Ungrounded wye connected CTs

[HamburgerHelper]

Can someone explain to me what happens when you have CTs for each phases connected to the relaying with an ungrounded wye connection? The neutral was left open and I suppose for a balanced system it wouldn't matter since the neutral would only carry zero sequence current. It was discovered as part of an investigation into a CT failure I believe. To me, it looks like an ungrounded wye would screen out any zero sequence current but you would have zero sequence voltages develop across both sides of the relaying. I don't think there is an issue with saturation. Can this cause that? Can the ungrounded neutral voltage bounce around so much to cause the CT to fail?

 

[scottf]

So that I understand correctly, are you saying that there was no ground reference on the secondary winding/wiring of the CTs?

 

[cranky108]

Sounds like you are saying you are missing your ground return.

Consider if you have a little unbalance current, your circuit would act like an open CT, as the current would have no where to go.
Is that the type of damage you saw?

 

[HamburgerHelper]

Scottf,

yeah, there is no path for zero sequence to get back to the CT neutrals. The cts to me shouldn't be able to source zero sequence currents because their neutrals are connected together but not grounded.

cranky108,

The part about that that doesn't make sense to me is why would they saturate if their is no path to even send the zero sequence current. When I look at it, it looks like to me that it should just screen out all the zero sequence current and only positive and negative currents should get to the relay. The guy that looked at it closely thought the same way that you do but I am having a hard time seeing a CT setup without a return neutral acting like an open CT. To me, it shouldn't cause saturation because the zero sequence voltage will be the same everywhere due to there not being a place to source/sink zero sequence current. The currents the relay saw looked like a phase to phase fault with a small amount on the third phase. All three phases that the relay saw look like they sum to something close to zero. There didn't appear to be any saturation but the waveforms I saw were only sampled at 4 time/cycle.

 

[scottf]

CT secondary windings should ALWAYS have a ground reference, otherwise, the secondary winding can float up in voltage towards the primary winding's voltage due to capacitive coupling. Not allowing the secondary windings to have a ground reference is a big no-no in my book and certainly could lead to failure, not to mention it's a potential safety hazard. What is the voltage class and type of CT?

 

[HamburgerHelper]

I think it was 138kv. Maybe 345kv. If I understand you, the ct may have had issues prior to the event ever happening if it was due to capacitive coupling. The ct failed test but I don't know if the ratio is off or the ct excitation curve no longer matches the test data.

 

[HamburgerHelper]

Scotty,

It is grounded. The outputs at the relay are tied together and grounded but the neutral back to the cts was broken.

 

[HamburgerHelper]

Cranky,

I think I agree with you. The cts would be trying as hard as they could to push the zero sequence current into the relay circuit until they saturated. The zero sequence voltage would have increased I imagine until something failed. Maybe, the part I find different about this is that they aren't open cts in the traditional sense and you would have something like this if you had any sort of component or frequency that saw a very high impedance in your secondary circuit.

 

[stevenal]

I've a bit of experience here, since I know of an intermittent open CT neutral that caused a mis-operation. If you consider your CTs as ideal current sources, the open neutral will look exactly like an open circuit to zero sequence. The CTs are not ideal current sources, though, and are more accurately modeled with shunt admittance paths. This admittance on the CTs of the un-faulted phases is your return path for zero sequence as it circulates through the relay coils and back to the CT sensing a SLG fault. This is how an out of zone line fault can cause a bus differential relay to trip showing a target on an un-faulted phase.

 

[HamburgerHelper]

How it looked to me is that zero sequence current can't be generated in the CT circuits. The ct tries to output as hard as it can the zero sequence current but all or close to all just ends up as excitation current. The waveform that was seen by the relay showed the current being carried on two phases, 180 degrees out of phase. One of the cts failed test and I don't know why yet but to me even if a ct hadn't failed, a SLG fault with the zero sequence component screened out should look like a phase to phase fault. Sort of like how a SLG fault looks on the wye side of a delta-wye transformer from the delta side. Anyways, that is how it made sense to me.

 

[scottf]

I'm having trouble, based on your description, understanding how these CTs are not open-circuited on the secondary. It doesn't doesn't sound to me like there is a current return path. Maybe I'm missing something.

 

[stevenal]

Link

I've linked a sketch. To the right of each ideal CT is the associated shunt excitation branch, which is carrying the return current. Note that the current through the relay on the un-faulted phases is 180 out of phase with the current through the faulted phase of the relay. I imagine saturation would be unequal so that current wouldn't divide equally through the un-faulted relay coils. The result would be as described: equal and opposite current seen at the relay.

 

[HamburgerHelper]

Stvenal,

Your sketch of the setup is correct as to my case.

I don't agree with your assement. If you move the excitation branch to the high side of the ct, you have two paths that the fault current has to go through. It has to flow straight through the excitation branch or through the ideal transformer to the relay and back through the transformer to the highside. For negative and positive sequence current, you would have to be extremely saturated for the least impedance path to be the excitation branch. For the zero sequence current, since the neutral is broken the excitation branch is the easiest path. The relays are saturating but the component split is why i beleive you have two phases carrying current 180 degrees apart. Also, the zero sequence currents are causing the saturation and if there were slight differences in the degree of saturation i don't think they would be much due to how the saturation curves level out after the kneepoint.

 

[stevenal]

So why would it be more proper to move the excitation branch to the primary side? I believe the most proper model includes these shunt paths on both sides, but the high side path was not necessary for my analysis. I'm unsure how sequence components help in the analysis. It is true the SLG fault will contain all three components in the faulted phase, but in the other two phases the components all sum to zero (assuming no load). CTs can only respond to the summation.

I see extreme saturation as highly probable, since on any un-faulted phase you have current on the secondary with no corresponding current on the primary (essentially an open circuit primary). As soon as one CT saturates, that will be the low impedance path that will hog the current.

 

[stevenal]

Link

If you prefer to see the magnetizing branch on the primary side, here is another look. I see no change in drawing it this way. The magnetizing path on the faulted phase remains high impedance since matching secondary current has a path to follow. One of the un-faulted phases will be the one to saturate.

 

[HamburgerHelper]

To me it is easier to visualize the excitation on the high side but you can put it on the low side as well. But, the point that I get from it is that everything thing has to either pass through as excitation branch or go through the relays. Looking at it using sequence components I beleive helps becuase it shows that the ungrounded ct voltage will swing out to whatever develops across the ct and that all three cts will be pushed into saturation because of the shared neutral being allowed to swing out to something close to the open ct voltage. IMHO, the relays will only see zero sequence current through capacitive grounding of the circuit or when the insulation breaks down someplace. That is why I beleive all or almost all the the zero sequence current passes through the excitation branch and never makes it to the relay. Basically, it screens out that component.

All the cts I beleive are saturated close to equally due to all three phases being connected with to the grounded relay circuit and having a common neutral that is swinging out. So, I don't think it has anything to do with one relay saturating before or more than the others.

I suppose it depends on the ct but when I look at the C800 curves, the slope levels out but it still has to go pretty far before it is below the relays burden, which with electronic relays and cables I imagine might be just a few ohms. So, to me it makes sense why the positive and negative sequence currents go to the relay and not through the excitation branch.

I suppose I can model it in ATP-EMTP sometime to prove it this week.

 

[stevenal]

everything thing has to either pass through as excitation branch or go through the relays

Look at the diagrams. Current through the excitation branches of the un-faulted phase CTs passes through the relay. Even a matched set of CTs will not saturate together because there is no such thing as perfectly matched CTs, and one will always have a different amount of residual magnetism. Once one has saturated, it presents a short circuit to the CT sensing the fault and the remaining un-faulted CT has no voltage across it to cause it to continue into saturation. This is similar to high impedance bus differential where one CT is assumed to fully saturate allowing the other parallel CTs to operate in the linear region during a through fault.

 

[HamburgerHelper]

Stevenal,

I modeled it and your right. The split without saturation is 1.0 in one phase and 0.5 pu returned on the other two phases. To have it go out and return on two of the phases has to be due to uneven saturation like you said.

 

[stevenal]

I'd be interested what the MVPs have to say about this. Waross, Davidbeach, ScottyUK, others? Any experience here?

 

[scottf]

I discussed this with one of our CT design engineers and he agreed with me that you likely had some degree of capacitive coupling happening between the primary and secondary windings of the CT which could have done some initial damage to the insulation. During a fault, the CT likely also saw very high voltage across the secondary as the voltage present is going to be close to the open-circuit voltage of the CT. Assuming these are protection rated cores, then that voltage could be in the 10's of kVp range.

You mentioned that the "CT failed test", so I'm assuming it wasn't a catastrophic HV insulation failure, but some kind of insulation failure on the secondary, which makes sense. I would assume that the insulation resistance from seoondary-to-ground would show lower than normal readings and/or the excitation test would show lower than expected saturation voltage. Failure modes could be turn-to-turn short on the secondary, secondary short to the core, or secondary short to ground (either high or lower resistance).