PT Primary Connections 2

[timm33333]

If we feed a transformer-relay (digital) with Y-Y connected PT’s, with primary Y of PT’s ungrounded and secondary Y grounded; will the fact that the primary of PT’s is not grounded cause malfunction of the relay? Thanks

 

[jghrist]

I should have said:
Unless you connect a neutral as well as a ground to the VT primary, it will not be shorting out the NER.

 

[timm33333]

Yes I had figured it out, thanks!

 

[LiteYear]

How is that shorting out the NER? The primary of a VT has a quite high impedance. A wye-wye VT, grounded on both sides, can't contribute any ground current to a fault unless there is a ground source on the secondary side, something I've never seen. That's all a bunch of hog wash.

Quite right, thanks for the criticism - my scenario only concerns the magnetising current of the VT. Yes, provided the VT is perfect and the magnetising current is zero, there can be no ground current on the primary without ground current on the secondary. But consider these three scenarios:

[ul]
[li]each VT contributes some magnetising current in parallel with the NER current. How many VTs does it take to become significant?[/li]
[li]once the VT is saturated, the magnetising current becomes very large. How much headroom do your VTs have under fault conditions?[/li]
[li]magnetising current at DC is very large. Do you have any DC sources that can contribute to fault current?[/li]
[/ul]

Once you've considered and eliminated those possibilities, and any others I haven't thought of, you can go back to your ideal VT model. Or you can just stop introducing additional ground paths on resistance grounded networks and sleep easy.

 

[krisys]

The VT magnetizing current is so negligible (in the order of few mA) that it need not be taken into account while dealing with the subject issue. We are comparing mA with the earth fault current thru NER which is the order of 200 to 400 Amps.

 

[davidbeach]

Wow.

Even on a high resistance grounded generator with only 5-10 amps of ground current the VT magnetizing current is several orders of magnitude lower.

On resistance grounded systems the VT primary should be rated for phase-phase voltage, so saturation is not an issue.

DC is going to cause enough problems all over the power system that any sources will be quickly rooted out and corrected. Half-wave saturation of a power transformer gets lots of attention and the DC (more likely even harmonics rather than actual DC) source gets removed to stop the power transformer from growling.

 

[LiteYear]

Good-o, glad you've got it all sorted. Where I'm from people use things like VSDs and very sensitive ground fault relays, buy their VTs from the cheapest supplier and use lots of them. Ergo, somewhere in the world it's a problem and has led to failures in the field. Fortunately not in your part of the world.

 

[timm33333]

DavidBeach, the manufacturer says that the test current that they will use for testing of NGR will be DC current, so it will pass through transformer primary as the transformer primary will act as short circuit to DC. It still does not make sense to me because transformer primary is inductor which will delay the passage of current and in the mean time the NGR will trip. What do you think about it?

 

[davidbeach]

Test the NGR when? Under what conditions?

 

[timm33333]

Instead of waiting for the ground fault to occur, they will intentionally inject external DC current into the NGR circuit. This external DC current will act as ground fault current, and will trip the NGR. The test will pass if the NGR trips on this external DC current.

 

[davidbeach]

So they test it and put it in service. If you want to test it again, open the disconnect to the VT. It's not like that's something that you're going to need to worry about very often.

 

[scottf]

Sounds like you're mixing a few concepts here.

A pure inductor does act like a short to DC. However, in reality that's speaking to the inductive reactance. The VT's primary also has a very high DC resistance...probably in the 1-5 k-ohm range depending on the voltage class.

With all of that said, it's always a good idea to remove VTs from the circuit when doing DC tests. It's more important if doing DC voltage stress tests, but also a good idea when doing DC current tests, especially if the voltage will be anywhere over a few volts. Dc can send the VT into deep saturation and could damage the unit.

 

[stevenal]

How does an NGR trip? Breakers trip, resistors melt.

 

[timm33333]

NGR feeds monitor and monitor sends 4-20 mA signal for tripping the breaker, as attached.

 

[LiteYear]

timm33333, for what it is worth, I am familiar with neutral-grounding resistor monitors and high-resistance grounded networks. I understand there is a continuous online test. I've measured the waveform produced by the SE-330 in particular. I understand, as does Startco, that putting the primary of your VTs in parallel with the neutral-grounding resistor by grounding the VT primary neutral is problematic for the reason they mention and many others. You're free to try to convince yourself otherwise but the facts remain. scottf is completely correct, but is not familiar with the SE-330 operation. It is not just as you describe.

 

[scottf]

Liteyear-

Can you provide some more information on why a VT connected in parallel with the NGR will cause a problem?

 

[timm33333]

If we wind the primary of PT in delta, then zero sequence voltage will not be passed on to the relay. The only practical way to pass the zero sequence voltage to the relay is if the PT primary is Y-grounded.

 

[LiteYear]

Can you provide some more information on why a VT connected in parallel with the NGR will cause a problem?

If you specifically mean with regard to a NGR monitor, then I will have a go. For an incomplete list of other potential problems, see my posts at 28 May 14 1:15 and 28 May 14 18:37.

NGR monitors attempt to confirm the resistance of the NGR is within tolerance (typically -50%/+100%) by continuously injecting a square wave voltage across the NGR. Naturally the rest of the system is in parallel with the NGR, and therefore must be discounted. A substantial DC component is used so that the system capacitance can be neglected, but it is assumed that the DC resistance to ground of the rest of the circuit is very high. In fact, NGR monitors are typically coupled with ground fault lockout devices that ensure phase to ground resistance is >1MΩ. At medium voltage levels the primary resistance of Y-Y VTs might be a few kohms, but there is no guarantee. Every VT combines in parallel, and it doesn't take much for them to make a significant impact on measuring the few hundred ohms of the NGR.

The fact that NGR monitors and ground fault lockout devices exist should suggest how important it is to keep stray paths to ground out of high-resistance grounded networks.

 

[davidbeach]

Or, it should suggest that NGR monitoring systems should recognize the realities of real life. In this day and age of numeric relays there is no excuse or justification for VTs with any primary connection other than a grounded wye. Period. Full stop.

On high resistance grounded systems the secondary of VT can go with a phase grounded rather than the wye point; anything else the secondary should also have the wye point grounded.

To hobble the rest of your protection simply to make it easier for an NGR monitor manufacture seems to be getting priorities screwed up. We have many NGRs and nary a monitoring system that precludes proper protection on the system.

 

[stevenal]

The diagram shows a CT (EFCT-X) monitoring the current. Not sure how the manufacturer expects DC to properly test this setup.

 

[timm33333]

It looks that DC current is injected out from terminal 6.