loading/expected life curves from the manufacturer before I did such a thing.

I worked for a crew where we did several multi megawatt projects connected to Georgia power. Their calculation was quite simple and just required the sum of inverter KVA to match or be less than transformer KVA. You will probably find a myriad of slightly different methods and calculations out there.

 

[ggunn]

I guess if I had to design a PV system with an AC rating of exactly the kVA of a standard size transformer I would take a look at power factor concerns, but in practice that has never happened to me. I always select the next standard kVA transformer up from the AC kW rating of the system.

 

[jaggedben]

Again, I'm not the engineer, but I don't see how ignoring power factor can be anything but conservative. Given that no grid-tied inverter that I know of can have a lower power factor by raising apparent power, as opposed to lowering real power.

 

[Carultch]

Again, I'm not the engineer, but I don't see how ignoring power factor can be anything but conservative. Given that no grid-tied inverter that I know of can have a lower power factor by raising apparent power, as opposed to lowering real power.

Some inverters have headroom built into their ratings on the datasheet, for non-unity power factor. For instance, SMA's 62.5kW Core-1 inverter, has a power rating of 62.5 kW, and an apparent power rating of 66 kW. This gives you flexibility to support up to 21.2 kVAR of reactive power, at full operating power.

Consider 8 of these inverters. Based solely on the 62.5 kW rating, this might seem to imply a 500 kW transformer. If you limit the inverter unity power factor, you'd be correct in selecting a 500 kVA transformer. However, if you need reactive power support, you'd need to add up the 66 kVA, rather than the 62.5 kW, for sizing all PV infrastructure. This yields 528 kVA, which would require a 750 kVA transformer, if limited to standard kVA ratings of the transformer.

One reason you might want to do this, is if the facility is penalized for poor power factor (percentage, rather than kVAR). Consider a behind-the-meter facility, consuming a constant 900 kW and 436 kVAR reactive, i.e. 1 MVA total. Put 500kW of PV on the building, and at full capacity with unity power factor, and the net load is 400 kW real, and 436 kVAR reactive. What started as 90% power factor, has now become a 67% power factor, even though kVAR remained unchanged. You can mitigate this problem by setting the inverters to support the prevailing reactive loads, which means taking advantage of the 66 kVA rating on the inverters. In this example, the net load is now 400kW + 266 kVAR, which is now a power factor of 83%. Not quite as good as the original 90%, but still a great improvement from 67%.

 

[jaggedben]

Right, but as I said above, for transformer sizing shouldn't one start with the inverter kVA rating and not the KW rating? And in that case is ggunn's simple method not conservative enough for the worst case? Does any of the stuff about facility power factor change that?

 

[electrofelon]

Again, I'm not the engineer, but I don't see how ignoring power factor can be anything but conservative. Given that no grid-tied inverter that I know of can have a lower power factor by raising apparent power, as opposed to lowering real power.

If you want to find a counterexample, just post it on Mike Holt and someone will find one.

This is form the datasheet of a Yaskawa PVI-50:

Rated AC Real Power/Apparent Power/Output Current: 50 kW / 50 kVA / 60.2 A
Overhead Mode: Real Power/Apparent Power/Output Current 50 kW / 55 kVA / 66.2 A

But yeah, its an option, and not something that it does automatically, I assume.