How do I select CT ratios?

[GoldDigger]

This is pretty interesting
http://store.gedigitalenergy.com/faq/Documents/750_760/CTsaturation.pdf

you need to get some pretty high fault/rating ratios before you have issues

Very interesting, I am thankful for the Conclusion!

 

[Ingenieur]

Very interesting, I am thankful for the Conclusion!

if you look at 10 x rating: very accurate, little distortion
at 200(!) x rating clipped at ~0.5 pu of fault

this would indicate at 20 times no real issues

if you look at the C ratings
if 100 i is linear up to 100 V
assume 20 x rating or 100 A
modern digital relays have Z <0.1 or less (can be estimated from va and i burden spec)
so V is 10 V, well below the 100 so linear and accurate
so in this a C of 20 should be sufficient
and relay should not saturate at 20 x rating

the new relays vs old electro-mechanical type make the ct black magic less critical imho

 

[mbrooke]

This is pretty interesting
http://store.gedigitalenergy.com/faq/Documents/750_760/CTsaturation.pdf

you need to get some pretty high fault/rating ratios before you have issues

I will read up on that.

What are your experiences with CT saturation? I've been researching it a lot and keep getting various extremes in terms of what engineering docs state. Some say its a very large concern others that its relatively rare. Still others only saying that its really a concern with differential line protection.

FWIW at the 115 and 345kv level my POCO typically uses 3000/5. This is typical for 40,000ka interrupting breakers.

 

[mbrooke]

if you look at 10 x rating: very accurate, little distortion
at 200(!) x rating clipped at ~0.5 pu of fault

this would indicate at 20 times no real issues

if you look at the C ratings
if 100 i is linear up to 100 V
assume 20 x rating or 100 A
modern digital relays have Z <0.1 or less (can be estimated from va and i burden spec)
so V is 10 V, well below the 100 so linear and accurate
so in this a C of 20 should be sufficient
and relay should not saturate at 20 x rating

the new relays vs old electro-mechanical type make the ct black magic less critical imho

Since this is a safe place to ask... how exactly is the C rating defined? I see C400 always specked, but never question it.

 

[mbrooke]

And ok, correct me if my line of thinking is wrong:


With standard over current and definite time relays (50/51) CT saturation is not a concern when you get over your definite time setting as you no longer care what the waveform is afterward or the magnitude.


But with differential you must have as little saturation as possible because one CT may give different wave quantities when saturated as apposed to another CT- thus the relay may operate on what it believes to be an inzone fault- thus the CT must take into account the highest fault current.


With that said, assuming I am right on these two self created axioms, where does 21P / 21G (step distance/line impedance) stand?

 

[Ingenieur]

Since this is a safe place to ask... how exactly is the C rating defined? I see C400 always specked, but never question it.

It is equivilent to ansi va burden rating
there are conversion tables

the number eg C100
100 is the max voltage across the ct sec before it saturates
becomes nonlinear

say the ct is putting out 50 A (10 x rating) at trip setting
and the loop R including leads and relay is 1 Ohm
V = 50
rule of thumb says 2 x this, so a C100 is required
if R = 0.5 then V = 25 and C50 is chosen

iirc they call it an accuracy rating (?)
it's defined by applying sec Z at a defined current range until output becomes nonlinear ie saturated

 

[mbrooke]

It is equivilent to ansi va burden rating
there are conversion tables

the number eg C100
100 is the max voltage across the ct sec before it saturates
becomes nonlinear

say the ct is putting out 50 A (10 x rating) at trip setting
and the loop R including leads and relay is 1 Ohm
V = 50
rule of thumb says 2 x this, so a C100 is required
if R = 0.5 then V = 25 and C50 is chosen

iirc they call it an accuracy rating (?)
it's defined by applying sec Z at a defined current range until output becomes nonlinear ie saturated

So if I use an SEL587Z for my bus protection I need at least a C200 because that is about where the pickup voltage will be set?

 

[JayP99]

get a Relay Engineer to issue recommended settings

get a Relay Engineer to issue recommended settings

You really need to get an experienced engineer in relaying to give you the settings.
You will expose yourself to liability if it blows up down the road for miscoordination.

Most of the comments are giving you the information bit by bit that you asked for,
but choosing it yourself could be costly.

That breaker is probably $30K, plus any other equipment or cable or people that
could get hurt from the selection of the wrong settings.

There is upstream and downstream coordination that must be calculated and considered,
it's not just load. If there is billing grade metering then that requires a different class CT.
The C100 is a relay(protection) quality only, not accurate enough for billing. It should cover most
relay protection scenarios.

You mentioned a differential, that requires specific matching of CTS in the circuit. Bus Differential
requires all CTs to be the same ratio. Transformer Differential requires HV CTS and LV CTs to have
the same ratio as HV/LV windings of the transformer it is protecting. There is also phase compensation
that may be required if there is phase shift from HV to LV windings.

You're in too deep - back away.
If you're just looking for knowledge that's fine. Leave the detail to a PE relay guy.

 

[mbrooke]

You really need to get an experienced engineer in relaying to give you the settings.
You will expose yourself to liability if it blows up down the road for miscoordination.

Most of the comments are giving you the information bit by bit that you asked for,
but choosing it yourself could be costly.

I hear you, but my concern is around the CTs rather then the settings afterwards.

That breaker is probably $30K, plus any other equipment or cable or people that
could get hurt from the selection of the wrong settings.

Unless the breaker is continuously operated at current well above its rating, wrong settings will not damage the breaker. The real concern is short circuit current and its associated transients/DC components which are taken care of by fault studies.

There is upstream and downstream coordination that must be calculated and considered,
it's not just load. If there is billing grade metering then that requires a different class CT.
The C100 is a relay(protection) quality only, not accurate enough for billing. It should cover most
relay protection scenarios.

No billings. There will be coordination with the transformer's protection relay.

You mentioned a differential, that requires specific matching of CTS in the circuit. Bus Differential
requires all CTs to be the same ratio.

Only high Z relays. An SEL 487B has settings to factor in miss-matching CTs.

Transformer Differential requires HV CTS and LV CTs to have
the same ratio as HV/LV windings of the transformer it is protecting. There is also phase compensation
that may be required if there is phase shift from HV to LV windings.

Both of these can be automatically taken into account by an SEL-487E's settings options.

 

[Bugman1400]

I am a PE relay guy (20+ years) and there seems to be a lot of mis-information in this thread. It is called the "Art" of relay protection because there are many ways to "skin a cat". It would take way too long to get into the details of the different methods for accomplishing the same protection on the same equipment. I have not seen the FLA rating of the breaker in this thread....perhaps I missed it. It is typical to use a 600A or 1200A breaker on a feeder ckt that will only have 100 to 200 A on it. This is because feeder ckts can be swapped or load added on to it in the future. The size (ratio) of the CT typically follows the rated FLA of the breaker but, not always. Sometimes it will follow the expected load (200A). This method provides the most protection sensitivity for weak systems. For systems that have substantial fault current, the CT ratio may follow the breaker FLA because detecting fault current is not a problem in a stiff system. The goal here would be reliability of the circuit and to allow downstream protection devices enough margin to clear before affecting upstream devices. As you can see, there are a lot of factors with determining the CT ratio and protection method.
A CT does not always saturate when the secondary nominal current (5A) is exceeded. It mostly has to do with the circuit that the CT is driving. If the distance between the CT and the relay is far and the wire is skinny, the impedance that the CT "sees" may be larger than its rating. There are other elements that can affect the CT ckt impedance as well. For example, if there are additional devices that are using the same CT ckt (ie metering, transducers, etc.). Also, the terminal connections themselves (loose lugs) can add to the CT ckt impedance. This type of saturation that can occur is analogous to an audio amplifier that is trying to drive too many speakers connected to one channel. The sound becomes more distorted as you turn the volume up and up.
One thing I do want to mention about the 587Z relay is that it not like the other relays. It is not an overcurrent relay. It is a voltage relay that is set to expect CT saturation to function correctly and typically does require all the CTs to have the same ratio but, not always.

 

[mbrooke]

Max breaker currents are typically 600-800amps on the 38kv breakers and the bus is fed via 3 40/50/60MVA transformers. The 15.5kv breakers typically see between 400-600amps tops, and the fault current is much more limited: fed from the 34.5kv system and the transformers are usually between 12.5 to 20MVA.

 

[Bugman1400]

My guess is all the feeder bkrs will be 2000 or 1200A bkrs with 2000:5 or 1200:5, C800 CTs. Let me know if you need a company to set the relays and/or commission the sub.

 

[mbrooke]

My guess is all the feeder bkrs will be 2000 or 1200A bkrs with 2000:5 or 1200:5, C800 CTs. Let me know if you need a company to set the relays and/or commission the sub.

1,200amp breakers. Where did you get the C800 # from? The plan is C400, but if not, let me know.

 

[Bugman1400]

1,200amp breakers. Where did you get the C800 # from? The plan is C400, but if not, let me know.

C800 is typical for substations whether the fault current is high or not. The C400s are slightly cheaper but, will saturate sooner than a C800. That used be a concern in the days of multiple electro-mechanical relays sharing the same CT ckt. Nowadays, its mostly micro-processor relays that have very little burden and most CT ckts are dedicated to one or two protection devices. The only other concern is the burden (impedance) of the CT wires and the terminal blocks. The C400s may be perfectly adequate in your application!

 

[Electric-Light]

The 600:5 and similar ratings assume that the relay or other device connected to the low current side of the CT is designed to operate with a 5A full span current.
That is why it says 600:5 and not just 120:1.
If your device wants to see a full scale current other than 5A, then you have to do a little more work in choosing the CT.
You can use a 600:5 CT with a device that wants to see 1A when the primary current is 120A. It will work just fine, but will be overbuilt for that task.
If you want to get a 10A output to your device, you could not use any of the CTs listed because they would be seriously saturated before that point.

None of these ratings directly address fault current. Using a CT instead of a suitable shunt for thermal trip and coil for magnetic trip is an entirely different design problem.

Use two, parallel.

What thermal trip? I thought the readings from the CT was used to tell the machine that uses the 120v from the PT to crank the motor to turn off the switch? At least that's my understanding of why some PT allows output load in kVA range.

 

[Ingenieur]

2 x rated 5 A should be linear and not saturated
use 100:5 C200 as an example
loop R including relay ~ 2 Ohm (much higher than typical with a digital relay)
v = i R = 200 = i 2
i = 100
the ct is linear to this point
even with 2:1 sf 50 A out or 10 x rating should not be an issue
not sure why you would need 10 A though

 

[Fnewman]

Also, these make of relays ask for amps secondary...

So if I want my 51 to pick up at 600amps on a 1200/5 CT, I would enter 2.5 amps?

1200/5=240

Desired amps / 240= amps secondary in the settings sheet?

Yes, it is as simple as that.

 

[Bugman1400]

2 x rated 5 A should be linear and not saturated
use 100:5 C200 as an example
loop R including relay ~ 2 Ohm (much higher than typical with a digital relay)
v = i R = 200 = i 2
i = 100
the ct is linear to this point
even with 2:1 sf 50 A out or 10 x rating should not be an issue
not sure why you would need 10 A though

Ingenieur is correct here on saturation but, the constant 10A becomes a thermal issue instead of a saturation issue. Unless the CT is rated at 2.0 SF, a constant 10A will certainly lead to failure.

 

[mbrooke]

Ingenieur is correct here on saturation but, the constant 10A becomes a thermal issue instead of a saturation issue. Unless the CT is rated at 2.0 SF, a constant 10A will certainly lead to failure.

SF=Service factor?

I am getting some conflicting info outside this forum, but from what I am being told is that a 1200:5 CT should never exceed 1200amps?

Also as a general rule a C800 CT is good to limit the wiring to 1.2 ohms? (Yes/no/maybe? )

Thanks to everyone who has answered thus far, its been a wealth of info :happyyes:

 

[mbrooke]

Yes, it is as simple as that.

Comforting to read. Which means I am going to get it right for once :angel: