Determining effect of harmonics on transformer loading 2

[rockman7892]

I am investigating the failure of a 2500kVA 13.2kV-480V transformer which appeared to have failed with internal winding failure. There was no failure analysis performed but there were visible burn marks on the windings and smoke was reported at the time of failure.

I was asked to perform PQ monitoring on the secondary of the transformer (480 Switchgear)in order to determine if several VFD's which are located downstream of the transformer could be causing harmonic issues which led to the transformer failure. The maximum current THD that was measured was 16% with the maximum Voltage THD being about 3.09%.

It is my understanding that the Harmonic Current Limits given in IEEE 519 are for application of the PCC with utility and are relatively meaningless when looking at harmonics downstream within an industrial facility. Are there any guidelines which dictate the maximum acceptable harmonic current distortion at a bus within an industrial system? My understanding has been that the harmonic current value is not really a problem by itself unless it is causing voltage distortion at the location. In this case the voltage distortion is only 3% so there does not appear to be an issue here?

Although there doesn't appear to be any significant harmonic distortion on the Bus, there is the possibility that the harmonic current is causing additional heating on the transformer. The maximum loading on the 2500kVA transformer during the monitoring period was about 1000kVA. Is it possible that 16% harmonic current could have caused additional heating on the transformer which caused its failure when only being loaded at 40%? Is there a "rule of thumb" for determining how much additional heating capacity harmonic current adds to the loading of a transformer in order to determine maximum transformer loading. I know IEEE Standard C57.110 " (IEEE Recommended Practice for Establishing Liquid-Filled and Dry-Type Power and Distribution Transformer Capability When Supplying Non-sinusoidal Load Currents")provides calculations for determining transformer capacity with harmonics but was wondering if there were any general guidelines or rules of thumb without having to carry out the calculations in this standard?

 

[prc]

The failure cannot be from harmonic currents. Trust it is a dry transformer. You have to calculate to get a real picture of overloading from harmonic currents. Sure for 16 % harmonics, derating cannot be more than 50 %.For a typical calculation see Annexure A of IEC 60076-16 -2011 "Transformers for wind turbine applications" A 1000 kVA transformer with harmonic load of 26 % 3rd ,8 % 7th ,5% 5th and 3 % 11th is getting derated by only 11%.
Typical K-ratings are as below:
K-1: This is the rating of any conventional transformer that has been designed to handle only the heating effects of eddy currents and other losses resulting from 60 Hertz, sine-wave current loading on the transformer. Such a unit may or may not be designed to handle the increased heating of harmonics in its load current.

K-4: A transformer with this rating has been designed to supply rated KVA, without overheating, to a load made-up of 100% of the normal 60 Hertz, sine-wave, fundamental current plus: 16% of the fundamental as 3rd harmonic current; 10% of the fundamental as 5th; 7% of the fundamental as 7th; 5.5% of the fundamental as 9th; and smaller percentages through the 25th harmonic. The "4" indicates its ability to accommodate four times the eddy current losses of a K-1 transformer.

K-9: A K-9 transformer can accommodate 163% of the harmonic loading of a K-4 rated transformer.

K-13: A K-13 transformer can accommodate 200% of the harmonic loading of a K-4 rated transformer.

 

[waross]

This brings to mind an interesting question.
The extra heating effects of harmonics due to increased skin effect and increased eddy currents is well addressed.
However, in the case of a wye:delta transformer, do the harmonics cause relatively high circulating current in the delta winding in addition to the other effects?

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

 

[rockman7892]

PRC

Yes transformer is a dry type.

I cannot access the standard that you recommended however C57.110 has a similar example. In that example for a 2500kVA dry type with 9% 3rd ,41% 5th ,20% 7th and 5 % 11th is getting de-rated by 48%. This example appears to show a much higher de-rating than the one that you listed above. Perhaps because of the larger 5th and 7th harmonic? Or perhaps these examples my not be exactly not be similar and I am overlooking something.

So when you say that the failure cannot be from harmonics current are you saying this because at face value there just isn't enough load and harmonic content on the transformer to produce significant heating? If the load was 2000kVA instead of 1000kVA with the same harmonic makeup would this be a cause for greater concern?

 

[VTer]

Hi Bill,

It is my understanding that 5th and 7th harmonic act as positive and negative sequences with respect to fundamental while triplen harmonics act as zero sequence that would circulate in a delta.

Rockman, yes the load on the transformer plays an important role. The harmonic content is based on % of the fundamental current and voltage, therefore if your fundamental current is relatively low in comparison to the transformer capacity the life expectancy loss is negligible. The total RMS current is a square root of sum of the squares of all individual harmonic currents. This current is what causes the I2R heating losses in the transformer. You also get some additional iron losses due to hysteresis caused by harmonic voltages that contribute to the heat generation. With a relatively low load that you mentioned I would not expect the harmonics to be the culprit of the failure as my initial guess. How old was the transformer and was the failure on primary or secondary winding? If on primary, where in the winding assembly was the failure, close to the start of the winding? Where there surge arresters on the transformer?

"Throughout space there is energy. Is this energy static or kinetic! If static our hopes are in vain; if kinetic ù and this we know it is, for certain ù then it is a mere question of time when men will succeed in attaching their machinery to the very wheelwork of nature". û Nikola Tesla

 

[rockman7892]

VTer

The transformer is 24 years old however there is what appears to be an older nameplate behind the existing one so perhaps its even older than that and was re-furbished, repaired, etc.... at one time.

The damage appears to be on the primary winding and on one of the phases the damage was very close to he start of the winding. The other phases have the damage further in along the winding. I'm attaching some pictures for reference.

There were no surge arrestors on the transformer. The transformer is being fed by a vacuum breaker located only about 30ft away. It is my understanding that a vacuum breaker in close proximity can cause severe overvoltage transients on the transformer windings due to the current chopping action of the vacuum contacts. I believe it is always recommended to have a surge arrestor to protect the transformer in these applications?

 

[waross]

Have I done this correctly.
Impedance = 5.87%
Zero sequence impedance = 17.61%
Assume that most of the harmonic is triplen; = 3.09%
Worst case.
3.09% / 17.61% = 17.5%
In the worst case the 3.09% voltage harmonic will cause 17.5% of full load current to circulate in the delta.
If there was equipment not running at the time that the harmonic voltage content was measured, it may cause worse results when online.

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

 

[rockman7892]

waross

Are you referring to a Delta-wye transformer with a primary delta winding?

Is the transformer zero sequence impedance always 3x positive sequence impedance like you indicated?

Are you saying that this 3.09% voltage harmonic at the transformer secondary will cause this 17.5% to circulate in the primary delta winding?

 

[davidbeach]

In a delta-wye transformer the zero sequence impedance is in the range of 85-100% of the positive sequence impedance.

 

[waross]

Thanks David.
I guess that means that triplen harmonics will cause some current to circulate in the primary delta, but not much.
I guess that my calculation was high by a factor of 4 or 5.


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

 

[VTer]

Hi Bill,

I did not deal with this before but my understanding is that the nonlinear load can be modeled as a current source at that particular harmonic frequency. If the OP is measuring secondary 3rd-harmonic current directly, would we not be able to simply sum these for each phase for total neutral current? Simplifying even further and ignoring the magnetizing branch it would be a simple turns ration transformation to obtain delta circulating current. For example and simplicity, if OP is measuring 10A of 3rd-harmonic current from X1 and all 3-phases are balanced, total neutral 3rd harmonic current would be 30A. This neutral current transformed to primary side of the transformer would be the delta circulating current. Primary copper losses can then be obtain by i2R using primary winding resistances. This is the way I would approach it but please correct me if you see errors with this approach.

Rockman, it is good practice to provide surge arresters on the transformer considering cost vs benefit. If the failure is close to the start of the primary winding and no surge arresters were present I would suspect electrical surges as potential cause of failure. I think most modern VCB have become better with the issue of switching transients but any time there is a sudden change of system parameters you will get redistribution of energy in the form of electrical transients hence the need for surge arresters. Was this transformer relatively deep in the system? Did you contact the utility and see if they had any events during the time of your failure? Were there thunderstorms in the area during the failure? I cannot answer your questions to the root cause of the failure but there are people on this site much more qualified to assist you. I am just trying to ask questions for additional information before someone else jumps in.

"Throughout space there is energy. Is this energy static or kinetic! If static our hopes are in vain; if kinetic ù and this we know it is, for certain ù then it is a mere question of time when men will succeed in attaching their machinery to the very wheelwork of nature". û Nikola Tesla

 

[prc]

Rockman, From the photos I would rule out overheating from harmonics. The middle phase looks clean and shining and burning and heating is in localized portions at end phases.
If you can say specific clause of C57.110-2008, I shall check again.
I rule out surges from VCB as this unit is of 1993 vintage and working trouble free so far. Also failure points are distributed and not at ends, as usually noticed in such cases. You are right, with modern VCBs problem is much less.
Dry transformers can be subjected to severe over voltages during switching in and switching off by VCB. The severity will be maximum when current is broken under no-load excitation current or during inrush current with minimum cable length between transformer and VCB.This phenomenon is known from 1970’s.Here are some references:
A3-302-CIGRE Session 2014- VCB Switching interactions with transformers and mitigation measures- Dullni –ABB
IEEE 1971- A guide to the application of VCB- A N Greenwood

 

[rockman7892]

Prc,

The example I am referring to in C57.110-2008 is in Section 5.3.

This transformer is not deep in the system. It is about 30ft from the main service entrance Switchgear which is fed with 13.2kV from the utility.

This particular failure happened when they were restoring power to the facility after a loss of utility (not sure why there was a loss of utility, very little info was provided). Upon re-energizing the transformer from the feeder breaker in the Main Switchgear the transformer failed immediately tripping the differential relay. One thing of note is that this feeder breaker has a 2nd tapped feeder on the load side of the breaker which feeds another 2000kVA transformer at a 2nd substation location (not sure at the moment what distance to the 2nd transformer is.) So I believe that when closing the breaker they were energizing both the 2500kVA unit which failed as well as a 2nd 2000kVA unit. Not sure how the 2nd transformer could impact any of this?

Seeing that it failed during energization is what seems to point to a voltage transient as the culprit in my mind since this the severity of the severity will be maximum during transformer inrush as you mentioned.