CT saturation VS Numerical relays

[sakaran51]

I have come across an article by one of the relay manufacturer that their numerical relay is unaffected by the CT saturation for any through fault.It can detect the saturation and prevent the malfunctioning of the relay,so they claim. Any comments?
Moreover,are the special type PS/X Class CTs really essential for the numerical relays?
I have seen ordinary protection CTs like 5P20 being used for numerical relays for Generator differential protection.

 

[slavag]

Yep, 5Pxx used for 87L, 87B, 87T, 87G, 87M, without any problem.
Class X. PS, TPS must be used for HiZ only.

 

[lume7006]

Hello,

The issue is related to several factors affecting CT performance:

a) the burden between CT terminals and relay. This means you have to know the values of leads resistance, CT resistance and relay burden.
They all together should be lower than CT burden nameplate.

b) CT's have to be for protection purposes, then they should be P, PS, TPS and if they were based on ANSI C-class.
The values of the burden have to be according to what you are expecting.

c) it is important to do some calculations based on IEC or ANSI standard in order to be sure CT's characteristics are good enough to each application.

 

[LiteYear]

I have come across an article by one of the relay manufacturer that their numerical relay is unaffected by the CT saturation for any through fault.It can detect the saturation and prevent the malfunctioning of the relay,so they claim. Any comments?

What is their method? The problem with "detecting" saturation is that it is largely indistinguishable from a real current. Without having some sort of expectation of the "right" result, it's hard to know whether saturation is occurring. The only method I'm aware of to overcome this is to inject a known signal on the secondary until saturation is reached (as detected by the secondary voltage no longer following the injected secondary current). The change in the threshold of saturation represents the influence of the primary current, so by using opposing primary and secondary currents and careful subtraction of the injected current, the primary current can be determined even in the presence of saturation. It's a tricky business though, and quite sensitive to the particular characteristics of the CT. As such, I'd only expect to see this technique if the CT (or a small range of CTs) was explicitly specified for the relay.

 

[davidbeach]

Modern differential relays can determine internal vs. external fault in the first 1/4 - 3/8 of a cycle. Saturation begins soon thereafter. Knowing where the fault is then allows the relay to operate, for internal, or go very secure and wait out the saturation for external. The fault on the boundary of the zone that starts as external and then evolves to internal may see delayed clearing.

 

[sakaran51]

Yes davidbeach,
That is how their numerical
relay functions, they say.

 

[scottf]

Liteyear-

The secondary current waveform of a saturated CT is very "noisy". I don't think it's that difficult for a relay to determine if the CT is in deep saturation.

I've always heard the relay manufacturers talk about being able to "filter" the saturated current waveform, but I have no idea how that do that.

 

[LiteYear]

Interesting scottf. Any further info on that? I thought a saturated waveform mostly just results in attenuated peaks. I've always found this paper quite authorative:

http://apps.geindustrial.com/publibrary/checkout/White Papers|GET-8501|generic

The lightly saturated waveforms are barely indistinguishable from a sinusoid, while the heavily saturated waveforms just look similar to 6 pulse waveforms to me. In both cases, far from what I'd consider noisy. Perhaps if you knew that your primary current was supposed to be a pure mains frequency, then you could detect it. Doesn't seem reliable though.

Interestingly enough, there's a paper here that claims detection based on statistical comparison with the previous quarter cycle or so. Not sure if this is successful in practice.

http://ieeexplore.ieee.org/iel5/61/5910683/05771582.pdf?arnumber=5771582