Over Current and Differential CT ration

[Electriman]

Good afternoon,

I was wondering if any one knows the proper CT ration for over current protection and differential protection.

I am thinking that over current CT ration must be selected such that it is around 20% greater than nominal load.

But I have no idea about diffrential CTs.

What does NEC recommend?

The application can be transformer protection and/or feeder protection.

Thank you in advance

 

[Bugman1400]

Typically, for bus differential protection the CT ratios are sized to the full load ampacity of the bus ie (2000/5) where you have a 2000A bus. The feeder CT ratios for the same bus will typically be sized for the full load of the cable or the max expected load of the feeder circuit plus some growth margin (125% to 200%). Obviously, there are many other things to consider. For example, its possible to have a low load bus (500A) but, the CTs may be oversized because the fault availability is high (ie 20,000 amps). So, if you had a CT ratio of 500/5 (100/1) then a 20,000 amp fault could cause 200 amp of secondary current on the CT cables which is a bit excessive and considered dangerous. The higher fault current could also lead to excessive saturation of the 500/5 CTs. Another thing to consider is the class rating and service factor rating of the CTs. The service factor for CTs is similar to the service factor for motors.

 

[Electriman]

Typically, for bus differential protection the CT ratios are sized to the full load ampacity of the bus ie (2000/5) where you have a 2000A bus. The feeder CT ratios for the same bus will typically be sized for the full load of the cable or the max expected load of the feeder circuit plus some growth margin (125% to 200%). Obviously, there are many other things to consider. For example, its possible to have a low load bus (500A) but, the CTs may be oversized because the fault availability is high (ie 20,000 amps). So, if you had a CT ratio of 500/5 (100/1) then a 20,000 amp fault could cause 200 amp of secondary current on the CT cables which is a bit excessive and considered dangerous. The higher fault current could also lead to excessive saturation of the 500/5 CTs. Another thing to consider is the class rating and service factor rating of the CTs. The service factor for CTs is similar to the service factor for motors.

Thanks,

Does it matter if it is used for diffrential protection or over current protection?

For example lets assume the application is a transformer protection and I have two sets of CTs on the primary one over current and one differential they are MR:600/5 and my load is 175 A at 12.47 kv. The CT class is C200.

 

[Bugman1400]

Thanks,

Does it matter if it is used for diffrential protection or over current protection?

For example lets assume the application is a transformer protection and I have two sets of CTs on the primary one over current and one differential they are MR:600/5 and my load is 175 A at 12.47 kv. The CT class is C200.

Typically, your overcurrent will be set at 150%-200% of the full load capability of the xfmr. If its 175A, then I would expect the pickup for the OC to be around 263A (150%) or 2.2 times the CT ratio of (600/5). I think your okay with that CT ratio. The C200 class of the CT is typical for distribution/industrial type applications and would be adequate to use for xfmr differential application as well. The exception is if this xfmr is at a generating plant where the fault currents may be much higher than typical and you may need a higher than C200 class CT.

 

[paulengr]

Another consideration is not to oversize the CTs because most relay taps are limited to 1-1.5 on the low end.

I also suggest using microprocessor relays always. They are much more reliable, no drift (it either works or it doesn't), virtually no burden on the CT, usually do some data logging, much more flexible, and usually you combine all your relay functions including trip/close buttons into one unit. Much cheaper. As an example for around $1,000 you can get 50, 51, 50/51N, 87, 59G, 29, over/u derfrequency, data logging, waveform level fault recording, multiple sets of trip curves (think maintenance switch), and various communication options. The biggest complaint is complexity but you can get ABB circuitshields and Littlefuse models that are even simpler than electromechanicals, up to the ultimate SEL and Basket relays that include so much capability I dumb them down from the defaults to avoid making maintenance too complicated. Manufacturer recommended test frequencies are around 6-8 years but watch for how you do it. Testing is a little different since its really just testing IO functionality and running self checks on CPU and memory instead of checking for calibration drift.