DG trip on volts/hz 3

[electricpete]

DG's are a little out of my area. I was asked to provide input on the following scenario:

700 kw 480 vac Kato / Reliance diesel generator.
Generator and system are ungrounded.
Generator was supplying MCC in islanded mode, with low load on the MCC, but load bank resistors were added to bring total generator load to approx 610KW for test purposes.

A large block of lighting fed from the bus auto-switched on. The amount may be equivalent of several hundred kva.. suspected reactive ballast inrush according to an engineer more familiar with the lighting system.

At approx the same time, a generator volts/hz trip occurred and undervoltage alarm occurred expected following generator trip).

No other protection or alarms occurred. The other protection/alarms include:
"Generator Overcurrent with voltage restraint" 51V
64F (field ground?)
overspeed, low-lube oil, high-cooling-water-temperature

It is suspected that the block of lighting load caused the diesel to slow resulting in low hz and therefore high volts/hz trip.

Any thoughts on this scenario?

I have 2 reservations:
1 - if it is truly inductive inrush load, then it probably would not slow down generator but would mess with the voltage control.
2 - Why didn't overcurrent trip? Would you generally expect to get an overcurrent 51V?... (I'm sure it may require knowledge of time current characteristics, and inertia etc, but maybe someone is familiar with typical behavior).

What testing would you recommend ?

What would be the purpose of voltage restraint of the overcurrent ?

=====================================
(2B)+(2B)' ?

 

[electricpete]

When I listed the other protection that did not trip, I left out one:
64 = ground detection circuit associated with the generator ac output.

=====================================
(2B)+(2B)' ?

 

[wayne440]

Are you saying the set was already loaded to 610KW, then "several hundred kva" of lights were simultaneously added to that load?

 

[electricpete]

Yes, the load was definitely exactly 610kw prior to the lights coming on. The load posed by the lights is somewhat unknown, but estimated as several hundred kva

=====================================
(2B)+(2B)' ?

 

[electricpete]

Clarifcatio in bold:
Yes, the load was definitely exactly 610kw and pf near 1.0 prior to the lights coming on. The load posed by the lights is somewhat unknown, but estimated as several hundred kva

=====================================
(2B)+(2B)' ?

 

[rasevskii]

You overloaded it...at least with kVA. It did not like that.

rasevskii

 

[rasevskii]

You cannot get more kVA out that what the nameplate says, no matter what the PF. The stator amps must have gone well over rated, the AVR went to max output, and the AC voltage collapsed which must be why 51V operated. Be glad you had working protections. Otherwise a new Genset.

That was an Electrical Tsunami...

rasevskii

 

[rasevskii]

Reading the OP again, you say that 51V DID NOT trip, in fact. What were the settings? Time Lag? Was anyone looking at the stator ammeters? The undervoltage trip had to be correct.

51V is basically overcurrent protection which is high set at normal voltage, but reduces on undervoltage to a lower setting according to a curve. The time lag may have been such that the protection "started" but did not trip as the UV operated first.

rasevskii

 

[wayne440]

I guess I'm missing something. I understood the OP to state that the set tripped on V/Hz, not 51V. A large block of lamps certainly have the potential to upset voltage control. Having more info about the lamp load might help-HPS, mercury, flourescent or ?

 

[electricpete]

good points.

we only had undervoltage alarm, not undervoltage trip. The only trip was volts/hz.

The preliminary plan is:
megger generator
check relay cal
Repeat dg run to check performance

=====================================
(2B)+(2B)' ?

 

[rasevskii]

You might want to add an UV trip with a suitable time lag. In no case should the generator breaker be allowed to remain closed if undervoltage occurs. This has to have the trip circuit activated over an aux contact on the breaker to avoid a standing trip when the unit is not running. Likely this is the reason that it is only an alarm at present.

rasevskii

 

[dpc]

I suspect the step load increase of the lighting caused the AVR to overshoot the voltage setpoint, maybe combined with reduced frequency due to slowing of the engine on kW overload. Also, the lighting ballast load can be rich in harmonics, and this can be issue for some older AVRs.

The purpose of the voltage restraint is to allow the relay to operate on sustained fault contributions from the generator that are not cleared by other relays. The fault current produced by the generator will rapidly decay for an extended fault. Absent field-forcing, the current is ultimately limited by the synchronous reactance of the generators and this is often greater than 100%, meaning the fault current contributed by the generator is less than the full load current rating. At very low voltage, such as during a fault, the relay curve shifts to a more sensitive pickup, allowing the relay to operate on current that are less than load current. These relays generally will be fairly slow to operated for an external fault.

David Castor

http://www.cvoes.com

 

[waross]

Take a good look at the AVR. If the UFRO jumper is on 50 Hz for a 60 Hz system, you will get high V/Hz if an overload slows the prime mover.
The UFRO feature typically allows a 3 Hz drop and then lowers the voltage as the frequency drops, avoiding a V/Hz trip. If the jumper is on the 50 Hz position you will drop 13 Hz before the voltage is reduced and if you have V/Hz protection it will probably trip.
I almost never see a V/Hz trip on a diesel set in that size range as the UFRO feature of the AVR maintains a safe V/Hz ratio.

Bill
--------------------
"Why not the best?"
Jimmy Carter