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 site engineer measured the rotor impedance today with the rotor still inside the stator.

Rotor impedance - 8.07 ohms average (at voltages varying from 10 to 70 V, 50 Hz and the readings are fluctuating from 7.74 ohms to 8.36 ohms). The rotor IR is 5 megohms at 500 V.

They have an earlier rotor impedance value measured at 9.8 - 10.0 ohms when the rotor was outside the stator for an earlier overhaul.

Question - will the rotor impedance change when it is inside the stator ?

I have advised them to thread out the rotor to check both the rotor and the stator. Will a RSO test clearly determine the rotor turn-shorts ?

Muthu

http://www.edison.co.in

 

[edison123]

Thanks Bill. I'll check the data available once I'm at the site. Now the unit is shut down and most of the engineers are enjoying the holidays till tomorrow.

Scotty - What is that IEEE std no. ?

Muthu

http://www.edison.co.in

 

[ScottyUK]

An RSO performed by a skilled operator will pick up most winding faults. The key is the experience of the guy doing the test, and of course it helps to have the results of previous RSO tests for comparison. I'm certainly not in the 'skilled operator' category and make no pretence to be, but I'd do the RSO before contemplating unthreading the rotor.

If you do pull the rotor, recommend that the client fits an air gap search coil into the bore. With the right tools it will provide valuable diagnostic information while the set is running, so things like faults which only become apparent at running speed can be detected.

The IEEE document is actually a book, not a standard: ISBN 0-471-61447-5, $120 from IEEE. On of my employer's better purchasing decisions.


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

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

 

[electricpete]

Question - will the rotor impedance change when it is inside the stator ?

I think it is logical to expect a higher inductance when the rotor is tested inside the stator than when tested alone, because the stator iron provides a low reluctance path for flux to complete the flux loop outside the rotor. (this rotor was apparently measured lower inside the stator)

Let's say we used the same current level for testing rotor alone and rotor within stator. That constant current provides a constant mmf. When we add the stator, the reluctance is lower and Flux = mmf/Reluctance is higher. V=N*dPhi/dt is higher.
L ~ V/I is higher.

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

 

[electricpete]

On the surface the low reading and varying reading appear to cofirm the suspicion of shorted turns. But I'm sure you are proceeding cautiously considering that someone different did the test...

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

 

[electricpete]

I know RSO is a standard test as was mentioned.

Aditya has provided data on maintenanceforums.com showing that the BJM AllTestPro tests with varying frequency (I/f, Fi/f)can be very good for finding rotor shorts (more sensitive than power frequency test). I think it is best when a comparison is made among identical winding sections. I'm not sure if that is possible with your 2 pole.

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

 

[electricpete]

I think it is logical to expect a higher inductance when the rotor is tested inside the stator than when tested alone, because the stator iron provides a low reluctance path for flux to complete the flux loop outside the rotor. (this rotor was apparently measured lower inside the stator)

Let's say we used the same current level for testing rotor alone and rotor within stator. That constant current provides a constant mmf. When we add the stator, the reluctance is lower and Flux = mmf/Reluctance is higher. V=N*dPhi/dt is higher.
L ~ V/I is higher.

A simpler way to explain this: Take a core wound around a hollow plastic tube. Inductance is low. Insert an iron core. Inductance is much higher. So in general adding iron increases inductance.

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

 

[electricpete]

Correction:
"Take a core wound around a hollow plastic tube..."
should have been:
"Take a coil wound around a hollow plastic tube.."

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

 

[edison123]

Thanks Scotty. I advised them to pull the rotor out since they had already removed the top end shields on the both sides and both side bearings were inspected. And the unit was due for a capital overhaul since the last one was done about 4 years back.

Thanks pete. I agree. My 'intuition' too says that the rotor impedance should go up when it is inside the stator. I base this on the fact the current drawn by an empty stator is much more than when the rotor is inside.

Could you possibly post that maintenanceforums link on RSO and BJM alltestpro ?

Muthu

http://www.edison.co.in

 

[electricpete]

Here is the thread in question

http://maintenanceforums.com/eve/forums/a/tpc/f/7161085912/m/5731057063/p/2

See Aditya's post 8 April 2008 08:23 AM with attachment labeled "11-KBX-203" and his followup discussion 10 April 2008 05:34 AM

Attached is the test data from that thread. Look at page 12/26 section labeled "15 – Winding Circuit Analysis of New Poles Before Installation (All Test AT-31)" It is a series of tests of new salient poles 1 thru 30 (poles are the vertical axis of the table) for varying frequencies (frequency is the horizontal axis of the table). Poles 2 and 8 have an anomaly which I believe is a shorted turn later agreed by the pole OEM based on the discussion.

The difference between poles 2/8 compared to the remaining poles is very small at lower frequency, but becomes very obvious at higher frequency.

I have a simplistic theory why the higher frequency test can be more sensitive:

R ~ N
L ~ N^2
Z = sqrt(R^2 + w^2*L^2)
Where N is total number of turns.

From examination of R and L relationships, we expect Z ~ N^m where m is between 1 and 2. m will be 1 for dc where resistance R dominates and will be 2 at very high frequency where inductance L dominates the impedance.

Now assume we start with initial turns N0 and we "add" DN turns (DN is a negative number in case of shorting) to bring the total turns to NF (NF = N0 + DN)

Where postscripts are as follows:
F = final
0 = initial (base case)
D = delta (change)
(these same postscripts will be applied to R, L, Z, N throughout the post)

Look at the quantity Z ~ N^m.
How does Z vary when N changes?
ZF/ Z0 = (NF / N0)^m = [(N0 + DN)/N0]^m = (1 + DN/N0)^m

Now let us assume that DN is small in relation to the N0 (this is the only interesting case for looking at sensitivity.... if DN is sizeable fraction of N0 there is no question we can find it with any of the methods).

In that case we can simplify/approximate:
(1 + DN/N0)^m ~ 1 + m * DN/N0 for DN/N0 << 1
ZF / Z0 ~ (1 + DN/N0)^m = 1 + m * DN/N0
DZ/Z0 = (ZF-Z0)/Z0 = ZF/Z0 -1 = 1 + m * DN/N0 – 1 = m * DN/N0
DZ/Z0 = m* DN/ N0
Although it might look complicated, this is simply the relationship between fractional change in Z and fractional change in N. Z is expected to increase by m times more than N.

Conclusion, for detecting a small number of shorted turns:
for dc m=1, the test is least sensitive
for high frequency ac m=2, the test is most sensitive
for power frequency, 1<m<2, the sensitivity is in between

**Note that the above analysis has neglected some effects that can occur at very high frequency: skin effect is one (depends on frequency and conductor size). Another thing is that the core losses can increase the apparent resistance far above Rdc as we increase frequency. The All Test brings additional variables based on phase of the impedance and change in behavior vs frequency into the analysis... I don't fully understand the interpretations

Note So there you have it – supporting test data and a half-baked theory. From what I gather Aditya has a lot of experience testing generators and their rotors, and thinks that is one of the big benefits of the All Test Pro – detecting shorted rotor turns. Although I don't think he uses it alone, he combines it with traditional power frequency ac test. (Maybe he will comment instead of me putting words in his mouth) And it should be noted the actual measurement accuracy of ATP can be fairly low as has been the subject of much discussion on that forum.

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

 

[electricpete]

I think it is logical to expect a higher inductance [ac impedance] when the rotor is tested inside the stator than when tested alone

There is an ASSUMPTION built into the above prediction.

The assumption is that the stator is wye connected and there are no grounds or shorts across the line terminals.

I'm pretty sure a generator this size would be wye, not delta, so probably no problem there.

However if there were safety grounds hung on the terminals, then it is like the rotor is a transformer primary and the stator is like a shorted transformer secondary. This could cause the ac impedance to decrease below the value that would be seen when testing rotor alone.

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

 

[edison123]

Thx pete for that mf interesting link. Need time to go through all the interesting posts.

Could you tell me that the stator has to be connected in wye for the rotor to show impedance? I regularly test delta wound stators with and without the rotor and the current is exceedingly high when the rotor is not inside the stator due to lack of back-emf.

Muthu

http://www.edison.co.in

 

[edison123]

.. tell me why the stator has to be ....

Muthu

http://www.edison.co.in

 

[electricpete]

Comparing rotor impedance measurement with and without stator..

If the stator is connected in wye (and there are no grounds or short circuits connected), then the only relevant stator effect is the iron and the answer is easy. That it what I initially focused on - adding iron makes inductance (and therefore impedance) go up.

But if the stator is connected in delta (or the stator is wye with grounded or shorted terminals), then in addition to the effects of stator iron, we need to also consider the effects of possible induced stator winding currents. IF (*) stator current were induced, that would cause a decrease in impedance and we have two effects acting in opposite direction... it's not obvious which one will win.

* Will there be currents induced? I didn't think about it carefully before because I didn't think it was likely a 60MW machine would be delta connected. With a little more thought, I suspect if the rotor is aligned with A phase, when peak A phase voltage is induced, B and C phase will each see half the voltage with the opposite polarity, net voltage around the delta loop would be 0. So it probably does not make a big difference.

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

 

[electricpete]

Actually the delta is a little challenge to predict... we need to judge to what extent are the 3 phases excited in a balanced fashion to know if the voltages will cancel.

The wye with shorted terminals (such as safety grounds) is easier. If rotor field creates max + voltage at A at a given moment in time, any voltage in b and c at that time would be negative (which increases the voltage difference to A phase). It doesn't matter balanced or not, there would be current induced.

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

 

[electricpete]

... to what extent are the 3 phases excited in a balanced fashion ..

by "balanced" I meant "voltages sum to zero".

For example if I energized the entire rotor with normal connections from single phase ac supply, we would expect voltages at any point in time to be a snapshot of a three-phase symmetric balanced set.... and therefore sum to zero. For example we would expect voltages something like:
Va = 1*sin(w*t)
Vb = -0.5*sin(w*t)
Vc = -0.5*sin(w*t)
The voltages satisfy a sum-to-zero balance.

Whereas if we energize only a portion of the rotor during the test, the flux distribution is more of an unknown, and we don't know if sum-to-zero voltage will apply.

The distinction is important for predicting circulating currents in a delta since the driving total voltage is deviation of the sum of voltages from 0.

The distinction is not so important for prediciting circulating currents in wye with safety grounds on all 3 phases...we will have induced currents whether or not the voltages sum to zero.

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