Generator temperature 1

[sngpl]

We have five horizontal, 3.3 MW each, air-cooled hydroelectric generators. Generator 2 has started to trip on high stator winding temperature, while running at full load. Its worth mentioning that this generator was rewound (not by the original manufacturer) a couple of years because its was damaged due to a fire incident. My questions are:

1- would the generator capability curves change after rewinding the generator?
2- is it possible that the unit got derated after rewinding?
3- I want to measure the resistance of the stator RTDs for calculating the temperature of the winding and also at the same measuring the actual stator winding temperature with an infra red heat sensor. This is to ascertain that the RTDs are functioning properly. If I disconnect the RTDs would it trip the generator? Because I think the protection system would see a very high resistance i.e. the open terminals of the RTD.
4- any further suggestions?

 

[GusD]

Hi sngpl

Interesting post.
I'll try to answer your questions in the order you post'em.
1-Generator capability shouldn't have changed after the new rewind.
Caveat-Assuming that they did not use excessive temperature to burn the old wiring.Temperature in excess of 600 Deg will damage the insulation in the core laminations.
2-the first answer applies to this question as well.No derating should be expected if normal rewinding practices were followed.
3-Unless your generator is of the open coil type,your Infrared camera won't do much good.As for disconnecting your RTDs to measure resistance,that should not be a problem if you bypass the alarms or trips momentarily.
Depending on the type of RTD you use,its resistance should equal the Temp of the winding within a few degrees.Normally there is some variation on the RTDs measurements depending where they are located on the winding.We use mostly 6- 100 Ohms Platinum RTDs in the windings and normally use the Hot RTD for alarm trip purposes.The RTD tables show that they should read 100 Ohms at 0 degrees .If you have Copper or Nickel or a combination of them ,than you would have to find the appropriate tables.They should be available from the RTD vendors or Books.
4- For sugestions, I would like to know if you have excluded the possibility of impaired cooling of the generator due to dust build up or other.(Should be checked)

A very important aspect of your problem is your mention of a fire.Could this fire that caused motor failure,created high enough a tenperature to cook core insulation.If that was the case ,than it could explain your lower HP output.(overheating).
I am curious to know if they performed a Core loss Test when the shop recieved the Gen. and after it was rewound.
Those values would tell you if there was a problem or not.

Hope it helps sngpl



GusD

 

[sngpl]

Hi GusD (Electrical)

Thanks a million for your reply.
First of all I am pretty sure the proper core repairs were carried out before putting in the new winding. I 'll have to dig into the reports and shall get back to you soon.

If we remove the side louver covers then the end turns of the stator winding can be shot at with the help of the infrared gun.

Regarding impaired cooling: as I have mentioned above that basically these are air-cooled generators and this particular generator is cleaner than the others because it was rewound recently.

Further questions:

What do you think could be the cause of this abnormal rise in temperatures?

WE have started running the unit at 2 MWs with unity power factor.

Thanks again.

 

[DanDel]

sngpl, what PF and load were you running the generator at the time of the trip?

 

[sngpl]

DanDel

The load on the unit was 3.2 MW with 0.97 lagging pf.

 

[electricpete]

You have a good idea to try to confirm the temperature reading. Although it may be difficult to draw conclusions from infrared. I would recommend that you measure all 6 winding temepratures on this machine and compare to your other machines.

If just one of 6 is abnormally high then it is either an indication problem or much less likely a very localized winding overheating. In that case I would also take direct resistance measurements at the rtd's to rule out instrument error. But still could be an intermittent loose connection causing high temperature.

Some other wandering thoughts:

Review of core loss test is a good idea, could have been overheated during stripping or during the fire.. Best if you can compare it to a previous core loss test either before winding stripping, or even better during manufacture.

If you are really lucky, a load test would have been done following rewind which would be great info to review.

I would not assume airflow is good just because it is recently rewound. I have seen a thick varnish applied after rewind with B-stage coils. It was so thick that it dramatically reduced the size of the air passages. Also might be good to double-check your filters. And although I don’t expect you’ll see anything noticeable, you might want to feel the velocity of exit airstreams to compare among machines.

If the info could be easily obtained, some interesting comparisons might be:
Compare air inlet temperature for this machine and others. Should be the same unless there is something heating up the inlet air for this machine…. But you never know.
Is unloaded temperature of this machine higher than unloaded temperature of other machines?
Is the mechanical input of this machine measurably higher in the unloaded (but excited) condition than other machines? (probably would be tough to measure)

Would be interesting to know what is the setpoint?
And was the temperature recorded at time of the trip?
And what is the design temperature rise or insulation class on the nameplate. It could be that your setpoint does not leave you much margin. How close do the other machines run to the setpoint?

Find out as much as you can from folks involved in the rewind. Did they try to recreate original design. If they changed something there is more possibility they intreoduced a problem. One other thing is also possible (although not a good practice) they may have tried to save money by designing a machine with less copper and higher temerature insulation.

 

[edison123]

sngpl,
when you say

"Its worth mentioning that this generator was rewound (not by the original manufacturer) a couple of years because it was damaged due to a fire incident", do you mean

(a) " the fire accident was couple of years back, you rewound the m/c then and it was running alright till you noticed this temp rise of late " or

(b) "the fire accident was couple of years back, you rewound the m/c only now and you noticed this temp immediately after rewind"

If it is (a), then check whether your RTD scanner is working properly. I recently saw a 20 MW Gen with winding temp as high as 110 deg C in the scanner readout when the actual resistance measurement of RTD's by a DMM showed only 85 deg c. This was seen in all the six channels of the scanner. The Instrumentation guys checked and said there was some drift in the scanner (whatever that is) which affected all the readings shown by the readout. So, check the actual resistance of your RTD's and convert to the temp using the RTD temp vs resistance chart. This can be done online without anxiety of tripping the m/c.

If it is (b), then your rewinder could have made a mess of the rewinding.

 

[jbartos]

Suggestions to sngpl (Electrical) Jul 21, 2003 marked ///\\
1- would the generator capability curves change after rewinding the generator?
///Yes, they change, at least marginally. Manufacturing tolerances do exist.\\2- is it possible that the unit got derated after rewinding?
///Yes, it might have, at least marginally.\\3- I want to measure the resistance of the stator RTDs for calculating the temperature of the winding and also at the same measuring the actual stator winding temperature with an infra red heat sensor. This is to ascertain that the RTDs are functioning properly. If I disconnect the RTDs would it trip the generator?
///It depends. Some RTDs are designed to annunciate others to trip. If you happen to have RTDs annunciating only, a quick bypass of RTDs is feasible; however, it will pose some risk.\\ Because I think the protection system would see a very high resistance i.e. the open terminals of the RTD.
4- any further suggestions?
///What is the rated value of the generator Power Factor on the nameplate? Normally, the generator high power factor operation is approaching the generator capability limits and should be avoided, if possible, at least on continuous basis. There might be some margin of the generator output power manufacturing tolerance. By keeping on running generator at PF=1 the generator may actually trip.\\

 

[sngpl]

electricipt
1- The core must have been damaged during the fire because I can see hot spots on the outer frame (in the pictures…….. because I was not there at that time).
2- The core has the same laminations... I think the core was restacked after cleaning it and reapplying insulation to it.
I am not sure if the core laminations, once over heated in the fire, would have different characteristics than the new core? ?
3- What exactly to look for in the load test after the rewind (if I can lay my hands on it)??
4- Unloaded temperature of the machine is not higher than the other machines.
5- How would you recommend calculating the mechanical input of the machine??
4- The only noticeable difference between this unit and the others is that its phase winding resistance is about 4 to 5 % higher than the others. I am not sure if this is because of slightly reduced cross sectional area of the winding or due to an additional length of a conductor.

edison123
a) is true.

jabratos
Could you please explain part 4 a bit further? I thought running a machine at or close to units power factor would eliminate any reactive components and thus help in less heating of the unit.

I haven't done the aforementioned tests on the RTDs yet.
I have also ordered an air velocity meter for measuring the intake and out let air velocities and temperatures on this particular unit for comparison with the other units

Thanks

 

[jbartos]

Suggestion to the previous posting marked ///\\Could you please explain part 4 a bit further? I thought running a machine at or close to units power factor would eliminate any reactive components
///True.\\ and thus help in less heating of the unit.
///Not true. By the elimination of the reactive components, the impedance to the generator internal voltage becomes smaller, the current in the circuit increases and RI**2 in Watts increases.\\

 

[sngpl]

jbartos

Is such a simple and precise answer..... thanks!!

So what would be the perfect pf for running the generator?

Can you recommend some good web links to help me understand this problem a bit further?

 

[electricpete]

"I am not sure if the core laminations, once over heated in the fire, would have different characteristics than the new core? ?"

Yes, heat due to fire or improper winding burnout can destroy the insulation between laminations. That allows more eddy currents to flow within the core which increases the core loss.

HOWEVER, the fact that your no-load temperature is normal rules out a core problem, because core loss is independent of load. (A core problem would cause hotter machine under no-load as well as full load).

"The only noticeable difference between this unit and the others is that its phase winding resistance is about 4 to 5 % higher than the others"
Assuming you have done your measurements right and corrected for temperature, that definitely indicates there was some kind of redesign, and apparetnly in a direction that would increase stator I^2*R heating temperature.

"What exactly to look for in the load test after the rewind (if I can lay my hands on it)??"
To me this would be an essential key piece of info. As first step, look for the temperature rise at full load during test and compare it to the temperature rise that you see now at full load. (Perhaps you correct for any small differences in power factor and voltage.) It will help you sort out whether the machine is running close to it's performance at time of rewind or whether it has degraded since then. Either way you have narrowed down and better defined the problem.

"How would you recommend calculating the mechanical input of the machine??"
What type of prime mover is it?

Still it would help you to tell us some things that give a picture of how much margin there is to trip and how much this is different than the others:
-what is the winding temperature trip setpoint?
-what is the measured winding temperature of this machine at full load?
-what is the winding temperature of other machines at full load under same conditions (or if cannot get same conditions tell us the conditions voltage and power factor)

 

[ScottyUK]

jbartos,

I'm not sure I follow your logic:

For a generator at constant power output, increasing the reactive load to either a positive or negative value will increase the MVA of the generator. For a constant terminal voltage, this will mean an increase in current. From I^2R, if stator current increases, so do the losses in the stator winding.

You don't mention losses in the rotor. Rotor losses are largely due to I^2R in the field. These losses increase as the field current increases, i.e. the power factor becomes increasingly lagging.

Reducing the reactive load reduces the overall heating of the machine by reducing both the stator and rotor losses.

It would be unusual for a generator to trip at unity power factor. Most units can accept a certain amount of leading power factor loading, as illustrated on the capability curve, but are normally designed to run at about 0.85 lagging.

 

[jbartos]

Suggestion to the previous posting: Visit

http://www.transmission.bpa.gov/orgs/opi/Power_Stability/FldCurPanelGE.pdf

for Figure 1 depicting a generator capability curve.
Notice that at rated PF~85% the abscissa reads ~230MW. Now, at PF=1.0 the abscissa reads about 275MW.
Ratio of 275MW/230MW=1.19565
~20% over the rated generator MW can trip the generator unless the manufacturer provided that much of design margin for MW.

 

[electricpete]

Scotty is right.

Heating tracks much closer with kva than with kw. Along the limit curve between 0.85 and 1.0 the machine has approximately constant kva and approximately constant heating.

In general, a fairer comparison is machine operating at constant power with two different power factors.
1.0 power factor results in lowest heating.
275MW at 1.0 => at the thermal limit
275MW at 0.85 => beyond the thermal limit
There is no logical reason to suggest that higher power factor creates more heating.

 

[jbartos]

Suggestion: The generator nameplate includes the generator rated values. It is built and tested for these values and many parameters are related to these values. The generator is very flexible machine since one may run it according to the capability curve. I used a simple electrical equivalent circuit for my explanation above, which often appear in books, software modeling, etc. namely, the generator internal voltage and generator synchronous impedance. If this model is not correct, why is it so much used?
References:
1. William D. Stevenson, Jr., Elements of Power System Analysis, 3rd Edition, McGraw-Hill, 1975,
Figure 9.4 Circuit diagram for a generator and motor. Ia is the current delivered by the generator and received by the motor. (Page 217)
2. Arthur R. Bergen, Power Systems Analysis, Prentice-Hall, Inc., 1986,
Figure 8.5 Per phase circuit diagram (round-rotor machine). (Page 241)
Obviously, more advanced books use different generator mathematical models; however, these are books for more advanced engineering analysis of generator and systems, e.g.
3. Yao-nan Yu, Electric Power System Dynamics, Academic Press, 1983
4. Charles Concordia, Synchronous Machines Theory and Performance, General Electric Company, 1951

 

[ScottyUK]

Hello jbartos,

My post did state "For a generator at constant power output...", which means that the MW value is a fixed value. The reactive power may be varied without affecting the real power output. Your example went from 230MW to 275MW, which suggests you are looking at the constant MVA curve, not the constant MW line. Normally the prime mover - engine or turbine - isn't able to deliver the full MVA rating of the generator. At our site, the 365MVA generator is only able to deliver 310MW at unity power factor because that is the limit of the turbine. The same generator will deliver 365MVA by supplying 310MW and about 200MVAr. Thus my constant MW assumption is valid for normal operation of the machine.

Define MVA as total volt-amps of the generator
Define MW as real power output of generator
Define MVAr as reactive power output of generator

If we agree that MVA = sqrt(MW^2 + MVAr^2), then we surely must also agree that when MVA = MW, MVAr must equal zero.

When MVA = MW, the power factor is unity. For any condition other than MVAr = 0, MVA must be greater MW. The stator heating is largely governed by I^2R in the stator, and the minimum value for stator I occurs when MVA=MW, i.e. when the power factor is unity. For any other power factor, the stator heating is greater.

The generator model is useful for predicting the behaviour of the machine, but it is worthwhile considering the fairly complex phasor diagram which illustrates how the relationships between the internal voltage E, the actual terminal voltage V, the stator current, and the machine load angle change as the generator MW and MVAr are varied. I think you will gain greater understanding of the machine by constructing such a diagram - there should be an example of a generator pahsor diagram in any half-way decent textbook on electrical machines. Construct it for a lagging, leading, and unity power factor load.

I have an old photocopy of a CEGB guidance note on constructing capability curves. If you have a fax number, I will gladly send you a copy. If lots of people want it, I might get it scanned and e-mail it on request.


Scotty.

 

[plc123]

Hi,
if temperature sensor is OK and positioned properly,
then may be useful to run this generator without load - magnetic flux is the same on / off load. If temperature goes up - then problem with insulation between stator plates or joint bolts installed without insulation tubes.
If not - check internal coil connections - could be bad contact. This contact usually diagnosted as spikes on
the screen.

Thanks,

Vladimir.

 

[edison123]

scottyuk,

could you pls mail me the CEGB guidance note on constructing generator capability curves to eewemf@myrealbox.com? thanks.

 

[jbartos]

Suggestion: Reference:
1. A.E. Fitzgerald, Charles Kingsley, Jr., Stephen D. Umans, "Electric Machinery," 5th Edition, McGraw-Hill, Inc., 1990,
Fig. 5-16 Construction used for the derivation of a synchronous generator capability curve on page 242