Transformer Winding Costruction 5

[gordonl]

A transformer designer has suggested to me that a disc winding is mechanically superior to a sheet (foil) winding. Any comments thoughts?

In reference, we experienced a 13MVA rectifier transformer failure, apparently due to failure of the low voltage due to mechanical stress between layers. We are looking at buying a spare set of coils for the six other transformers should another failure occur. Our choice is now to stick with the sheet design with nomex for more insulation strength, or switch over to a redisgned disc winding. (The load is approx 150% for 3 minutes every five minutes)

 

[SphincterBoy]

Yes, in fact disk refers to "helical". It is inherently
stronger than "sheet" (or layer), and will move much less
after/during faults.

Definitely specifiy "helical" or disk windings.

The (square) copper conductor is wrapped in a cellulose insulation (much like an insulated cable), which is then wound helically around cardboard forms, with cellulose spacers keeping the windings separated from one
another.

The sheet winding is where a flat copper or aluminun sheet
is sandwiched between epoxy-impregnated insulation, and wrapped concentrically (one winding on top of another), until
the desired ratio is achieved. Commercial (2500 KVA and below) and residential transformers use sheet winding.

There are a few manufacturers that employ the helical wind, namely Waukesha Electric Systems, Delta-Star, and Sunbelt.

In fact, it might be wise to even specify a maximum %Z
change after fault, and ask for certified test results
along with bid price, otherwise you won't have a leg to stand
on when you reject a low bidder (the sheet" wound).

 

[gordonl]

Could you please elaborate on the %Z change after fault, this idea is new to me, and what would the certified test reports be for, with the bid?

 

[busbar]

Some related discussion at thread238-14487

 

[jnims]

A disk and helical winding, though similar, are not the same. Each winding has a number of vertical stacks of rectangular, individually insulated conductors.

In a helical winding, each stack is a single turn (usually with multiple conductors per turn). When it is wound, the stack of conductors are wrapped around the cylinder with spacers between each turn. Think of it as a spring with space between coils. The only change in conductor alignment is for transpositions.

In a disk winding, each stack, usually called a section, is made up of two or more turns. Sections are connected by alternating inside (close to cylinder) and outside crossovers. A cross section would look something this

x xxxx x
x x x x
xxxx xxxx

They are generally stronger with respect to short circuit strength than sheet windings.

 

[SphincterBoy]

Thanks jnims for the good explanation.....my info was
based on info passed down by various manufacturers. Good
info!

Gordonl: The change in impedance after a short circuit
test is (I recall) called out once in the ANSI C57 specs.
The ANSI specification is pretty "loose", to make room
for the sheet wound manufacturers.

The disk-wound (jnims) and/or helical-wound, if the
design has been short circuit tested, will have (should
have) a small change in impedance, after the fault. The
change in impedance is a result of winding movement.
Ideally, one would like to know exactly how and where the
windings moved during the fault. It's better to equate
these movements to changes in impedance.

The difference (before and after fault) may be provided
in the short circuit test report--you'll have to specifically
request this in your bid specs.

The Waukesha Electric Systems' Wisconsin factory has
performed these expensive short circuit tests on several occasions, and will provide you with these test results.

Their tests indicate the impedance changes less than 0.5%!!

I entertained one bidder who exlaimed "Our 5% change in impedance is pretty good!"

I have had my share of 'lemons' in the past, and therefore
I wrote my specs. to make certain I would be purchasing
the strongest units money would buy, since electricity
is a critical infrastructure, and equipment failure
is a possibility.

In my opinion, a substation class transformer should perform
flawlessly (with PM, of course) for at least 30 years, and observing how manufacturers are turning 'durable' goods into
'less durable' to save money is troubling, especially in
power apparatus, prompting me to tighten my requirements.

This may not be possible in the warmer regions in our country, since insulation age is determined by the heating
it undergoes.

I would like to know who, if anyone has utilized DuPont's NOMEX, which has an exceptional heat tolerance, which I beleive would make it immortal, as far as lifetime.

I specified NOMEX some years ago, and quoted price was
beyond affordability.

 

[gordonl]

We used Nomex because cost wasn't a factor, we needed the transformer back fast, and this was the best way to improve strength without changing the design. I talked to a fellow from Niagra who said they used Nomex in their sheet winding rectifier transformers. My transformers were part of a turnkey job where cost was king and the supplier had free hand, now we're paying for it.

 

[SphincterBoy]

Sorry about your failure. I can probably send you
a copy of my bid specs......just say the word.

I also know that some 'used' transformer shops have a
nasty habit of not performing impulse or doble testing, even
though it was specified.

We had to make prior arrangement with one such shop, to make
certain the impulse & doble tests were performed before
we got stuck with the unit.

Guess what? The used unit we bought failed the impulse test, resulting in the shop having to untank and rewind the
transformer. Not to worry.....the transformer failed the test in the factory, not on the pad. We even inspect our
bucket trucks in the factory setting, before we take delivery.

You probably are going to have to witness all factory testing
from now on, prior to shipping.

NOMEX (like all DuPont products) is patented, but I am wondering when the patent will expire, and maybe the
price will become more reasonable.

 

[electricuwe]

Be careful with the term disc-winding:

There is also a type of transformer-design where disc-wound coils of primary and secondary are interleaved. This one is absolutly the best regarding short-circuit ruggedness, but very expensive to manufacture, difficult to cool and only used rarely today.

In German the terms are:
"Scheiben-Spule" (disc coil)for the design described in this thread

and
"Scheiben-Wicklung" (disc winding)for the design with the interleaved primary and secondary

I do not know if the terms translate word by word and unfortunatly I don not have an English Textbook at hand but I can recommend the following:

Karsai, Kerenyi, Kiss: Large Power Transformers (1987)

 

[jnims]

electricuwe

Thanks for adding to the discussion. All of my limited design experience is with medium power transformers, is based on one manufacturer's practices and is getting pretty dated. I have never seen an interleaved winding, but have heard of its use.

 

[jnims]

I also meant to add a comment to SphincterBoy's remarks on factory testing. The engineering firm I work for now has had several instances of one unit (in a two unit order) failing the heat run when the sister unit passes. In the past, our bid forms have asked for a deduct to eliminate one of the tests. We will probably discontinue this in the near future.

 

[SphincterBoy]

jnims and electicuwe.....many thanks for great comments.

I encourage everyone to become a Doble Engineering client, and submit reports of failed transformers to Doble.

I read a Doble newsletter, briefly discussing a client's
report of a failed impulse test at Delta-Star's Virginia
facility. It's good to know such things, especially
when Delta-Star may be low bidder.

I've heard about interleaved windings, but don't know
of any major manufacturer using them in the U.S., except
Howard in Louisiana (cannot confirm this....salesmen
have nasty habit of lying).

 

[gordonl]

SphincterBoy,

I would appreciate your specs, my email is:
gmlees@algoma.com

 

[prc]

What all varied comments!
Let me comment just to the point ie what is the best winding type for the LV of the said rectifier trf?
You have to check with the manufacturer of the unit as he only knows the space available.
The relative strength of the winding to with stand short circuit forces arising from fault currents:
sheet(foil)winding is the best ,disc wind next.
helical and layer wdgs are relatively weaker from short circuit withstand strength angle.
There is no need for nomex insulation unless the temp rises are abnormal.Nomex will not help the trf in short circuit strength but prevent inter layer insulation failure from temp over run . The designer can easily check these points.
Dont worry about the change in %impedance.You need not check this after every fault. This is an indication of the shift in winding from short circuit forces .IEC gives the limiting values and this is checked when the trfs are subjected to special test viz short circuit test ie applying full voltage on HV terminals with LV shorted for a duration of 0.25 seconds

 

[gordonl]

Nomex insulation is phsically as well as thermally superior to straight paper. Our fault seemed to have been caused by the edge of the sheet cutting the interlayer paper, thus the switch to Nomex.

Does anyone know of article/s which compare winding construction?

I think SphBoy had a good point in using %Z change of equal units as a comparison of winding design strength.

Thanks to all, much appreciated.
Gord

 

[electricuwe]

I would like to add two important comments to this thread:

Beside the type of winding design (sheet or disc or whatelse) there are several other important factors in transformer design which are crucial for the short-circuit-performance of a transformer. See the book recommended in my post dated Feb 5th. So you can have designs with good or with poor behavior for all types of winding design.

I have to diagree with dpc: A significant change in impedance is the sign of mechanical changes in the transformer and a transformer showing significnat change in impedance has already gone half the way of being scrapped. Of course, on transformer which is deemed to be of good design its not necessary to check impedance after the transformer was subjected to a short circuit, but on a transformer where this is in doubt its a good idea to check the impedance after a fault.

 

[prc]

My posting was only to the specific problem raised by gordonol.
I am not sure whether Nomex(aramide paper) is superior in physical properties to normal cellulose kraft paper.
Pl see IEEE transactions on Power DeliveryVol 9 No.4 Octo 1994 pages 1892-1906"Background information onHigh Temperature insulation for Liquid immersed Power Trfs-Working group"where properties of Nomex and cellulose are compared.
Tensile strength as per ASTM D-828 Machine Direction :
2 mil cellulose paper- 30 psi

2 mil aramide paper -25 psi

So tensile strength of Nomex is less than calendered Kraft paper.
To increase the interlaminar strength of sheet winding ,a
better or cost effective way may be to use epoxy resin coated kraft paper which will make the entire winding solid block preventing relative movement between layers,increasing mechanical strength etc. I fully agree with elecrticuwe the strength of winding depends on the expertise of designer and quality of manufacturing and processing .So a disc winding also can be made equally effective in short circuit withstand srength by proper design.A good designer will be trying to reduce the mechanical forces from fault current rather than build on the force with stand strength.
But generally a sheet winding has better short circuit withstand strength than a disc winding except for axial compressive forces.
The case presented by gordonol : it seems the failure is not due to fault currents and the frequent over loads cannot cause much forces compared to fault currents.Probably edge preparation ,less thickness of interlaminar insulation etc can also create problem.

 

[prc]

Gordonl's request for papers on winding construction;There are some IEEE and CIGRE papers on the subject,but may be of more interest to trf designers.A good general tutorial on the subject is Appendix B to IEEE C 57 .125-1991 Guide for failure investigation ,documentation and analysis for Power Trfs,where features of various types of trf windings are explained ,stregths and weaknesses with regard to short circuit withstand strength ,surge withstand strength etc are explained.Incidently this std is an excellent document for those on trouble shooting &failure investigation of trfs.
For a detailed study on short circuit performance of trfs the following classic books may be of use:
1)Short Circuit Duty of Power Trfs-Giorgio Bertagnolli,1996-ABB Transformatori,Leganano,Milano,Italy.
2) Waters.M-The Short circuit Strength of Power trfs-Macdonald&co,London 1966.

 

[prc]

Sigle Phase short circuit impedance measurement as a diagnostic tool-
My observation that this will not be of not much relevance is with respect to the specific case thatwe were discussing ,based on my limited experience with trfs. This test and interpretation are explained in two IEEE stds:
1)C57.125-1991,Appendix A ,page 39
2) IEEE std 62-1995 Guide for Diagnostic field testing of Electric Power Apparatus-part1 oil filled Power Trfs
clause6.1.4.8 "a change in the short circuit impedance of trf indicates a possible winding movement with in the trf.Since the overall measurement accuracy is no better than 1%,using 0.5% accuracy meters, changes of +,- 2%of impedance are not considered significant.Changes of more than +,-3% of impedance should be considered as significant"

The problem is in normal trfs such a degree of variation will occur with substantial distortion/relative movement of windings(not with the type of relative movement that we saw in trf under discussion)viz buckling of inner winding,axial displacement between HV &LV windings.In field, trfs used to trip much ahead of such situation getting developed.If any one can narrate thier experience of this test as a diagnostic test to detect incipient fault,it will be helpful,except of course as a means to check damage during transportation of trf from works to site.
Specifying short circuit test in bid spec:Utilities in Europe and esp in Indian sub continent do specify such test as a special test in trfs subjected to frequent short circuits eg Traction supply trfs,furnace trfs etc. The problem is manufacturers will not have facility to do this test and trf has to be taken to special labs for test (KEMA,CESSI ,CPRI etc) Test cost is also high ,almost 20-60% of trf cost.For details see IEC 60076-5(2000)

 

[gordonl]

prc,

Thank You for all of your help. Before going with the nomex I took a piece of each and ripped them, the nomex was harder to rip than the kraft. We did use, as did the original design, epoxy coated kraft paper, as well adding epoxy to the nomex. I beleive the movement was radial, as you pointed out the sheet was quite strong axially because of the epoxy.

prc, are you suggesting the redesign stick to sheet? The original design had no gap between sheets, for the rewind we used a small gap between sheets, and we instructed the winders to watch the edge of the strip as they wound. We could stick to the mods, but I'm going to need to do some research and maybe put the question to the designer again if he beleives the disc would be superior for this application.