statement about false tripping of ground relay (window CT) 2

[electricpete]

IEEE242 Buff Book states the following regarding window (core balance) CT feeding overcurrent relay for ground protection:

"The flux produced in the current transformer core is proportional to the ground-fault current, and a proportional current flows in the relay circuit. Relays connected to core-balance current transformers can be made quite sensitive. However, care is necessary to prevent false tripping from unbalanced inrush currents that may saturate the current transformer core, or through faults not involving ground. If only phase conductors are enclosed and neutral current is not zero, the transformed current will be proportional to the load zero-sequence or neutral current. Systems with grounded conductors, such as cable shielding, should have the current transformer surround only the phase conductors and not the grounded conductor."

I am only interested in the bolded statement (I only included the rest for context). What does it mean?

To my way of thinking, any saturation of window CT means the output is lower and less likely to trip. I don't see what scenario they are talking about, unless perhaps they are talking about unbalanced inrush into a circuit that includes capacitance to ground.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[davidbeach]

Good question. What you have bolded is a description of sources of false residual than can cause false operation of residually connected ground elements. But how those things could impact a core balance CT is a good question.

 

[RalphChristie]

Pete, just clarify:

Core Balance (Ring) CTs been use as:

Residual Ground Fault Protection. (use three CTs, each CT around a phase conductor)
Sensitive Ground Fault Protection. (use a single CT around the three conductors)

Both methods use a residual current principle, but I assume the statement refers to Sensitive Ground Fault Protection.

Although the currents may balance, the magnetic fields may not cancel out resulting in incorrect operation. I want to add another possibility for this to happen and that is if the conductors are not spaced equally through the window of the CT.

[red]Failure seldom stops us, it is the fear for failure that stops us - Jack Lemmon[/red]

Make the best use of Eng-Tips.com
Read the Site Policies at FAQ731-376

 

[electricpete]

Thanks guys.

There is a separate section on residual-connected CT's. This section I'm sure appliles to the situation where all three phases go through a single CT.

I have heard about the effect of the positioning of the leads and I have also seen a clamp-on ammeter reading change as I move it. As mentioned, it would not have anything to do with my original question about the the bolded sentence (saturation). But I wanted to followup because it never made sense to me.

Integral H dot dl = Ienclosed
If we look at the path around the CT core, Ienclosed does not change as we move the conductor, so why does that affect the CT output?

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[PWR]

I believe you have identified an error in the standard. Saturation of the core for this type CT installation is not an explanation for nuisance tripping.

 

[dpc]

As far as the position of the conductors within the core, it can definitely have some effect, as can the angle of the core wrt to the conductors.

I think the reason is that there is always some leakage flux from the conductors that is not fully linked with the core. Every transformer has some leakage flux. As long as the leakage flux from each conductor is approximately the same, the flux in the core remains low. But if one conductor is up against the side of the core while the other two are in the center, it would seem, intuitively anyway, that the flux linkages could be different. But that's basically a guess.

 

[electricpete]

I like the idea of applying the concept of leakage reactance.

Leakage reactance is flux linked by one winding but not the other. Therefore it creates a voltage drop in one winding but not the other.

If I view the airgap between primary conductor and secondary-conductor/core as primary leakage reactance, it creates a voltage drop in the primary circuit. That is pretty much irrelevant of the CT because the primary circuit determines the current almost independently of the CT (CT’s impedance as seen in primary circuit is very low).

So to make a difference, I have to view it as secondary leakage reactance. Flux from the secondary winding that is not linked by the primary. That doesn’t initially sound right but I have to think about it some more. Any comments are welcome.

Another thing I was thinking about: If I viewed the clamp-on or window CT secondary as a circular coil of wires (instead of a circular coil with toroidal windings), then the explanation would be a lot simpler. I can simply say that secondary voltage is dPhi/dt. Since field drops as 1/R with distance from the primary conductor, the highest Phi (and highest dPhi/dt) occur if I place the conductor directly in the center. Unfortunately, that simplified model doesn’t seem to apply here.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

Correction:

"If I viewed the clamp-on or window CT secondary as a circular coil of wires (instead of a circular core with toroidal windings), then the explanation would be a lot simpler"

Maybe I didn't use the terms circular and toroidal exactly right. I think you know what I mean?

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[dpc]

I'm still digesting it, but this may be of some interest:

http://www.powl.com/main/Uploadpdf/00605755_GndPaper_wCopyright.pdf

 

[fsmyth]

Well, I grew up believing that there was nothing critical
about the position of a wire inside an enclosed coil.
I always assumed (yeah, I know) that if the perpendicular
magnetic field surrounding a wire had weaker coupling to
one side of a coil, it would be offset by a stronger
coupling to the other side.
I would very much like to know if this is incorrect.

I have constructed, bought, and removed current transformers
from old equipment, and have never noticed a difference when
the primary wire position was not centered. I will add the
caveat that almost all of the time, I calibrated the receiver end and/or burden myself.

I have constructed a few items of test gear-- wattmeters,
current meters, 3-phase balancers, and such, sometimes using
store-bought CT's, and sometimes home-made.
Here are some of the references I used:

http://www.ensico.si/files/wip1_e.pd

http://www.simpsonelectric.com/pdf/manuals/186.pdf

energy.tycoelectronics.com/getDocument.aspx?PRD_ID=448

but if any of them (or the others) referenced a primary
loop orientation, I missed it.


I have built and used a Rowgoski coil (instrumentation):

http://www.pemuk.com/pdf/ias_2000_pem.pdf

The only time it refers to the wire position is in relation
to the end of the coax (this is, simply put, a piece of
coax cable looped back to itself, anchored, and
calibrated).

<als>

 

[dpc]

The instructions for the various ground sensor CTs that I have seen always advise keeping the wires near the center of the opening and perpendicular. There is generally a minimum clearance of 1" or so recommended.

In general, the toroid is very forgiving of primary conductor placement, I think. But with the ground relay, there is extremely high current relative to the rating of the CT, so even a slight difference in flux linkages could make enough difference to be a concern, especially since the relays are generally set with low pickups.

The asymmetrical current is another interesting issue for transformer energization and motor starting because it will not be the same in all three phases. But this wouldn't have anything to do with the geometry of the CT and primary conductors, I don't think.

 

[electricpete]

"The asymmetrical current is another interesting issue for transformer energization and motor starting because it will not be the same in all three phases."

But currents will still sum to 0, if we model the motor as a linear L/R system (will not sum to 0 if saturation causes harmonics). I can provide the math proof, if anyone is interested. Although I'm not sure if you were making any comment about summing to 0.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[fsmyth]

dpc, I cannot see how the current magnitude can have
any bearing on wire orientation as long as it is within
the CT's B/H curve (i.e., out of saturation, and in some
predictable region).

If orientation was at all important, it would seem that
the manufacturers would ship or offer some sort of wire
shims or spacers. They may exist, but I have never seen
any such. Also, in all the utility service drops I have
ever seen, it looked more like the installers were more
concerned with ease of installation (using mortar joints
in brick, beams on metal buildings, etc. than attempting
to center the conductor(s) in the CT. On most, some
feed conductors are centered, most are against one side
of the CT, and quite a few are at an angle to the CT.

I just did an un-scientific, un-calibrated experiment;
using a CT removed from a Best 5kVA UPS (no part number,
1 5/8" inside by 2 1/4"x1/2", a 1k burden, and a scope,
there was NO difference while moving the wire from one
side to the center to any other point inside the hole,
or at any angle relative to the coil. No particular
reason for using that one, other than it was handy.

Looks like its gonna take some serious persuasion to
change this old fart's mind. But it /can/ be changed.
<als>

 

[electricpete]

electricpete wrote:
"But currents will still sum to 0, if we model the motor as a linear L/R system (will not sum to 0 if saturation causes harmonics)."

That parenthetical statement was pretty silly. If it's an ungrounded wye, it has to sum to 0 no matter what (unless there is some path to ground)

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[RalphChristie]

Ok, my two cents....

We use this kind of scheme a lot on our mining equipment. Most of the systems are 3.3kV or 6.6kV high resistance grounded, (limits the ground fault-current to 25A) but there are also some on certain 400V solidly grounded systems. In the few years that I have done testing, faultfinding, etc on these, I have seen:

Conductors not spaced equally through the window do sometimes create nuisance trips. (More on the higher voltages)
Unbalanced currents during starting, and especially on big motors, also give nuisance trips. This is true on all voltage levels.
I do believe through faults will have the same effect, but can't confirm it. (has not seen it due to the fact that we have a rather long trip-delay)
I have seen sensitive earth fault operation when only one conductor (not involving ground) was alive due to a broken OCB-mechanism.

Pete:
A few comments in my opinion on your bolded statement:

I do not think CT-saturation will create a nuisance trip, rather slower operation or no operation at all. During saturation there will be no, or a lower output, from the CT.

Operations during such conditions are true. I have witnessed some.

I would rather believe operations during such conditions are due to unbalanced magnetic fields and not necessarily current.

[red]Failure seldom stops us, it is the fear for failure that stops us - Jack Lemmon[/red]

Make the best use of Eng-Tips.com
Read the Site Policies at FAQ731-376

 

[dpc]

Yep, I've seen all kinds of CT installations, including CTs just supported by the cable itself. Usually they work OK.

But I've also had cases where nuisance GFR tripping was cured by simply getting the conductors off the side wall of the CT and into the center of the opening. This is consistent with the manufacturer's instructions. How this squares with Maxwell's equations, I'm not sure.

Pete: I don't know for sure what I was talking about either. Obviously the net current will still be zero, even with dc offset. I was probably thinking about the effect of the dc currrent on the CT, but that would probably not explain nuisance tripping.