Long distribution line w/o neutral

[SphincterBoy]

We have a long, three-phase distribution line without a neutral, which powers six delta-connected transformers.

The system has NO ground fault (GF) protection, but simple inverse time phase overcurrent protection.

Whenever a line segment comes down, it remains energized (high-impedance fault), without the breaker going to lockout.

Can I utilize negative-sequence protection to better protect the line, and to make certain it goes to lockout whenever a line comes down?

If so, what settings should I set the negative-sequence trip value for? I cannot find any literature for negative sequence protection whatsoever.

 

[DanDel]

If this is an ungrounded system, then one line going to ground shouldn't be a fault at all, high resistance or not.

If it is a grounded system (without carrying a neutral), can't you install zero-sequence protection? Just set it to a low value based on the highest zero-sequence impedance at the furthest point in the system.

Negative sequence protection can be used to trip on an open line, check out SEL (Schweitzer) 351, or almost anything they make, for negative sequence protection.

 

[Kiribanda]

SphincterBoy,

I donot think it is adviseable to use any current detecting protective scheme on an ungrounded system.

As you also know, during an earth fault on an ungrounded system, the voltage of the healthy phases will go up and that can be used to trip the feeder by making use of an over voltage (ANSI 59)relay via a suitably rated voltage transformer. Since the system is ungrounded you donot need a trip/ lock scheme. An indication is more than enough to investigate the fault by your maintenance crew.

Otherwise you can install a Zig-Zag transformer on the ungrounded system and then the current through its artificial neutral during an earth fault (ANSI 50/51G) can be used to trip/ lock the feeder breaker. An NGR can also be added to this scheme to limit the ground fault current depending on the nominal voltage of your ungrounded system.

So the final scheme depends on whether your system is a cable or an overhead line.

Hope this helps.

Kiri

 

[SphincterBoy]

Thanks for the comments!

There are no phase-ground loads anywhere on the line.

All single and three phase loads are connected line-line.

There are NO grounded points whatosever.

 

[stevenal]

Assuming the source is grounded:
For a SLG fault, I0=I2. No advantage in using I2 for this situation. I2 can be used to differentiate ph to ph faults from load. If there are no line to neutral connected loads, ground fault protection can be set very low (sensitive earth fault). If there are line to neutral loads, GE and ABB have products that identify high impedance faults by their waveform.

 

[davidbeach]

Assuming the source is not grounded:
Three VTs on the line, connected in a grounded-wye, can be used to detect the voltage shift caused by a down line. The secondary of the VTs could be connected in wye and used with a 3-phase voltage relay and you would then know which phase was down.

Or the secondaries of the VTs could be connected in broken delta and a single voltage relay (with a stabilizing resistor) used to measure the resulting voltage. This would tell you that you have a line down, but not which one.

Far better though, would be to provide a ground source for the system and let the ground fault draw enough current for detection that way.

 

[SphincterBoy]

The feeder source is solidly grounded. Also, I just so happen to have three VT (metering accuracy) connected at the feeder source.

How low do I set the Io setting for sensative ground fault?

 

[davidbeach]

How low do I set the Io setting for sensative ground fault?

It depends.

You'll have to do the calculations on your circuit. With no line-neutral loads you can go much lower than you could if there were line-neutral loads. You probably can't go low enough though to find all down lines without risking unnecessary trips. Down lines can have surprisingly high resistance to ground and surprisingly low fault currents.

 

[stevenal]

ftp://ftp.areva-td.com/NPAG/Chap9-122-151.pdf.

 

[SphincterBoy]

Since I have the VTs at the source, could I make use of distance relays as opposed to either sensitive GF or negative sequence relays?

 

[davidbeach]

Distance relays won't help much if the fault doesn't draw enough current. A ground relay is probably your best bet unless you want to go with something that can detect the arcing characteristics of a down line with high resistance to ground. You will be able to detect faults above some minimum current level (below some maximum fault resistance level) and having no phase-neutral loads your minimum current level will be much lower than it would be if you had phase-neutral loads.

 

[stevenal]

Guess the direct link doesn't work. Try this one: ftp://tdeftp_public:tdeftp_public@ftp.areva-td.com/NPAG/Chap9-122-151.pdf

Distance protection is handy when coordinating with other distance protection. Directional ground perhaps?

 

[stevenal]

Damn. Okay, go here and click on Chapter 9:

http://www.areva-td.com/static/html/TDE-AGF_Product-Product_Detail2_1056536208199.html

 

[jghrist]

You probably do not have cables to pass through a core balance CT for sensitive earth fault protection, so you will be limited to using the residual from the phase CTs. The minimum setting available in microprocessor relays is typically 10% of the nominal CT current. I wouldn't go less than that anyway. You might get a false residual current under heavy load or inrush because of CT errors.

I don't really see any advantage to using ground distance or directional ground overcurrent relaying.

With no neutral, the ground fault current may be quite low even for a solid ground fault. Have you done any fault studies to determine what the phase-to-ground fault current would be? The minimum setting for residual current ground relaying may be too high to detect a ground fault at the end of the line. If this is the case, you may have no other option but to try one of the relays mentioned by stevenal that detect high impedance ground faults.