Abnormally high current during transformer turns ratio test 1

[Rwelagedara]

Dear All,
We are performing a transformer ratio test on a 1 MVA,33kV/400V transformer using a three phase transformer ration meter from a reputed manufacturer.The ratio meter is also showing the current (mA) in each phase in addition to the turns ratio value. During the testing, we found out that the ratio values are in acceptable range in all three phases. But current in one phase(phase A) is abnormally high. This high current is around 120 mA, whereas the current on other two phases is just around 12 mA. Is this an indication of an issue . We have tested many transformers before but this is the first time we see this kind of a very high current in one phase. Please assist.

 

[RedSnake]

Have you tried to change place (phase) (rotate) the delivery/measuring cords to find out if it is the instrument or the transformer?
/A

“Logic will get you from A to Z; imagination will get you everywhere.“
Albert Einstein

 

[Rwelagedara]

Dear RedSanke,
Thanks for pointing out that. Yes we changed the probe and still the error is in the same phase of transformer. So we are sure that the issue is not with ratio meter.

 

[RRaghunath]

If ratios are fine in all three phases, I feel it is nothing to do with the transformer.
12mA in B&C phases Vs. 120mA in A-phase suggests there may be a factor of 10 somewhere in the meter or CT or something else!!!

 

[RedSnake]

I am not sure what is possible to measure on your transformator, I haven't myself worked with transformators that big.

But if it is possible I would measure the resistans over each winding separately.
A-B
A-C
B-C
they should all have the same resistans
and then
N-a
N-b
N-c
and they should also have the same resistans.

Best Regards Anna



“Logic will get you from A to Z; imagination will get you everywhere.“
Albert Einstein

 

[Palletjack]

Are these leads marked H1,H2,X1,X2?
Or are they just red and black leads?
I assume this is an electronic unit vs a hand crank unit.

So big black on X0, big red on X1, little black on H3, little red on H1?

 

[edison123]

Do a surge comparison (Baker) test. It will show shorted turns nicely.

Muthu

http://www.edison.co.in

 

[Rwelagedara]

Thanks all.
Palletjack , the ratio meter is a digital electronic one from Raytech. The markings are very clear.
There are two similar transformers.And this deviation in current is observed in only one transformer.
We measured the DC resistances too, but they are in order.

 

[Palletjack]

I have a lot of questions. With what you’ve given so far I wouldn’t energize quite yet..
It could smoke and go boom


Are there any tap changers on this transformer? Can you energize it with a lower voltage, say 480 and check it out?
Closed loop/short somewhere in the windings?
How new is it? Has it ever been energized before?
Is this a commissioning test?

Try a different test set also just to confirm XF or test set validity.

 

[prc]

A turn to turn short is to be suspected in A phase. To confirm, conduct single phase excitation current measurement at 10 kV as explained in C57.152 standard.

 

[prc]

Being only 400 V transformer, only 230 V is required instead of 10 kV.

 

[electricpete]

It's been a long time since I tested transformers.

My first reaction would be to trust anything prc says about transformers.

I was wondering how a shorted turn that is not significant enough to show on TTR but significant enough to cause a 10x increase in magnetizing current. My simple first thought was treating it as inductor, removing a small fraction of the total turns doesn’t change the inductance very much (L~N^2). But then thinking some more, I suspect the shorting turns creates a loop which allows current to flow around that loop, which increases current in the rest of the winding (similar to wrapping a shorted winding on the same core... it presents a load which causes a corresponding load current in the energized winding). So yes prc's comments make sense to me now.

Another thought I had was about the magnetic circuit. If the core had some gaps at a joint in one leg, the turns ratio would be unaffected but the reluctance would go down and the exciting current for a given voltage would go up in that leg. I can't say I've ever heard of that in real life, but from my understanding of the theory it would match the symptoms. If anyone wants to shoot that down, please do (I won't be offended).

Is it a new transformer or was the motor previously in service?
If previously in service, did it trip or just tested for routine testing?
Is there any previous factory or in-service test data to compare?

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

 

[unclebob]

If both the ratios and the winding resistances are ok, its difficult to imagine how a turn to turn short is present.
Is this a new transformer?

There are times when I miss the old cranky TTR model.

 

[RedSnake]

Well if you made a transformer like this you could, not exactly 10:1 or 8.25:1,
but you still would have seen it when measuring the resistans in the windings.

Best Regards A

“Logic will get you from A to Z; imagination will get you everywhere.“
Albert Einstein

 

[electricpete]

> If both the ratios and the winding resistances are ok, its difficult to imagine how a turn to turn short is present.

> but you still would have seen it when measuring the resistans in the windings

Agreed. I think TTR is supposed to be a bit more sensitive than resistance for shorted turns based on typical instrument errors. I imagine you might get some insight into whether it is even credible to miss a short between adjacent turns by looking closely at the actual readings, the accuracy of the meter, and if available the total number of turns in both windings (as tapped). Or else just look to see are the suspect winding TTR and resistances are outliers compared to the others. A short in HV winding is more credible to be missed in TTR than a short in LV winding due to lower fractional change for one shorted HV turn than for one shorted LV turn.

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

 

[electricpete]

Another thought came to mind thinking about winding taps. If you had a turn to turn short on a portion of the winding that is not used based on existing tap settings, then that would not show up in TTR or resistance but would still add a loop-current load to cause the observed increase in current. You can of course adjust the taps to make sure the electrical tests are checking all the windings if you have not already done that.

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

 

[unclebob]

To be honest, I never really care about these exciting current values given by TTR testers.
I prefer a 10 kV test applied on the primary as proposed by PRC to get some significant current values.

 

[electricpete]

hmmm, good point unclebob. Before when thinking about what might be different in the core, the only thing i could come up with was a core joint gap, which is a pretty rare/unusual thing. But it seems a lot more likely the core characteristics are being affected in one leg by something else like residual magnetism from last deenergization. That could bias that leg to a different part of the B-H curve which changes the effective permeability. I can see that would be more a factor at low test voltages used in TTR, whereas the high test levels of a standard excitation test can overcome the residual magnetism.

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

 

[waross]

Is the high current on the center core?
If so it may have something to do with the different magnetic path of the center core as compared to the magnetic paths of the outer cores.

Bill
--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[electricpete]

Core geometry factors lead to the center core leg winding drawing lower current than the other two during a standard single-phase excitation test.
Two high currents and a low current is a typical pattern for that test. We might expect similar on the TTR currents.
One high and two lows doesn't fit the pattern.
That's my thoughts anyway, always subject to correction.

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