AVR Droop CT Harmonic Current Response

[niallnz]

I have situation of when a 280kVA standby generator is supplying the site load (mains fail) the voltage oscillates by approximately 5V at around 2Hz. The site load is about 135kW @ 0.98 leading, and consists of a UPS, air conditioning plant (VSD driven) and the usual basic building load. The generator is a typical diesel driven Stamford 4pole alternator with an MX321 AVR. The site has two generators (N+1 redundancy) and it doesn't mater which generator is supplying the load the voltage will oscillate. If the two generators are run in parallel supplying the isolated site load the voltage oscillation is not apparent.

The UPS and the VSD's have enough filtering to keep the voltage THD to 3% when running on the generator, but the current THD is around 20%.

Do you think it is possible that the AVR is responding to the current harmonics?

As far as I can tell the generator is operating with its capability area. Could there be a resonance interaction with the harmonic filters on VSD's?

Your thoughts please
Cheers Niall

 

[waross]

First try tweaking the stability adjustments. At no load, adjust until the voltage becomes unstable and then trim the adjustment until the voltage becomes stable. If that is the present setting, try adjusting to add a little more stability.
Then try switching off the UPS and the AC in turn to see if one of then is responsible for the instability.

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

 

[niallnz]

The AVR's on both generators were reset using the manufactures procedure. The no load voltage setpoint of both AVR's was set so that they are identical (232.0V L-N). Made no difference. The speed of the oscillation I think is too slow to be an AVR stability issue.

Turning off the UPS (bypass mode) is not the clients prefered test option due to the sensitivity of the load. Minimizing the AC load didn't make any difference either.

Cheers Niall

 

[ScottyUK]

Set up some recording instruments to trend the system load, current, and voltage. If you can get governor data too that's even better, although I think this is a bit too fast for a governor problem. The VSD and UPS rectifier will both act as constant power loads, so as the line voltage drops the current will increase for any given load condition. You may well have some instability / poor dynamic response in the AVR because of this 'negative impedance'* load. If you have a low frequency voltage rolloff then this would interact adversely with the VSD load by causing current to rise as voltage falls.


* Not my name for it, before anyone starts arguing!


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[niallnz]

The full system is this. Cummins diesels with Cummins EFC's controlled by Woodward DPG 2401's (actually a Barber Colman governor but Woodward bought them out). Synchronising and Load sharing is by Woodward EGCP-2's. The Real load sharing is very good, no problems with it at all. When the two generators are running in parallel they share the load approx 65kW each and the small leading VAR load (15-20kVAr) swings slowly from on generator to the other but never runs away. It is a lot harder to tell if the frequency of reactive power swing is the same as the voltage oscillation when a generator is running by its self as all I've got to go on is the EGCP screens which don't up date that fast.

Cheers Niall

 

[catserveng]

You might try looking at the field output and the voltage bias input at the same time, see if the EGCP2 is trying to drive the voltage or if the bias input is stable and the field output is varying due to the AVR control functions.

With the unit running standalone you could also disconnect the voltage bias wiring and see if the problem goes away or stays.

Also see how much of a change in the field output from rated voltage no load to when the load with a leading power factor is applied. If it's really a leading power factor the field output will drop slightly. Does the generator voltage stay at rated or rise up a bit? Is the voltage trim in the EGCP2 active? Some Cummins packages have the loss of field function active and set pretty tight, have seen stability issues because the AVR or controller thinks the leading PF is a loss of field and shuts off excitation and doesn't latch, casuing an oscillation.

Hope that helps,

Mike L.