Semiconductive coating for improving corona

[jimbofitz]

Does anyone have information about semiconductive materials avaiable for use on transformer core steel (or Metglas in this case)? I am hoping that if i use this type of material on the edge of the core i can increase its effective radius to improve performance. My company isn't familiar with any of these materials nor am I. Any help is appreciated.

Thanks

 

[wolf39]

What if the semiconductive material doesn't stick to the core but dissolves in the transformer oil and starts circulating? Also, remember that the transformer core is vibrating and thermal expansion/contraction takes place. How can you be sure that the semiconductive material keeps its position for the transformers livetime?

Usually the LV winding is the one nearest to the core. Under oil there should be no corona problem in this area if the correct distance between core and winding is maintained and an insulation cylinder is provided. For grid connecting transformers with their two HV wye/wye windings you require a LV tertiary (compensation) delta winding which always is placed next to the core. To my knowledge there is no need in the transformer business for special measures as you suggested. Also, semiconductive material has a conductivity which is only a tiny fraction of conductive metals, this including the core steel.

If you want to eliminate or soften the influence of core steel edges, you must stack your core layer by layer with lamination which recedes by varying dimensions both sides. Then you get core legs which are almost perfectly circular. This is of course uneconomical and would allow you to reduce the gap between the grounded core and the LV winding only very little.

Regards

Wolf

http://www.hydropower-consult.com

 

[jimbofitz]

Wolf,

Thanks for your thoughts. For a little more background information this will be a potted transformer assembly. We are size and weight challenged because it hangs underneath an aerostat at elevations up to 15000 ft. The extreme altitude changes don't allow oil to be used in this application. It is an odd design in comparison to any transformer I've seen in my limited experience. The coil construction consists of a HV primary with low voltage windings on the inside and outside. I don't think I left enough margin on my coil to have proper space from the HV winding to the edge of my core steel since the LV winding build is small. There aren't many sharp edges, the copper foil of the LV winding is among the few. I pass all of my Hipot testing in the prototype but corona testing has been interesting.

We exclusively use rectangular core cross sections at my company mostly due to shock and vibration requirements of our products. I like the idea of round core cross section but it won't happen... and we already have the core material in house.

The thought of using the semiconductive material is to increase the effective radius without introducting higher core losses by shorting out laminations with a pure conductor. I'm not sure if it is a real solution but it doesn't hurt to ask...

When you refer to an insulation cylinder are you referring to the insulating medium that the copper is wound on (i.e. core tube)?

Regards,
Jim

 

[wolf39]

jim:

Is your transformer fitted in an airship gondola and does the term "potted transformer assembly" mean that the transformer in question is of an air-cooled design with resin-moulded coils? Give us technical data like kVA rating, voltage levels of LV and HV windings, etc. Why is the LV winding split into two separate coils?

A rectangular core cross section is having a poor space factor in comparison with a stepped core design. You can reduce the diameter of the inner coil (and the outer as well) if you change your design from a rectangular to a stepped core. This saves weight, which is benificial for the application described.

To give you one example:
A core has a cross section of 100 square cm. If rectangular you get a core dimension of 10 cm x 10 cm (core lamination factor not being considered). Edge to edge you have a distance of 14.1 cm. A perfectly circular core has a diameter of 11.3 cm. Perfectly circular is not practical but a core with 5 or 6 steps still allows you to reduce coil diameters considerably.

It is possible to wind a LV coil onto an insulation cylinder but this would limit heat exchange. For oil-cooled power transformers the inner LV coil insulation cylinders usually are placed away from the core and away from the LV winding by using spacers. This improves the cooling.

Regards

Wolf

http://www.hydropower-consult.com

 

[prc]

Semi conducting epoxy resin impreganted tape is a good solution that is being used by many manufacturers.

 

[jimbofitz]

The subject transformer is encapsulated in an epoxy resin and is forced air cooled. There is a housing underneath the air-frame and the payload is exposed to the elements. The power delivered is 80 kVA. There are two isolated 3-phase secondaries; 240V and 200V. The primary varies from 5700V to 6400V at 400Hz. The Corona requirement is 8500V prior to inception.

I like the space savings of the circular core cross section but it's not a viable option right now. The core material for this assemby is Metglas and I haven't seen a circular cross section of that variety yet. And to further confuse this discussion the core is gapped for a specific inductance. The transformer serves dual purposes: power delivery and a line compensating inductor. We wanted to have seperate line-line inductors but the weight savings of a dual-purpose transformer assembly was irresistible for the custormer


I performed testing last week on a partially built prototype and found my corona inception voltage increased with the core inserted. I was pleasantly surprised but I still fear the final assembly will be a challenge.

PRC,

Do you have a specific manufacturer to recommend for the semicon tape? I see Von Roll makes a variety of them but I wouldn't know where to start.

Regards,
Jim

 

[edison123]

jim

For Isola products (in pdf),

http://www.vonroll.com/products-and-solutions/documentation/brochures/

But I am not sure how would apply these in dry type trafo windings since they are meant for MV rotating machines.

By circualr core, wolf meant rectangular core plates (in sectional view) of various widths filling up inside a circle to increase the core area, which in turn increases the voltage per turn resulting in lower no. of turns.

 

[wolf39]

Jim:

edison123 described excactly what I meant.

I'd like to hear from prc how, when and why semiconductive tapes are used in a transformer design. The Isola Roll tapes edison123 mentioned are used solely for application on the end winding portion of generator stator windings. The purpose of these tapes is not to provide a shielding but to control end winding surface sparking which can develop when high voltages are present.

I can't imagine how semiconductive tapes with their high surface resistance can provide a proper shielding like metal surfaces do.

One way to overcome possible corona from the core leg edges is to put four circle segment to the rectangular core in order to get a perfectly circular cross section. Then wrap one turn of metal foil around this cylinder and ground it. You have to make sure, however, that both ends of this foil are well insulated from each other in order to avoid a short-circuited turn. In case you have problems with the core cooling, provide air gaps between circle segments and rectangular core.

In your case, with one of the two LV windings placed next to the core you shouldn't have corona problems because the LV winding is acting like a shield. Just make sure the the axial length of all three windings are about identical.

Regards

Wolf

http://www.hydropower-consult.com