Source of Generator Current Unbalance? 1

[davidbeach]

I was looking at an event report triggered when a generator was tripped off-line this afternoon and would like some conformation - or alternate explanations - for what I think happened.

Generator is question is a 2MW diesel engine driven gen, 480V, new generator installed on the engine following a messy failure of the previous generator. Generator was running in parallel with the utility at the time. The relay that tripped the generator was between the generator and any other connections, load or utility. The voltage at the bus was rock steady and the angular separation of the three phases was 120[°][±]0.5[°]. Therefore the currents should have been very close. But, it wasn't even close - 440A, 450A, and 505A (in round numbers, there was some fluctuation).

As I understand the situation, when the generator is running in PF/VAr control mode, power into the shaft, from the prime mover, determines power out and the rotor settles at the necessary angle relative to the utility to meet the requirements of the power transfer equation. The equation can be evaluated on a three-phase basis or it can be evaluated phase-by-phase. With a nearly perfectly balanced system voltage and a common internal voltage, assuming identical impedances, each phase should solve to the same value, should it not? The excitation system will control the amount of reactive power, but again, assuming identical impedances, there is nothing that would act differently on different phases, is there?

My contention is that there is a serious problem somewhere between the CTs of the relay and the neutral of the generator, something that is presenting quite different impedances to the different phases. Is there anything beyond the relay (remember nearly perfect voltages) that could cause a current unbalance when running in parallel with the utility in PF or VAr control mode? Any control system (AVR/Excitation/etc.) failures that could do this?

Thanks.

 

[Hoxton]

Whether it is in pf mode (in parallel with utility) or in voltage control mode (single running) does not affect the power (kW). When in parallel with the utility, the engine fuel sets the power generated. I assume that the set has some form of kW control, to control the power when in parallel with the utility.

Assume you have a 2500kVA alternator, full load current at 0.8pf is 3007 amp. So this set was not at full load (kW or kVA)

What did it trip on? What was the relay or device that tripped, or was it a breaker with inbuit protection? Were the voltages balanced?

Was this the first run since repacing the alternator?

 

[ScottyUK]

Hi David,

What relay function operated - negative phase sequence?

I think you can discount AVR and exciter problems from causing this problem: AVRs fitted to small sets simply don't have the level of field forcing required to modulate the individual phases as the machine rotates.

Have you looked at the obvious things such as busbar joints and the neutral link bar? Are the CTs matched and can you arrange a mag curve test and ideally a primary injection for the CTs to prove it?

How close were the nearly perfect voltages? If the local grid is particularly strong and also slightly imbalanced then it may be able to influence the generator terminal voltage to cause the current imbalance you see.

Is the generator direct coupled to the system or does it have a transformer? Is the transformer the standard dYN1? How is the generator grounded - solid or impedance?


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

Sometimes I only open my mouth to swap feet...

 

[waross]

I assume that the set has some form of kW control, to control the power when in parallel with the utility.

This would be the position of the fuel control mechanism and a mechanical stop on the maximum fuel position of the fuel injection system.

David;
The CTs are indicating unbalanced currents. One of my first reactions would be to check the CTs and associated wiring. The fault currents associated with the "messy failure" of the previous generator may have damaged the CTs or wiring.
Is the relay new or is there a possibility that it is out of calibration?
Is the air gap on the alternator concentric?
Respectfully

 

[jghrist]

The terminal voltage will not be affected a lot by small (in relation to the rating) unbalanced currents. If the load is unbalanced, the generator will produce unbalanced currents to feed the load.

What is the connection of the transformers between the generator and the utility. A grd Y - grd Y connection will allow the generator to serve unbalanced loads on the utility system. It may take a larger than expected share of the utility unbalance because of the relatively low zero-sequence impedance of the generator.

 

[electricpete]

Would be interesting to know what relay tripped.

Assuming you have the time waveforms in digital format, it might be interesting to construct the sum of the three phase currents and see how closely they sum to 0.

In the world of motors, current unbalance can be much higher at low load than at full load. I have seen it many times. You might have 20% unbalance at no-load 5% unbalance at half load and 2% unbalance at full load. The exact explanation, I was never sure. Actually come to think of it, I have a reference as well showing that behavior. Let me dig it up.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

Here is the figure I mentioned, from "Energy Efficient Electric Motors"

http://home.houston.rr.com/electricpete/eng-tips/CurrentUnbalanceVsVoltUnbalanceLR_NL_FL.ppt

Could also be unbalanced excitation or shorted rotor turn. As you may know, shorted turn can exist in a sync generator rotor for a long time without any damaging effects.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[jghrist]

Maybe I'm not getting something. This is a generator, not a motor, right? Generators can serve unbalanced loads. This doesn't mean anything is wrong with the generator like fuel control mechanisms, unbalanced excitation, or shorted turns.

 

[davidbeach]

The trip was negative sequence current. There was metering data that showed the same imbalance. I don't know if that came off the same CTs or not, but two independent sets of AD converters produced the same result, so it isn't a relay problem, might be CTs, will have to find out. Messy failure of the previous generator was the result of something coming undone in the generator and things on the stator coming in contact with things on the shaft (I think it was part of the fan) and ripping out a bunch of end turns as well as pulling apart the fan (or what ever it was, not poles).

Connection to the grid is through a YNyn transformer; both transformer and generator are solidly grounded. The transformer is the site service transformer, site loads are on the same side of the transformer as the generator.

Lots of good questions, things to check out, thanks. Gotta head out now, I will check back later.

 

[electricpete]

jgrist - I agree 100%. That was my first thought when I read the post. Then I scanned responses and saw you covered it so I added a few things.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[davidbeach]

Sure, generators can serve unbalanced loads; but when running in parallel with another source the power transfer is a function of the two voltages, the angle between them, and the impedance between the two sources.

I've seen very unbalanced generator currents before while parallel with the utility and was able to account for all of it based on utility voltage and angle. This is part of a program where the utility uses customer standby generation and the last time we had enough negative sequence current to trip (standard setting) it was at a location where the phase voltages were off by more than 10V (on a 480V base) and the phase angle separation ranged from less than 116[%deg;] to more than 121[°]. This time, to the resolution of the relay and the other meter, the voltages didn't deviate by more than a volt or two and the phase angle didn't deviate by more than 0.5[°].

Yesterday I had the one event report to look at on a different computer, right now I am downloading all the event reports from the relay to see what there might be of interest from previous runs.

 

[waross]

If possible and easy, you may want to check the open circuit voltage of the generator phases, both phase to phase and phase to neutral. If you see even a small voltage variation it will probably be an indication of a problem with the generator.
Remember that trying to get a correlation between line voltages and currents is analogous to trying to infer the current inequalities in a distribution system by measuring and comparing the voltages. With the wye/wye, you may be able to work backwards. Calculate the terminal voltage of each phase based on the phase currents and the voltage regulation curve of the generator. See if the calculated voltages are a reasonable match with the measured voltages. Check that no conductors take a different magnetic path than the others. Looping around the wrong side of a large iron support member will create a crude wireless reactor and increase the impedance of the affected conductor(s).
Check for bad connections. All these issues are possible candidates.
I suspect that this is connected to a distribution circuit rather than a direct line to the substation. I have seen too many unequal currents, unequal voltages, and neutral phase shifts to be too worried about a close voltage or current balance on a distribution line. With the YNyn connection I would be surprised any time the currents are exactly equal.
Good Luck
Respectfully

 

[davidbeach]

I'll see what we can find out regarding site conditions next week. The site is primary metered with multiple transformers (could mean that the YNyn transformer mentioned above isn't utility and therefore might be Dyn1) down to utilization voltage. Metering data available in the database ends Wednesday morning, so I looked at Tuesday afternoon and found that the whole site has about 5% current unbalance.

I think we need to get some detailed metering on the load of the transformer that connects the generator to the distribution circuit to see if the load is a badly out of whack as the generator currents were.

 

[jghrist]

I wouldn't be surprised by an 8-9% phase unbalance on the distribution circuit. The unbalance on the generator may be more than that on the circuit because the generator zero-sequence impedance is low. This is a typical problem with YnYn generator transformer connections.

This shouldn't be the cause of a neg-sequence trip, however. Usually, this causes problems with zero-sequence tripping. Without the current phase angles, you can't calculate the load sequence currents. Are these phase angles available? What is the setting of the neg-sequence current relay?

 

[davidbeach]

Negative Sequence current was set at 0.6A secondary. This is one of more than a dozen sites, all with the same negative sequence settings. The only other times we have seen negative sequence actually trip was when the bus voltages were way out of whack. I've request that something like an RPM be put on the secondary of the service transformer next week to see what the load current looks like without the generator. If nothing else, the continuous standing negative sequence current isn't good for the generator.

 

[catserveng]

We are seeing this more and more on newer generators, and the responses from manufacturer when asked about these issues has not provided many answers.

We are seeing relatively small voltage imbalances equate to fairly large current imbalances with units in parallel to larger units or to the grid. We see the most problems in utility systems that are having caoacity problems, and using DG units to provide system relief.

I don't have answers, as a lowly field technician I can only tell you this is becoming commonplace, and I think it is likely related to the reduced amount of iron and copper for a given kVA rating, plus the fact that system imbalance seems more pronounced than we saw 10 years ago.

The 0.6 amp secondary setting for negative sequence seems high based on recent experience, we are seeing operational problems at 7-8% negative sequence currents, mostly notable increases in excitation levels and stator temps.

I deal with one manufacturer mostly, but at a recent EGSA meeting, seems we aren't the only ones seeing these problems.

 

[ScottyUK]

I'm surprised you see high stator temps as a result of high negative sequence currents - the NPS relay is designed to protect the rotor where induced current at double line frequency circulates in the rotor forging where it generates heat. If the stator temps really are picking up the effects of NPS heating then the rotor is probably in trouble.


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

Sometimes I only open my mouth to swap feet...