Motoring effects on a steam turbine generator 1

[edison123]

A 60 MW, 3000 RPM steam turbine generator went into motoring and the protection system failed to catch it (DC failure). The machine was tripped manually and the it was resynchronized. But the generator vibrations shot up.

The machine was stopped and the generator and the turbine bearings were inspected and found ok. It was then restarted and the vibrations were found to increase with the excitation. On tripping the excitation, the vibration went down immediately obviously indicating electrical issues.

Questions to ScottyUK, epete and other generator aces.

Will the motoring action affect the generator rotor winding ? What about the stator winding and the core ? What are the checks to be done on the generator after such motoring ?

Muthu

http://www.edison.co.in

 

[edison123]

The generator is rated at 11 KV.

Muthu

http://www.edison.co.in

 

[ScottyUK]

Hi Muthu,

Few more questions first:

How long was it motoring for?
What type of exciter: sliprings; rotating rectifier; other?
Did the field remain energised?
Did the condenser break vacuum?
Is the vibration problem immediately apparent or does it build after a period of time?


FWIW I'd pretty much discount any possibility of problems on the stator and core - if there's damage to the generator it will likely be on the rotor or in the auxiliaries.


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

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

 

[edison123]

Thanks Scotty. See below the field report. You think the rotor got buggered ?

Muthu
www.edison.co.in

 

[dpc]

Motoring protection is really there to protect the prime mover, not so much the generator. On a steam turbine, the problem can be overheating in the low pressure section. The generator should not have problems running as a motor as long as it was running within its capability curve. It would be interesting to know if the loss of field protection also tried to trip the unit.

 

[ScottyUK]

If there was no slowing of the unit as stated then the excitation presumably remained available in order for the unit to run as a synchronous motor. I would guess the excitation is inter-tripped as the GCB opens, but each vendor has their own slightly different ways of doing things. That's good because it means the machine didn't run as an induction machine on the amortisseur, which would eventually heat up the rotor.

As dpc has already noted, the reverse power relay is primarily to protect the turbine LP rotor, but if vacuum was maintained - as it appears to have been - then there should not be any danger of the LP blading being damaged by frictional heating due to windage.

How fast does the vibration come up with excitation? A very fast response would suggest assymmetry in the rotor field, while a slower response would suggest a thermal bend. Both could be attributed to a shorted turn or partial winding failure. I'd get the unit down to a stationary condition and megger the rotor at 2x working voltage from winding to the rotor forging, then do an accurate DC resistance measurement and compare with OEM data. You're probably happy enough doing that - only comment I'd make is make sure you know where and how the OEM measurements were taken.

Do you have records of what balance weights were fitted to the machine and in which locations? Check they're all still there.

Have you got any support from the vibration analysis guys or are you on your own with this? Also what data does the machine historian collect? Most DCS and turbine control systems can be persuaded to talk to a Bently Nevada 3300 system (or anything of later vintage) via a data link, and that lets you get a lot more data. Get it added to the historian so you can trend machine parameters alongside vibration.


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

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

 

[edison123]

Thanks dpc and Scotty.

That bomb got lobbed at me today afternoon. There is lot that is left unsaid in that communication to me. I talked with a low level engineer and asked him about the vibration behavior with excitation. He told me that the vibration drops immediately upon switching off the excitation. That and the increasing vibration levels with load seem to indicate some rotor turn shorts.

I have been asked to study the issue including any new vib analysis. They have the machine shut down now. I plan to be at the site this weekend. I plan on doing a megger, IR and AC impedance measurement.

And I will come back here with more questions.

Sorry to bother you during the holidays.



Muthu

http://www.edison.co.in

 

[ScottyUK]

Hey Muthu, don't apologise. You're working in the holidays. Me, I'm going to enjoy Christmas being not on call after having (voluntarily) done eight Christmas Day callout covers in a row. I used to let the guys with young families have undisturbed time with their kids; now it's my turn to be undisturbed.


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

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

 

[davidbeach]

Turn to turn fault or ground fault on the rotor? No field current would mean no asymmetric magnetic field. Rotor over heated due to lack of excitation?But I'd also try to determine whether it is machine or prime mover. The loss of vibration on loss of excitation seems to strongly suggest the machine though.

 

[electricpete]

Yes, definitely no need to aplogize. Anyone reading and responding to your post is doing it because they enjoy it and want to do it, not because you made them do it.

I agree with above comments. Reverse power protection is for the prime mover.

If power was reduced to low level before securing the turbine, there is no reason to suspect any unusual swings of the generator against the system that might create a pole slip. (Even if turbine was at full power and its valve suddenly closed, the reverse angle swing shouldn't exceed the initial angle... nowhere near any stability limits). All this assumes the excitation remained properly controlled.

Similar to the comments above, it is natural to start thinking about whether the excitation system worked properly. Maybe whatever malfunctions occurred with the DCS and the tripping functions may have also affected the excitation control system and loss of excitation protection (?). For example voltage signal fed from DCS to AVR ?

Both underexcitation and overexcitation are problems. Overexcitation - overheating of rotor conductors based on I^2*R and heating of stator iron based on core losses. Underexcitation/loss of excitation ... as mentioned a loss of excitation turns it into an induction motor or generator. It was mentioned it is a problem for armortisseur windings... I think it can also be a big problem for the rotor iron and retaining ring in a high-speed smooth rotor ...rotor teeth have no laminations and act like rotor bars of a SCIM, the retaining rings act like end rings of a SCIM. For smooth rotor generator, loss of field culd also create heating of stator end-iron due to axial leakge flux when retaining ring drops out of saturation and leakage between stator end iron and rotor increases.

For that matter, the DCS malfunction may have also created problems for temperature control of hydrogen / stator cooling water, seal oil (?)

By the way, the units of vibration I assume are microns? Why are there two numbers per location per direction?

There is no obvious smoking gun or target to go after that I can see. Best keep an open mind going in there.

You know better than me what and how to test.


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

 

[edison123]

Thanks pete. The vibrations shown are in microns/ mm per sec. I also suspect that the machine lost the excitation and probably as induction generator (a nightmare scenario).

I will come back next week with more details.

Muthu

http://www.edison.co.in

 

[ScottyUK]

Edison,

Running as an induction machine is definitely something to look for. The report said the unit didn't slow down, but would be interesting to see if it did actually drop to a sub-synchronous speed. Look at the data from the machine historian, but do not treat the data as gospel truth: sometimes sampling periods and deadbands can make things appear to happen (or not happen), and also to change the apparent sequence they occured in.

ePete,

Yes, agreed about effect on teeth of high speed rotor. Should have mentioned that myself - thanks for the good explanation. How serious would stator end heating when the stator currents involved are relatively low and the rotor field is absent?


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

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

 

[waross]

Merry Christmas, or seasons greetings if you prefer.
Looking at the sequence of events:
1> DCS failure. Investigation may reveal more than one function has failed.
2> Unit failed to trip.
3> Unit motoring.
4> Steam shut off manually.
5>Breaker tripped manually.

The following issues are suggested:
Multiple DCS failures.
The excitation did trip but the steam and breaker trips failed.
Following the loss of excitation the unit motored as an induction motor for long enough to damage the rotor windings.
It may be well to review the configuration of the DCS safety routines, both the software and the hardware implementations.

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

 

[rbulsara]

Bill's thinking appears most plausible.




Rafiq Bulsara

http://www.srengineersct.com

 

[waross]

Thanks Rafiq.
"....I have stood on the shoulders of giants."
An attempt to summarize the posts of those who preceded me.
Yours
Bill

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

 

[rbulsara]

Bill:
That in itself is an important skill and virtue of a consultant, sifting through the pile of information and zeroing on the important ones.

Regards,
Rafiq

Rafiq Bulsara

http://www.srengineersct.com

 

[electricpete]

How serious would stator end heating when the stator currents involved are relatively low and the rotor field is absent?

First I will mention that I'm a motor guy, not a generator guy (in spite of edison's kind words). I'm sure you and others here know more about generators than me. But I believe (with at least 51% confidence) that there would be a concern for heating at stator endcore if the stator remained connected/energized at no load for extended period upon loss of field. Even if that's true, how severe in 2 minutes? Probably not severe since steady state P-Q limits are applied on underexcited operation to address long term effects. And how likely that stator damage would case vibration? Not as likely as rotor damage. So it's probably among the lower probability items to investigate, and a good clarification.

My simplistic viewpoint starts with the P vs Q generator capability curve. We have stator core end overheating possible when leading Q is more than a limit (even when P=0). Low field current is the more severe condition. Zero field current would seem to be an extreme of lowering the field current... although the curve was clearly not intended to cover that scenario so it may not be correct logic.

Beyond trying to extrapolate the P-Q capability curve into new territory, we try to look for the physical mechanism which I certainly don't completely understand. An excerpt from "Protective Relaying for Power Generation Systems" (go to pages 254 and 255)

http://books.google.com/books?id=i9... heating in the round rotor generator&f=false

Out of that reference I get two concepts:
1 – The reluctance path for endleakage flux is affected by flux level in the retaining ring. Low excitation decreases reluctance and increases axial endleakage. (2nd full paragraph on page 255)
2 – The mmf driving flux through that reluctance (of retaining ring in series with air) is assumed by this reference roughly proportional to to airgap flux (last 2 paragraphs on page 255).

Which of above 2 effects is more important? I vote the 1st one since retaining ring is in a strategic location and it's reluctance can change by many orders of magnitude. In contrast the airgap flux and mmf does not change by much when connected to a constant voltage system (only the voltage drop accross the synchronous stator reactance). So far, it seems to support my conclusion about effect of long-term operation.

You'll note whenever this references cites the effects of underexcited operation, they refer only to retaining ring reluctance and airgap flux. They never talk about relative phase of the respective leakage flux from rotor end connections and stator endturns, which I have heard mentioned as an aggravating factor in some other references. The phase relationship is easily understood by looking at a stationary location in the airgap: rotor flux in phase with rotor current, airgap voltage lags rotor flux by 90 degrees (by Faraday's law), and assuming small stator leakage reactance, stator current leads airgap flux by 90 degrees for a pure capacitive load (P=0, Q<0), and therefore stator current in phase with rotor current for pure capacitive load, so leakage fluxes add at a given location. This aggravating effect is completely absent in the loss of field scenario as I think Scotty was pointing out. I don't know the relative importance of all these factors compared to each other.


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

 

[electricpete]

typo correction:
"stator current leads airgap flux by 90 degrees for a pure capacitive load"
should have been
"stator current leads airgap voltage by 90 degrees for a pure capacitive load"

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

 

[ScottyUK]

ePete,

I've quoted the following which supports your suggestion from "Operation & Maintenance of Large Turbo-Generators" - published by the IEEE, it is a useful reference book.

Unexcited Operation (‘‘Loss-of-Field’’ Condition)

Operation without field current is potentially dangerous and can occur under a number of circumstances. The following are the most common two:

Loss of field during operation. If for some reason the field current goes to zero while the generator is connected to the system, the machine starts acting as an induction generator. The rotor operates at a speed slightly higher than synchronous speed and slip-frequency currents are developed. These penetrate deep into the rotor body because they are of low frequency (this does not represent the skin effect discussed in case 2, below). Severe arcing between rotor components and heavy heating may result. The ends of the stator core also experience heating due to stray fluxes in the end region, more severely than for operation at underexcited power factor. Protection is commonly provided to prevent or minimize the duration of this mode of operation, by the so-called loss-of-field relay.

<The other case was inadvertent energisation, which wasn't germane to this discussion>

There's also a useful diagram on page 46 of this book which is worth studying to better understand the stray flux problem and why it occurs at leading power factors.


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

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

 

[electricpete]

Thanks Scotty, that's helpful info to me.

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