Parallel Generators/Grid Grounding Scheme (NGR) Questions (w/ diagram)

[tumbleweed1]

We've got a 4160V system with 4 generators FLA=370A (grounded Y) paralleled to a utility transformer. All gens and the util TX have 25A/10s NGRs on the neutral point. All NGRs have relays and monitors. All feeders with motors have zero sequence CTs, are ungrounded.

I understand neutral currents occur because of a few things:
- Ground faults with capacitive charging current discharge
- Variations of voltages between sources causing unbalanced currents to travel through the system

I've read a ton of published articles and threads on here regarding HR grounding. My questions are as follows:

I have calculated 3phase system charging current to be about 7A per generator (assuming 0.23uF/phase for generator capacitance, 0.5uF/phase surge caps, 136pF/ft*1000ft cable length), ~28A total for all generators.

1. To handle scenario 1 (4 gens max online) Is it wise to have each NGR sized to handle the total charging current from the whole system? Or would 25A be suitable as, ideally, the GF relay would trip accordingly?

2. Its my thought that the current through each NGR would be equal if they were sized equally and there was a fault on a feeder branch. How does one prevent tripping of an unfaulted feeder/generator from occurring?

3. In scenario 3, when only the utility is supplying the load feeders, is it sufficient to use Zero Sequence CTs to determine presence of fault currents to trip off necessary feeders? I guess the NGR on the TX will limit the current, correct?

4. Would a zigzag transformer connected to the main bus be useful for scenario 3 or for any other operation?

As you can tell, Im quite new to grounding schemes and would appreciate some input from those that understand this. Thanks!

 

[collies99]

My thoughts:

1.Using 25A rating is OK as long as I^2t at 28A or Ifault max us not greater than I^2t rating of 25A ngr at back-up clearing time. I fault will be maintainef until back-up protection trips if the primary protection fails.

2. A fault,on a motor branch feeder has to be coordinated with the gen ngr tripping time. All on-line gens and utility former will source the fault current and be seen as a fault on the bus or the motor feeder. I am ASS-U-Me-Ing that there is a breaker feeding each motor branch feeder.

A fault between any generators upstream of the motor feeder breaker will trip all generators and utility transformer.

A fault on the motor feeder requires coordination with the motor feeder pn. This requires selectivity.

3. As noted in 2. utility supply xformer needs coordination with motor branch feeder protection.

4.Zigzag not applicable in this configuration as,you have wye grounded system.

 

[collies99]

Design question:

1. Are you using Gen differential for a fault between the Gen neutral an its unit breaker ?

2. Are you installing a stn service high side breaker ?

3. Are you having main bus differential protection ?

 

[tumbleweed1]

Sorry, the drawing is not 100% accurate. Yes there is a breaker on every feeder, generator, util tx, all VFD-controlled motors and 2MVA station service TX.

Yes there is Gen differential. As well as neutral sensing CT. Would it make sense to have a ZSCT on these feeders as well if we're already monitoring the neutral?

Yes high side breaker.

No main bus differential.

Each motor feeder has ZSCT and 3ph relay CT.

 

[tumbleweed1]

Another question, if we have some value of resistance on the neutral, and each generator is in service, so effectively 4 NGRs in parallel, wouldnt this lower the overall effective neutral resistance in the zero sequence circuit?

Wouldnt this be where a ZZTX w/NGR on neutral alone on the common bus be utilized to be a single point of GF current reduction and sensing?

 

[collies99]

Each ground source will contribute fault current to the fault location, in this case the 4 generators and the utility transformer. When you reduce the 5 sources to an equivalent single source, yes it effectively reduce the overall resistance. The ground fault impedance/resistance will be whatever it will during the event, low to high resistance.

To limit the ground fault current on a system you can either increase,the ground path impedance as is already done (neutral resistor/reactor et al) or install a delta system at the other end of the scale, as being ungrounded ... Which means the protection scheme will be different than what is proposed.

should be ok.

 

[odlanor]

During operation of 4 generators and the utility transformer, a maximum ground fault inside any generator shall be 50A !!
A stator core damage from arcing fault of 50A should be eliminated less 0.7seg to provide negligible arc burning of the core.

 

[tumbleweed1]

Where do you come up with 50A?

And where do you decide 0.7s? I suppose a guy could do some math integration based on the amount of total power for the amount of seconds and decide the amount of joules that the windings are subjected to. But even then, how do you know how many joules it is good for?

 

[dpc]

The total ground fault current will be the sum of all of the NGR ratings - basically. So it will be nore than 50 A. At 25 A I think you are in an unhealthy no-man's land between HRG and LRG. You also can have harmonic current, principally third harmonic flowing between the transformer neutral and the generator neutral that your resistors must carry on a continuous basis - not 10 seconds.

I'd recommend this book:

http://www.groundingdesignbook.com/

 

[tumbleweed1]

I'll look into that book. Have a link for a torrent download for it??

What do you think of slide 58-60 in this IEEE presentation powerpoint:

http://www.hv-eng.com/2010IASGrounding.pdf

 

[odlanor]

ERRATA

During operation of 4 generators and the utility transformer, a maximum ground fault inside any generator shall be 4*25 + 1*25= 125A !!
A stator core damage from arcing fault of 125A should be eliminated instantaneously to provide negligible or light arc burning of the core.

 

[odlanor]

see PROTECTIVE RELAYING FOR POWER GENERATION SYSTEMS - DONALD REIMERT chp5.22.3

 

[collies99]

We use neutral grounding transformer and secondary resistance sized to about 3Z0 which in effect reduces current to about 15-20 amps, but each unit is feed into a delta winding which blocks zero sequence from parallel Gen and the HV bus.

For bus connected generators, diesels non integrated units, they connected delta in its winding generator. Ground fault protection are detected using a zero seq voltage detector network which are just an open corner delta connected VTs with a voltage relay with an inverse-time curve.

 

[dpc]

Have a link for a torrent download for it??

If you are doing this for a living, the cost of that book should not be an issue.

 

[tumbleweed1]

Haha- this is true. But to reference that book two days ago was regardless of cost- more like availabilty.

 

[mosesnbklyn]

I would recommend running a simulation using software. Generators have significant impedance and multiple units in parallel, with possibly different qty running will change the fault current. This is not easy to do by hand.

Some sizing criteria from experience:
-The max NGR current is 1 unit operating, rest of units disconnected
-The min NGR ohm rating should be determined from calculation software with all equipment running. You have some choices for selecting the max. total system fault current contribution, that will ultimately determine the min. rating; one local utility uses 400A contribution from the customer on the distribution circuit as the upper limit. Its a little bit of trial and error for the first time with each voltage system.
-you should also get the technical limitation of the generator insulation/windings from the OEM.

If you need to find a consultant who has ETAP, ASPEN or other simulation software check with your local or nearest IEEE PES chapter:
Link

M. Nissen,P.E.
Senior Electrical Engineer
Waldron Engineering & Construction

M. Nissen,P.E.
Senior Electrical Engineer
Waldron Engineering & Construction

 

[tumbleweed1]

We actually have ETAP, but I havent spent the time to set it up and figure everything out.

Is there a recommended method to setting it all up? I mean i can easily model the system, add the data in for the generators etc. Place a fault on a line somewhere and do a load flow? But in the end there's so many variables with modelling that misconstruing the results could be as detrimental as making a decision without modelling.

It's a larger Stamford generator. I have the typical generator data sheets. Is there something more that you think the manufacturer could provide as far as the limitation to the winding insulation?

 

[mosesnbklyn]

Run a short circuit simulation as I described, you can also get the utility fault current/impedance upon request (it may take some time).

Another equipment limitation that should be checked for would be the lightning arrestors that the utility uses; if they are rated for Y-g system, using NGR's causes neutral shift and may exceed the LA ratings during fault conditions. ETAP will provide this, the utility might add a 1kV switching surge (on 15kV systems) as a buffer, so it might get tight. I would contact the utility engineers - they will ultimately have to approve of the system.

M. Nissen,P.E.
Senior Electrical Engineer
Waldron Engineering & Construction