Surge test on generator field coil

[motorspert]

Hi
We have been asked to perform a surge test on some salient pole field coils at 1500V. I am familiar with voltage drops using an AC supply at 50/60hz. Also aware of tests at higher frequencies (400Hz). But should I be concerned that user a standard surge tester (Baker type) we will overstress the first turns? The coils are single layer strap type copper and the test wil be done on the rotor.

We will perform some tests at lower voltage to see the distribution, but interested if others have any experience.

Many thanks

 

[electricpete]

But should I be concerned that user a standard surge tester (Baker type) we will overstress the first turns? The coils are single layer strap type copper and the test wil be done on the rotor.

I don't have experience with doing the test on generator field coils, but I think similar considrations apply as to motor stators (my experience). Based on that, my comments:

Surge test (IEEE522) is a potentially destructive test. Accordingly, attention should be paid to the proper selection of test voltage, considering the condition of the coil insulation (cured or uncured), stage of assembly, etc. Also (like a hi-pot) you wouldn’t test unless you had good IR/PI readings. The traditional assumption is that if the test passes, no damage has been done.


=====================================
(2B)+(2B)' ?

 

[jpts]

When you have a voltage pulse with very fast rise time, voltage drop over first turn is much higher than for the rest turns. If the turn insulation is not designed for this kind of voltage, you might stress the turn insulation too much. You can have even 50..70% of the applied voltage over first turn if the pulse if fast enough. Maybe you can check is that too much for the insulation before doing the test.

If you could test the coils when they are dis-assembled from the rotor (no ferromagnetic material around the coils), then you would have more or less linear voltage distribution over turns.

 

[electricpete]

When you have a voltage pulse with very fast rise time, voltage drop over first turn is much higher than for the rest turns. If the turn insulation is not designed for this kind of voltage, you might stress the turn insulation too much.

You can have even 50..70% of the applied voltage over first turn if the pulse if fast enough. Maybe you can check is that too much for the insulation before doing the test.

Also that is what I was referring to by “proper selection of test voltage”. The shape of the pulse (how long to reach the peak) is fixed by standards (IEEE522). We select the peak magnitude which determines the voltage rate of change (for a fixed shape) and the associated turn insulation stresses.


=====================================
(2B)+(2B)' ?

 

[electricpete]

Now that I think about it, motorspert probably knows all about testing stators so I’m sure I haven’t told him anything new.

I guess the real question would be how much voltage can be safely applied… would we use the same type thumbrules based on machine voltage rating as used for ac stators? Beats me. I would be a little nervous about that knowing that the pulse may act a little differently. Would look for some standards and experience to guide me. I have none to share though - Sorry. In retrospect I probably haven’t said anything useful for motorspert.


=====================================
(2B)+(2B)' ?

 

[motorspert]

Thanks guys
All useful info. We will do them away from the rotor, so that sounds like it could work.

I share your nervousness electricpete but am also keen to experiment! So, we will take it slowly. Luckily it's bare copper so we can put a scope on pretty easily.

 

[ScottyUK]

Have you considered an RSO test? Much lower voltage than a Baker test, so less chance of damage. This link gives some background if you haven't used it before. It is well-known in British plants but doesn't seem as well-known in some countries. It doesn require some experience in interpreting the results, especially for the less obvious faults.

 

[electricpete]

That’s a good point. You got me thinking. I think I have an idea of the reason for different strategies in turn testing on ac stators and dc windings.

There is a fundamental difference between RSO and surge test in that RSO detects shorted turns and surge test detects weak turn insulation (insulation that might fail under stress in the future but is currently bad). So, surge test is a much more senstive test, and necessarily is also potentially destructive to accomlish that sensitivity.

Why do we tend to use RSO test on generator rotor (only detects shorts) but surge test for ac machine stators?

I think it ties to the consequence of turn to turn short.

On an ac stator, as soon as turn to turn short occurs, there is immediate high circulating current due to auto-transformer effect. It will melt copper and quickly escalate to a ground or phase fault causing a machine trip. There is no benefit to RSO test because if you have shorted winding your machine has most likely already tripped. But there is advantage to determining weak turn insulation because it can result in trip almost immediately if it does short in the future.

In contrast on a dc winding, turn to turn short results in practically no change in current (perhaps current increases in inverse proportion to the fraction of total turns shorted).. The machine can operate for long period of time with that short present with little effect (often only increase in vib is the most noticeable effect). We can use RSO to detect/confirm that shorted winding. There is no need for invasive / potentially destructive surge test because turn short is not as severe a fault on dc winding as it is on ac winding (we have time to respond).


=====================================
(2B)+(2B)' ?

 

[electricpete]

must have... edit ... button...

There is a fundamental difference between RSO and surge test in that RSO detects shorted turns and surge test detects weak turn insulation (insulation that might fail under stress in the future but is currently bad).

should've obviously been:

There is a fundamental difference between RSO and surge test in that RSO detects shorted turns and surge test detects weak turn insulation (insulation that might fail under stress in the future but is currently not shorted).

=====================================
(2B)+(2B)' ?

 

[ScottyUK]

Nice summary ePete. Shorted turns usually result in uneven heating of the rotor and a resultant thermal bend which shows up as a change in the vibration signature. Sometimes the vibration goes down, which can confuse things!

It would help to understand what the OP is trying to diagnose - insulation strength I guess as a surge test was proposed initially.

 

[petronila]

Hello motorspert,

I am familiarized with this kind of test and I.M,O the dc surge test is not destructive. Note that usually the field voltage is low and the D.C. test voltage is 1.7 (2 V + 1000 ) so if the machine is new you can test at aprox. 1700 to 1800 Volts (considering that field winding voltage don´t exceeds 110 V) if the windingb is not new 1500 volts will be high (I will test it with 1000 V), our rule of tumb is perform surge in old windings at 60% of new voltage test.

Some considertaions about the field test (salient pole rotor):

If the machine is assambled then you can perform a resistance test and compare with ratio V/A, this reading will give you an idea about the winding condition. Another consideration is if you perform the test to conected field winding unless you have severe short circuit, you are not going to find nothing significative. Is possible that the field winding pass but really you are not making a comparisson (the winding is single phase). In this type of salient poles the best way for testing is disconecting the coils by pairs and compare one with the adyacent (also you can compare the ohmic resistance of each pair of coils), to do this coil comparison follow tester manufacturing direction for single phase windings.

Regards

Carlos

 

[motorspert]

@petronila - thanks for your comments, thats's useful information - they are new coils and this may be a test that some OEM's perform. I will let you know how it goes.

@scottyuk et al - we are just proof testing a new set of coils, we normally test for interturn faults bt applying 220V 400Hz, this confirms the coils are ok, but the coils are only effectivelty strength tested at 0.4V/turn. This customer wamnts a more demanding test.