Zero Sequence current during XFMR inrush? 1

[JBinCA]

Greetings All,

I'd like to solicit opinions on the following question - Should there be any significant zero sequence (ground) current during the energization of a XFMR?

The event which gives rise to my question is a ground fault trip that happened when the XFMR was energized with no appreciable load. The XFMR is D-Y, 480/120-208, 300 kVA (364 FLA). It is fed by an ICCB (Insul. case BKR) with a 500 amp Long Time Pick Up (LTPU) and a 360 amp Ground Fault Pick Up (GFPU). The breaker has a microprocessor based trip unit which recorded a GF current of 448 amps! There is no other load on the ICCB. There is not neutral current sensor on this feeder as it feeds only the Delta connected XFMR.

I know that XFMR inrushes vary widely in magnitude based upong residual flux, the point in the sine wave when energized, etc. I also know that the inrush currents have a strong 2nd harmonic and tend to have a significant DC component that decays in a second or so. I have seen multiple recordings that show the DC offset balanced between a couple of phases (one offsets positive while the other ofsets negative), so it seems to me that the resultant vector sum off all currents would still be 0. I'm also aware the the DC current can lead toward CT saturation, and that this can be a problem with sensitive protections. It doesn't seem to me that a 360 Amp GFPU on a 364 amp FLA XFMR is particularly sensitive, though.

Any thoughts would be appreciated.

Best Regards,

JB

 

[davidbeach]

Yes, there can be a considerable amount of zero sequence current in a transformer inrush. Add to that the saturation of the very small CTs built into the breaker and you have more zero sequence being seen by the trip unit.

 

[cranky108]

I would think the zero sequence would depend on the configuration. And with a Delta-wye transformer, you should see no zero sequence on the primary side. But if your energizing a wye-wye transformer I can agree, and mostly if the breaker poles are not well alighned.

I often apply a low set ground relay on the delta side of a transformer just to quickly pickup on winding to ground faults.

The other thing to consiter is that inrush is not just something that happens when first energizing a transformer, but can also happen after a voltage dip.

 

[JBinCA]

Davidbeach - Thanks for the response. Are the CTs in ICCBs not rated similar to protection class CTs? I can't remember reading it specifically, but I would think that they would be rated such that saturation would not occur until after 20 pu current.

As to the general question, I understand how a trip unit's residual current measurement will "see" zero sequence current incorrectly when one or more CT's saturate. But, can you explain what the source of the actual zero sequence current in the Delta windings would be?

Cranky108 - Thanks for your response as well. Good point aboutthe inrush after voltage sags. I am operating under the assumption that those would general be of a lesser magnitude than inrush upon energization when all other factors are equal.

 

[dpc]

Regarding CTs in ICCBs - these are not hardly even considered CTs - they are current sensors specifically designed for use with the trip unit in the breaker. I'm not aware of any standards that apply to these - other than the UL requirements for performance of molded case circuit breakers, which is what the ICCB is considered to be in the US.

There could certainly be some sampling error or artifacts in the trip unit's A/D conversion or numerical processing of the data. If possible you may want to use another monitoring device to measure the inrush current more accurately.

And dc current will saturate any CT quickly even at low current levels.

You did not mention the time delay setting on the ground fault trip. You may need to increase this if you haven't already done so.

 

[JBinCA]

dpc,

Thanks again for the reply. I have not gotten accurate data on the GFD settings yet. I have been assuming that increasing the delay band is the right way to correct the problem.

Regards,

JB

 

[dpc]

It could screw up your coordination with the upstream breakers, but that may be something you have to live with on this application.

 

[davidbeach]

Find a plot of transformer inrush current and look at the wave forms to visually add the three phase currents, they don't sum to zero. The zero sequence is large transient currents flowing through the excitation branch, a circuit that is almost alway omitted with drawing transformer circuit diagrams, and other than during inrush that omission is a very good approximation. During inrush you need a more complete circuit model of your transformer, there is more there than just the delta connected power windings.

 

[jghrist]

I'll have to respectfully disagree with you, David. The attached Excel file shows the energization of a 100 kV, 30 MVA transformer. Peak inrush is 944A. Peak residual IG is very low (24A). The three phase currents do add to essentially zero. I think there is no zero-sequence in the excitation branch because of a lack of zero-sequence flux.

 

[davidbeach]

jghrist, that one doesn't look too bad, except that the "zero-sequence" is all DC. I believe I've seen far worse, but that is just from memory and I don't know where to find all the inrush plots I've seen. Maybe what I'm remembering is what the CTs see rather than what the transformer does. Run that through some cheap (relatively) "current sensors" on a low voltage breaker (C20 if you're lucky) and see what you get. All that second harmonic is really hard on a easily saturated sensor and you will introduce a fair amount of false residual current.

 

[Bronzeado]

Hi folks,

In general, inrush currents is quite different for each transformer energization, as this phonomenon depends on various conditions (system impedance, residual flux, instants of closing braker contacts, voltage magnitude, etc.)which can not easily be repeated. So, zero sequence current (and so zero sequence flux) may or may not occurs.

I myself have experienced various transformer trips by neutral overcurrent relays during energization (with no fault). Sometimes, these trips happen not only in transformers being energized but also in transformers already in operation due sympathetic interaction.

Regards,

H. Bronzeado

 

[slavag]

Hello.
I think we need separate sevral things.
From protection point of view.
If ground/earth fault protection connected by Holmogreen connection and operated, problem isn't ZS component, it's unbalance current : reasult of partial saturation of CT, etc. Low current and time setting.
From my point of view, sum of all 3 phases isn't ZS current, it's unbalance current ( sorry for term).
Please put ZS toroid on all of 3-pf and energaized XFR, any problem.
I think ZS current possible only in case of ground/earth fault.
Regards.
Slava

 

[Bronzeado]

Slava,

That's is a good point!

In reality, the neutral inrush current is the result of the imbalance of phase currents during transformer energizing. This unbalanced currents cause different flux in each limb of the core, which may not sum zero (three-limbed core). As a consequence, this flux leaks the core and goes throught the tank, being responsable by the transformer neutral current.

As a model, we may think this flux as being produced by a part of the flux produced in each limb having similar characteristic (same magnitude, frequence and sense). So, we may call this flux a "zero sequence flux".

By the way, positive, negative and zero sequence (harmonics as well) are only maths and has no physical mean. But we, as engineers, need them to simplify our studies and explain some phenomena.

Regards,

H. Bronzeado

 

[slavag]

100% Agree with you, with each word!!!!
Best Regards.
Slava

 

[rovineye]

Slava and Bronzeado,
Your description is what I saw when we initially had our P123 Broken Conductor protection set too low/too fast. At times the protection tripped during xfmr inrush.

 

[crossconduction]

With regards to transformer inrush. I have an application involving D - D 450 to 450 transformer. The 3ph transformer is formed from separate single phase transformers. The transformer is feed by solid state source transfer device. The SCRs sometimes cross-conduct during source transfer, meaning the new source SCR turn ON while one SCR of the old source remains ON. One way for an SCR to stay ON with gate logic saying go OFF, is the presence of a DC current component. My question is, with separate transformer cores (D-D), if the volt-second characteristics of each core vary slightly, if there is an effective zero-sequence current which can promote an imbalance leading to a inrush response. Once inrush occurs, the DC component keeps the SCR locked ON. This is seen with the transformer unloaded.

The second question is, I've read that adding a resistive load on the secondary will decrease the inrush recovery time, SO, how does adding resistance load on the Secondary increase the rate of decay of inrush DC current component in the Primary current? Yet, Documents from GE Protective relaying, do not list secondary (load) as affecting inrush decay time.

 

[dpc]

Large transformers are typically energized with no secondary load.

For primary relaying applications, we are mostly concerned with worst-case, so an open secondary would give the longest decay time and is always a possibility if not a probability.

Secondary load just changes the time constant of the primary circuit, since the secondary load is reflected into the primary.

 

[Bronzeado]

Hi folks,

The effect of secondary loads is to reduce the primary voltage and, therefore, the flux generated in transformer windings. This will lower the inrush current peaks.

Regarding the inrush current decay, it depends basicaly on the resistance in the circuit between the source and the transformer being energized. However, there is a electromagnetic phenomenon called "sympathetic interaction" that prolongs this decay.

This phenomenon, which has not been normally considered by power system engineers, occurs when a transformer is switched on a busbar of a weak system on which there are other transformers connected to. If the inrush current (of the transformer being energized) is of high magnitude, the voltage drop across the system resistance caused by this current will saturate the transformers in operation, which will drain (sympathetic) magnetizing currents of high magnitudes. The peaks of this "sympathetic" currents will be of oposite sign of the inrush current peaks.

In this case, the system resistance will be the entity responsable to trigger the sympathetic interaction phenomenon and play the paradoxical role of "villain" that saturate transformers insteady damping it.

Best Regards,

Herivelto Bronzeado

 

[crossconduction]

Heriv, regards, so inrush decay profile is a LR time constant of the primary side (with L on secondary being reflected to the primary side which is influenced by R on secondary side) ??

Let me put it a different way: what are all the factors which can reduce the time it takes for inrush to decay? Do any of those methods include adding resistive load to secondary side?

The third issue is I'm still trying to figure out if and how the DC step response of energizing the primary winding inductance along with the 2nd harmonic of inrush current can create a DC current keeping the SCR ON (due to holding current) when the gate logic is telling it to be OFF during a source transfer. I can deal with the SCR physics, I need help identifying the source of any (even brief) Step or decaying dc current component.

Regards and appreciation to all for responding.

 

[Bronzeado]

Hi Crossconduction,

1. "what are all the factors which can reduce the time it takes for inrush to decay?"

- When just one transformer is to be energized onto a busbar (without other transformers connected to), the resistance in the (primary) circuit between the source and the transformer is fundamentally the unique responsable entity for the decay of the inrush current (negleting the transformer and generator resistance).

2. "Do any of those methods include adding resistive load to secondary side?"

- I have never seen methods that use resistance in the secondary to reduce the inrush decay. On the other hand (and side), circuit breakers with pre-insertion resistors have been used in transmission systems to mitigate transformer inrush currents.

Regarding your third issue, I wonder if you can send much details on that or measurements you have done.

Regards,

Herivelto Bronzeado