Impending Generator Failure

Status
Not open for further replies.

big john

Senior Member
Location
Portland, ME
Have a 12MW enclosed generator that is suddenly producing a very significant "hot electrical" smell.

All instruments are reporting "normal" values for temperature, current, voltage, etc. No protection has operated.

Did winding resistance and insulation resistance on both windings, plus surge on the stator. The only unexplainable reading is the rotor resistance was about 25% low based on nameplate voltage and current. The two caveats are the resistance is NOT temperature corrected, and this is a 28 pole machine so one shorted pole would not account for that much difference.

The rotor is basically inaccessible. You can barely even see it, so there's no way to do pole-drop.

Any suggestions at all for further testing? Something is going to fail here, and I have no idea how else to find it.
 

Ragin Cajun

Senior Member
Location
Upstate S.C.
If it fails, the damage would likely be beyond repair. If you shut it down and fix it, the repair will likely be possible, and take less time than getting a new one.
 
(thinking aloud, as it were, and I haven't finished my coffee yet)

Load bank test? When the load is known and controlled, if one leg is different from the others it should show in the currents/voltages. Granted that even a 5MW load bank is not small or cheap; OTOH temp rental on a 12 MW backup generator and the teardown isn't either.

Comes down to it though, you either take it apart before it fails (on a schedule) or after (as an emergency), only question is how long to wait.
 

big john

Senior Member
Location
Portland, ME
This is grid-parallel, so it's really easy to put it under load, but of course the issue is doing that puts it one step closer to failure.

I agree, that this thing is very likely getting disassembled regardless, my concern is collateral damage of running 'till it burns:

There is rotor ground detection, so that might save the bearings.

And there is differential protection on the stator so maybe it wouldn't obliterate itself.

That said, we are talking hundreds of thousands of dollars if this is disassembled and no fault is found, so either way this is an expensive gamble.
 

JFletcher

Senior Member
Location
Williamsburg, VA
super wag here, but if there are no outward signs of genset problems, what about associated gear failure? Is it electric start? A hung starter about to blow up would smell horrendous too.

What kind of genset is it? 12MW is a biggun; even a 20cyl EMD 710 only puts out about 2.5MW.
 

junkhound

Senior Member
Location
Renton, WA
Occupation
EE, power electronics specialty
....25% low based on nameplate voltage and current...

are you accounting for inductance ? All 3 phases 'low'?


oil leak anywhere ? Was there a shutdown period recently - mouse nest, etc?

At what voltage did you run insulation resistance ? What is gen voltage? Have you run a PD test?
 
The only reason I suggested a load bank is that the load is constant and should match fairly closely across all three phases; can you assume that for a grid-tie load (I haven't thought much about it)?

With a constant and known R (load), you can measure I and V of all of phases and see how closely they match. Bonus points for rotating the load connections and testing again. One phase consistently producing a different voltage/current from the others suggests a difference in the windings or connections (like metering motor currents). That said, I don't know what's an acceptable balance in generator currents. Also granted that any testing takes time and costs $$$.
 

big john

Senior Member
Location
Portland, ME
What about associated gear failure?
Have looked for that. Only associated gear in the doghouse is a mess of CTs, and none of them look bad, but obviously that means little. The only reason I'm shying away from that is they all control metering and protection, which is reporting everything is fine.
What kind of genset is it?
Low speed 4160V hydro.
Are you accounting for inductance ? All 3 phases 'low'?
This was the DC field winding test, so minimal inductance. All three stator windings tested balanced and matched historical data.
At what voltage did you run insulation resistance ? What is gen voltage? Have you run a PD test?
4160V gen. 5kVDC stator meg, with a 9.6kV surge test. 500V rotor meg.

No PD because there are safety concerns about being in the doghouse with the energized stator when we don't know what's about to fail.
How is the machine cooled? Could there be a failure of that system?
Air and water. I haven't seen the trends, but have been told no unusual temperatures and they haven't had any temperature alarms or trips.
how much will it cost if it is NOT disassembled, and a fault is found?
That's the gamble. If by some trickery this is an ancillary failure not a winding failure, they might get off easy. At this point all we know is *something* associated with this unit is going south.

Thanks for the replies so far.
 

drcampbell

Senior Member
Location
The Motor City, Michigan USA
Occupation
Registered Professional Engineer
Low rotor resistance might be a result of shorted turns. Is the rotor current higher than expected?
I don't suppose you have an infrared thermography borescope in your toolbox? (I don't even know if such a thing exists -- might be wishful thinking on my part)
 

Besoeker

Senior Member
Location
UK
Do the windings have thermistors or other means of monitoring their temperature? On a machine of that rating I'd be surprised if it didn't and maybe you answered that point already.
 

meternerd

Senior Member
Location
Athol, ID
Occupation
retired water & electric utility electrician, meter/relay tech
Have a 12MW enclosed generator that is suddenly producing a very significant "hot electrical" smell.

All instruments are reporting "normal" values for temperature, current, voltage, etc. No protection has operated.

Did winding resistance and insulation resistance on both windings, plus surge on the stator. The only unexplainable reading is the rotor resistance was about 25% low based on nameplate voltage and current. The two caveats are the resistance is NOT temperature corrected, and this is a 28 pole machine so one shorted pole would not account for that much difference.

The rotor is basically inaccessible. You can barely even see it, so there's no way to do pole-drop.

Any suggestions at all for further testing? Something is going to fail here, and I have no idea how else to find it.

I think using nameplate current and voltage, you'd be calculating impedance, not resistance. I would expect resistance to be quite a bit lower than full speed running impedance. The only way to know if it has changed is to have the original winding resistance. Not commonly supplied with motors. I'd say, if all of the winding resistance values are the same value, you should be looking elsewhere. Could be a winding core lamination problem. Is the airflow path clear of debris? All connections at the starter and the motor pecker head tight and no evidence of being overheated?
 
Last edited:

GoldDigger

Moderator
Staff member
Location
Placerville, CA, USA
Occupation
Retired PV System Designer
Change "motor" to "generator" in the previous post?
This is a wound rotor, hence current and voltage appear on nameplate?

Sent from my XT1585 using Tapatalk
 

big john

Senior Member
Location
Portland, ME
Low rotor resistance might be a result of shorted turns. Is the rotor current higher than expected?
All we can do is compare exciter current draw to MW production for another identical unit, and I don't think that's valid, so I don't really have a way of saying if it's higher than expected.

Could be shorted rotor turns, but I think the different-than-expected resistance is a temperature correction problem: I'm pretty confident a fault has not affected 25% of the existing poles.
Do the windings have thermistors or other means of monitoring their temperature? On a machine of that rating I'd be surprised if it didn't and maybe you answered that point already.
RTDs in the stator backiron. All reading normal.
I think using nameplate current and voltage, you'd be calculating impedance, not resistance.
That was in regards to the DC field winding, as far as I understand that is primarily a resistive circuit, with very little reactance.
...This is a wound rotor, hence current and voltage appear on nameplate?
Synchronous generator, the voltage and current discussed were from the rotor.
 

ATSman

ATSman
Location
San Francisco Bay Area
Occupation
Electrical Engineer/ Electrical Testing & Controls
I think using nameplate current and voltage, you'd be calculating impedance, not resistance. I would expect resistance to be quite a bit lower than full speed running impedance. The only way to know if it has changed is to have the original winding resistance. Not commonly supplied with motors. I'd say, if all of the winding resistance values are the same value, you should be looking elsewhere. Could be a winding core lamination problem. Is the airflow path clear of debris? All connections at the starter and the motor pecker head tight and no evidence of being overheated?


Big John,
Have you addressed this question, if so I missed it.
 
Status
Not open for further replies.
Top