Zero Sequence Current flow thru Gen neutral 1

[rconnett]

I have reports from a site I'm involved with regarding unwanted current flow in a generator neutral.

The generator is a solidly grounded 2000 kW/2500 kVA wye connected diesel unit, directly feeding a 2500 kVA step up transformer at 480 volts. The transformer is 480 volts delta (secondary - connected to generator) to 12.5 kV wye (primary - connected to system with a resistance grounded primary).

No current flows until the 480 volt generator breaker is closed, conecting the generator to the transformer, at which time we see about 30 amps flowing on a generator (there are 12 units parallelled at the 12.5 kV level and all exhibit similar behavior, except one which has about 80 amps flowing). The current was measured with a clamp on ammeter directly in the generator neutral.

The ground fault relay trips when this occurs.

I recommended to my clinet that we look at the harmonic current content of the neutral current. - The 3rd harmonic current will look like zero sequence.

Here's my question: If the delta connection traps the triplen harmonics, where's the return path to the generator neutral?? Any other ideas on where to look??

 

[digitrex]

I tried very hard to think what could be the possible cause of generator neutral current and cannot find a logical explanation.

I am trying to relate this to magnetic inductions by the surrounding cables (esp the single-core). May be I need you to provide more information:
1. Are the cables from generator to transformer single-core type? And how long?
2. Is the generator neutral grounding cable armoured or screened type? If yes, does armour-wires or cable screen pass through the CT of ground fault relay?


Other comment: I think your synchronizing facility is located at the 12.5kV breaker. Why not leave the 480V generator breaker closed all the time(not to be used)?

 

[alehman]

I too am at a loss. There must be another path to gound in the 480V systems. It might be interesting to insert a small grounding impedance at one of the generators and see if it changes the ground currnet. Perhaps there is a stray magnetic field that is influencing the ground CT.

 

[mc5w]

You could also use 2 grounding resistors attached to the neutral. One connected to ground all the times. What you do with the other one is to amplitude shift key it with a contactor and repeat cycle relay. You then follow the varying current to its source.

What you can also do with the 2nd resistor is to connect it to a 400 Hertz 120Y208 volt generator from Marathon. You would need to order it with 2 bearings and belt drive with a Goodyear Gripbelt. The generator can be connected for 240Y416 volts as well. The IBM 360 used direct drive motor flywheel generator sets for uninterruptible power - the induction motor would normally drive the generator at say 450 Hertz which sags down to say 380 Hertz during transfer swithc operation. Paper mills back in the 1930s also used a similar concept for uninterruptible power and the synchronous motors were built to tolerate the voltage and frequency sage until a transfore switch restored power.

Finding the source of the fault current could be a problem if there is too much insulation capacitance. If insulation capacitance is resonating at a harmonic voltage or line frequency you could be in trouble. A microfault could also be your problem - a fault near the neutral point has very little voltage motivating it and is very difficult to find with out using a 400 Hertz or amplitude shift keyed signal.

Mike Cole, mc5w@earthlink.net

 

[mc5w]

By the way, your fault current source could be system capacitance which could possibly be resonating or almost resonating at line frequency or a harmonic.

Mike Cole, mc5w@earthlink.net.

 

[rconnett]

Thanks to all for the input. The problem was that there was an inadvertant second neutral ground on the generator neutral. This ground bypassed the ground relay CT in the main bonding jumper (system grounding / earthing conductor) which was improperly routed with the phase conductors. I say improperly, in that it was not triplexed - it was on the outer edge of the (non-conductive)cable trench which held the phase conductors, thus forming a ground loop. To aggravate matters the conductors were arranged:

AAAAAA BBBBBB CCCCCC G

(Not as designed, specifed or as required by NFPA 70/National Electric Code)

The induction which resulted induced a current flow in the ground loop which flowed thru the ground fault CT.

This was discovered by moving the ground conductor to the interior of the cable bundle & the ground current immediately dropped to 3 amps from 30 amps.

The contractor is going to re-configure the conductor bundle as originally designed, with ABCG conductor bundle(s) arrangement in the trench. for each of the 12 generators. (Hopefully next time he will pay attention to the design detail). Also, the client will be removing the second grounding jumper.

Thanks to all for your input.

 

[Alex68]

Dear rconnect,
Every generator produces a little ammount of zero sequence voltage depending on the winding construction.
The two neutral to ground connections constitute a path for the zero sequence paths. In your particular case the Io current is exactly the 3% of the phase current in all the load conditions.
The best solution should be to remove one of the two connections, but this is not possible for diplomatic problems. So I have already suggested your DPL (GV) to rise the protection setting.

Regards
Alex68

 

[mc5w]

I am glad that you found your source of excessive ground current. National Electrical Code does allow isolated phase and neutral for underground installations but perhaps a better conductor arrangement in the trench would be

AAA BBB CCC GG AAA BBB CCC or

AABBCC G AABBCC G AABBCC or something like that.

How long is your distance from generator to transformer at 480 volts? At a long enough distance you might want to transpose conductor so as to equalize to phase to phase reactance and to reduce induction of current in adjacent control circuits:

First third of trench: AAAAAA BBBBBB CCCCCC G
Next thid of trench: BBBBBB CCCCCC AAAAAA G
Last third of trench: CCCCCC AAAAAA BBBBBB G

This is similar in concept to quadplexing the phase and grounding conductors but with fewer "twists".

Cleveland Electric Illuminating Company has a high capacity quad circuit 138 KV line with transpositions in it going from their Eastlake plant to a substation that sends 34,500 volts and 11,500 volts to the eastern suburbs and downtown Cleveland, Ohio.

You might want to reconsider resistor grounding with amplitude shift keying and a 400 Hertz generator for troubleshooting. You could build the resistor using a bank of 277 volt baseboard heaters. Resistor grounding will also greatly limit the damage that would be caused by an arcing fault. Sounds to me that your ground fault protection relay would work fine with resistor grounding if the resistor allows enough fault current.

Mike Cole, mc5w@earthlink.net

 

[rconnett]

The distance is pretty short - about 30 feet.

The point of resistor grounding is more for machine protection. The generator vendor is responsible for this, and my scope stopped at the generator terminals. I was responsible for all the relaying at the plant, & can't really get into their scope. I did advise them to do the resistance grounding, however but they elected not to & contractually I can't push any harder than this..

Thanks for the suggestion!