Solar power, inverters, and current transformer saturation.

[narkshark]

Hello everyone,

My question is something that a few customers have asked me over the last two months, and possibly many more times on the technical support team.

My understanding is that current transformers (CT) are recommended to have at least 10% saturation on them. So, a 200A CT should have at least 20A of current or else the readings may not be accurate and/or may just read zero.

What are customers to do when their solar power inverter is producing low amounts of power at dawn, dusk or during adverse weather conditions? Do we have a recommended procedure for that circumstance? Should they write off those times as just not having reliable readings? Are their specialised CTs available?

Thank you for any input.

[Hal_Etheridge · Accepted solution]

Check the data sheets for the current CTs.  You may find the limit is much lower than you think if they are measurement class CTs (instead of relay class).  I know that the ION meters do handle and measure some amount above the nominal rating so maybe you can take advantage of that by using under rated CTs to get better accuracy in the low end - it really depends on what dynamic range you require in the actual application and assuming that the CTs still stay reasonably linear above their rating (which may or may not be the case depending on the CT).

There are CTs that are rated to much lower current levels before they get inaccurate but that will probably cost more.  My suggestion is to critically examine the actual requirements for your application and compare them against the current CTs to see if they match.

Another possible suggestion is to change the compensation curve at the lower range.  Maybe a modified version of the transformer loss compensation framework would work in this case (assuming you are using an ION device where this is programmable).

This is an interesting problem.  Please post back with your solution.

[Yut]

Hello Conway,

     I have some installation in 50MW solar farm using ION8800 which has starting current of 1 mA no problem when cloudy. One thing to do, use just right size of CT and recommend the customer to use our high end meters such as ION8800 ION8650 or the ION7x50 are fine.

K. Yut

[narkshark]

When you say a starting current of 1mA, does that take into account the accuraccy of the CT itself that is attached?

Does ION7x50 include the ION7350? I happened to see that come up a lot.

[jimbradley]

All the utility standards (ANSI, IEC, Measurement Canada, etc) have a starting current spec where the energy meter must register something within a time period.  Not really an accuracy spec but more a holdover from mechanical meters to ensure the rotor wasn't stuck.

On the accuracy side there is a spec for load current ranging from a x% of rated load to full load (or higher).  Often there is a step in the accuracy where the lighter load is a "looser" spec, eg. 1-10% of rated load is 0.5% and 10-120% of rated load is 0.2 % ( I made these numbers up).

All these specs are meter only, i.e. the test currents are injected directly into the device under test.  The CTs accuracy are a separate issue.  As we all remember form 1st years physic errors in the case will be summed, i.e. Total errors = Meter Error + CT Error.

I'm pretty sure that 7x50 refers to the 7550 and 7650 as these share a hardware platform.  The 7350 would typically be referred to as apart of the 73X0 family, also a common hardware platform.  The 73x0 family will not meet the same specs as the 7x50 family.

Jim

[Grant_IPACS]

Just thinking out loud...

At really low production levels, if there is not enough current to cause output in the CT, wouldn't the meter then register 0Wh for the period?

And if it does manage to get an output from the CT, then some Wh would be registered?

If the question is accuracy, isn't that just a factor of the CT & meter class / accuracy?

Or am I missing something?

Grant