Rotary UPS Grounding question

[mparenteau]

All,

I am the owner's rep for a project in design, which is to install two 2.1MW, 15kV class rotary UPSs (coupled with diesel engines for extended outages). Upon loss of utility, we obviously lose our ground reference.

Two options:

1) Zig-zag w/ resistor (LRG). We would have to upgrade our arrestors since they are L-N rated.
2) Neutral reactor for each machine.

I would size the NGR (option #2) such that the SLG is limited at the machine terminals to no more than the 3 phase fault current available from the machine (expansion accounted for). If I meet the IEEE standard for an effectively grounded system (X0/X1 < 3), do the panelists see a problem with this approach, assuming we wouldn't replace our arrestors? Even if we choose to expand our system, there shouldn't be an issue with ground current circulation due to matched sets. Perhaps harmonics could be an issue, but doesn't seem likely. Am I missing something? I know that IEEE C62.92.2 recommends option #1, but am seeking some further input.

 

[cranky108]

Are your loads at 15kV? If not you would need a transformer, which might allow you to keep your ground reference.
But not enough information to give a good answer.

 

[waross]

Both a suggestion and a question:
Suggestion:> Rather than a zig-zag transformer use three distribution transformers in wye delta. Off the shelf, no special order nor long lead times. Rapid availability of replacements in the event of a transformer failure.
Question:> Are there any comments or comparisons of zig-zag versus wye:delta?
Is there any down side to using a wye delta bank?

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[mparenteau]

Thanks for the comments thus far:

1) We have no direct 15kV loads. We have about 60, Delta-Wye grounded transformers which are part of the electrical service entrance(s). We have both 480/277V and 208/120V secondaries.

2) We didn't consider a Wye-Delta grounding transformer, but doesn't mean we couldn't. Again, this would be the reference ground for all of the units for a power loss condition only.

3) Unfortunately, our distribution is comprised of approximately 30 S&C switches, so there is no metal-clad switchgear (with medium voltage power circuit breakers) to speak of in the existing system.

 

[waross]

With the wye:delta, not to be confused with a delta:wye, the delta voltage is unimportant.
A properly sized current limiting impedance may be inserted either in the line from the wye point to the ground electrode or in the corner of a broken delta.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[mparenteau]

I'll answer my own question in this thread:

According to IEEE C62.92.2, separate neutral grounding reactors is considered an outdated method of gen grounding. So, we'll go with a grounding transformer w/ a resistor and deal with the existing equipment issues as needed.

Mike

 

[wcaseyharman]

Just some thoughts - you might add grounding resistors to the generator neutral to ensure it has a ground reference starting up, it will also allow easy detection of generator ground faults during startup.

One of the downsides of grounding at each gen is your ground fault magnitude changes if one or both of the generators is online. An external grounding back would keep the ground fault level fixed.

I assume you’ll have a breaker or ganged switch to your grounding bank which would be open during normal grid connected ops, otherwise you’ll feed ground fault current to utility faults, and the utility would have to study and possibly re-coordinate their circuits with a remote ground source.

 

[waross]

I would. if at all possible, common all the neutrals and then use one grounding location and grounding device.


--------------------
Ohm's law
Not just a good idea;
It's the LAW!