CT accuracy 3

[electrolitic]

10P20 CT it means a protection CT with composite error of +/- 10% and that error will be maintained up to 20 times the rated primary current of the CT. Is it correct? Is it constant?
More and more current measurements/monitoring using the modern protection relays instead of dedicated instruments + dedicated measurement CTs are being used.
Are the measured values, even close to the rated primary current of the CT, subjected to the +/- 10% error?
Thanks

 

[LionelHutz]

I doubt it's constant. You can likely get a relaying CT that has a 1% error up to the CT rated current and still has the 10P20 rating.

 

[waross]

The rating describes the accuracy of the CT under test conditions. One of the conditions is full rated burden. As the burden is reduced the error is reduced.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[electrolitic]

Sirs
I hope so. Is it possible the evaluation of this error at normal operation condition?
Is there a standard or a paper to be consulted?
Thanks

 

[oldfieldguy]

Typical protection CT's that I've tested fall well within 1% accuracy during low current tests. I expect this accuracy up to nameplate rating, especially with the low burdens associated with most indoor installations.

In installations where the client required both metering apparatus and microprocessor-based relays with extensive metering functions, I've not noticed appreciable difference in measured data over a period of time. However, I do recognize that this is anecdotal. However, the accuracy of microprocessor-based equipment is causing a lot of people to rethink the traditional protection ves. metering line of demarcation in non-revenue venues.

old field guy

 

[electrolitic]

Thank you all
I will continue to use protection CTs (5P20 - 15VA instead of 10P20), Multilin / SEPAM series protection relays and no measuring CTs. The problem is how to prove to the client the guaranted accuracy. To find a safe and simple way to get additional measures (to be compared) in a 15kV class Swgr is not so easy. I hope I can find related papers or articles before the comissioning.
A Happy New Year.

 

[Runsor]

An old handbook states that a 5P CT at rated current and rated burden has less error than 1 % and 1 one degree. For a 10P CT the amplitude error is less than 3 % and angle error is not specified. I guess these are IEC specs. In practice the errors are smaller at least if the ratio is big and the burden is small.

Happy New Year

 

[stevenal]

We commonly spec CTs that are dual rated for both protection and metering.

 

[ZapSib]

Limits at nominal current for 5P/10P are given in IEC 60044-1. this one is being replaced, but I assume it is still given in the new norm.( sorry, I don't remenber numbering).

 

[electrolitic]

Stevenal
I always did it but the modern protection relays feature a lot of measurement/monitoring related to current measures and there are just inputs for protection CTs. My question was derived from this reality. As oldfieldguy said above: "However, the accuracy of microprocessor-based equipment is causing a lot of people to rethink the traditional protection ves. metering line of demarcation in non-revenue venues."
Anyway, thanks.

 

[dpc]

Keep in mind that protection CTs and metering CTs have very differing jobs to do. For protection, we want CTs that have good fidelity at fault current levels which can many times the normal load current. A compromise CT rating that is extremely accurate at normal load currents but saturates for faults is a dangerous compromise. New digital relays have very low burdens, but CT internal resistance and CT lead resistance has not changed for the past 100 years or so. A wimpy CT can still easily saturate for high magnitude faults, regardless of the relay burden. And for faults with high X/R ratios, the dc offset must also be considered. Our tendency for protection is to use higher ratio CTs to reduce saturation concerns, since the resulting lower "tap" settings in the relay do not introduce much burden penalty, unlike old electromechanical relays.

Also, as you probably know, for a multi-ratio CT, the stated accuracy only applies across the full ratio. If it is tapped down, the accuracy goes down.

 

[stevenal]

Sounds like I was misunderstood. The CT you attach to that multifunction box can carry ANSI ratings for both protection and relaying. For example: "1200:5 multi-ratio C400 relay accuracy class bushing current transformer capable of metering applications with an accuracy class of 0.3B-1.8 on the full winding."

 

[ScottyUK]

stevenal,

Have you had any problems with meters being damaged by through-faults? I'm no metering expert but I always understood that protecting the meter was the reason why saturation not too far above nominal current was desirable, which is obviously a conflicting requirement with a protection CT where reasonable accuracy is required when fault current flows in the primary. Or do you mean that the dual-rated CTs used with protection relays having sufficient accuracy to be used for measurement applications but which can handle the high secondary currents during faults? Apologies if I'm just not reading your post correctly.

 

[stevenal]

ScottyUK,
I don't believe ANSI has anything like the IEC "instrument security factor." My assumption is that ANSI meters are as rugged as relays and are not damaged by fault current. We have not had an issue with meters on protection class CTs or with multifunction instruments on dual rated CTs.

 

[electrolitic]

Sometimes my simple question becomes complicated. Surely is caused by my bad english.
The question it was: A feeder protection relay, GE Multilin 350, as an example, performs IOC, TOC, a lot of protections based on current measurement. Also provides metering of power, phase currents, thermal model, oscilography etc and displays some of them.
This relay has just 03 phase inputs from CTs and one more for ZCT. Of course these CTs shall be protection CTs, as for example, 10P20 (+/- 10% up to 20 In). I asked: Which is the accuracy of these CTs when the current is close to the nominal value? Can I trust on the normal running results or I need to provide a warning label:"Do not consider some readings of current. The composite error is +/- 10%". Of course the answer would be no but I was asking for a number.
Runsor answered that 5P20 has less than 1% and 01 degree. 10P20 less than 3%.
For me, a correct specified protection CT 5PXX will be enough and where accuracy is not critical I will not use measurement CTs and dedicated meters.
Thanks

 

[davidbeach]

Essentially a protection CT is accurate enough to use to provide the metering used to run the system, but if money is changing hands then a metering CT should probably be used.

 

[LionelHutz]

Almost any relaying class CT should be accurate enough up to rated currents that it can provide decent metering. FWIW, I can easily find ANSI rated CT's where both the relay class and metering class data is given making it easy to pick a CT which will accurately meter and provide protection. Are there no IEC rated CT's with both ratings?

 

[stevenal]

I thought I said that twice already.

Dual rated IEC CTs might be tough to find, since I understand metering CTs are required to saturate under fault conditions in order to limit current to the connected meter.

 

[Skogsgurra]

Stevenal
I think that is a valid concern. I once started a fire in 11 kV switchgear when I caused a 1600 kVA transformer to saturate (malfunctioning softstarter drew DC from transformer secondary) and draw a lot more current than anyone thought possible. With a DC saturated core, around half of the u[sub]k[/sub] was lost. If that CT also had saturated, we wouldnt have had the fire. Fire started in an old Ferraris meter for internal use, no energy vending, in series with protective relays.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[LeifEricson]

A bit late to the discussion, but Runsor is right on the accuracy limits of 5P and 10P classes at Ip. Also, instrument security factor is an optional specification for IEC measuring CTs so it shouldn't technically be a problem to get a dual rated protection/measuring CT (although it may be a practical issue depending on the specifics of the CTs you use).