12 pulse transformer secondary voltage difference 2

[Drivesrock]

Hi all. I am specifying some 4MVA double secondary transformers - 13.8/2 x 0.69kV, 6% impedance, 60Hz for a 12 pulse rectifier feeding a common dc bus with VSDs. I have a specification that requires max 0.14% voltage difference between the 2 secondaries but my transformer manufacturer says he can only get to 0.57% for this power rating.

I know all about this problem of unbalanced secondaries as on a contract a few years ago that I was on, <2% in 2x600Volts secondaries meant that the secondary with 10Volts higher voltage carried two thirds of the total current (and therefore kept blowing that rectifier's fuses) and we had to get reactors fitted to the higher voltage secondary.

What is a realistic minimum difference that can be demanded from the transformer manufacturer or will work without causing current inbalance?

 

[Marke]

Hello Drivesrock

It depends on what is on the output of the rectifier.
If you goe straight into a capacitive input filter, then the voltage is very critical.
If you have a DC bus choke on the output of each rectifier, then the imbalance is not the same problem.
An inductive input filter is strongly recommended to help to reduce the harmonic currents flowing in the transformer. This is obviously of concern due to the 12 pulse approach being taken.

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[waross]

I was browsing an old textbook yesterday, and I came across a scheme to parallel two disimilar transformers with a tapped reactor. The reactor connected the winding ends, and the power was taken off the adjustable tap. I understood that you could adjust the load balance by moving the tap position. I imagine that the impedance would only have to be equal to the difference between the transformers. You may also be able to use a fixed reactor on the primary of one of the transformers.
yours

 

[CarlPugh]

The voltage is kinda low, but you could consider putting the rectified outputs in series. For any reasonable voltage difference there will not be any problem. The textbooks never mention the subject, but under the correct conditions, there is some additional filtering of the DC due to the connection.

 

[Drivesrock]

Thanks Mark E, Waross and CarlPugh.

The transformer manufacturers can't get any better than the 0.57% informed for this 4MVA rated transformer due to the low number of turns of the windings and from another source I learn't that less than 1% difference should be OK as I have got input line reactors. I am checking if an interbridge reactor on the dc link is to be supplied by the drives-rectifier manufacturer for the common dc link for the multiple (400kW) variable speed drives fed from this.
The 12 pulse ac line voltages (connected in parallel after the rectifiers) must be as close as possible to avoid a large current inbalance.
Series rectifiers would be a solution but the current capacity of each bridge must the adequate for the total current of both secondaries, so in my case, each 2MVA secondary @690V will give about 1600A per phase thereby meaning my now 1600A bridges in parallel will become 3200A each in series. However, the bridges in series will also sum the voltages and as the VSDs can't have that we'd have to halve the secondary voltages but that will double the current to maintain the rating! so the 3200A rectifiers will become 6400A each - too big and expensive! Also, this would mean many cables per phase or busbars. Basically inviable.

The solution is to have the transformer secondaries with voltages as close as possible, an impedance of about 6% and input reactors - perhaps an interbridge reactor also.

If things go wrong then suitable line reactors will have to be designed to correct the current inbalance.

Thanks again.

 

[n21]

i am not sure the application but if you are looking for control why not consider a mag amp?

 

[teng22]

This maybe not a direct answer to your question, but it may help your understanding.

For the 12 pulse rectifier transformer the first thing to realise is that there is a limited number of turns ratios that can be used. As one secondary will be Delta connected & the other Wye connected, the first step is to match the open circuit voltages using the "magic ratio" of square root 3 turns between these 2 secondaries.
for instance
Wyeelta LV Turn Ratios
15:26 and 30:52 giving 0.07% difference in open circuit voltages between the secondaries
19:33 and 38:66 giving 0.28% difference in open circuit voltages between the secondaries
30:20 turns giving 0.41% difference in open circuit voltages between the secondaries
So you will see that this is a big limiting factor on the transformer design.

This said, the biggest influence on the current share is the impedance matching between the two LV windings under load conditions. If the impedances do not match, then no matter how good your turns ratio is you will not get acceptable current sharing.

Typically the impedance between HV & LV1 and HV & LV2 must match within 0.5% for a transformer with "closely coupled" LV windings (low impedance between LV1 and LV2).

This figure will rise for a transformer design having "loosely coupled" windings (high impedance between LV1 & LV2), but the principle is the same.

The choice between closely coupled and loosely coupled LV's is also influenced by other factors, including short circuit withstand requirements.

For good open circuit voltage match, the LV turns must have a "root 3" ratio between them.

For good current share (and matched on load voltages) the impedances between each LV and the HV must be closely matched.

These parameters are possible with good transformer design from experienced transformer manufacturers. But as these are not "off the shelf" units then expect to pay for their expertise.

It would be possible to try and correct poor impedance matching externally to the transformer, but this would of course incur additional and unnecessary expense.

 

[CJCPE]

Would better results be obtained with two Y secondaries with zig-zag windings for +15 and -15 degree phase shifts? That would mean that all of the windings for each secondary would be identical, but they would be connected differently. The place where turn ratio error would come in would be in the number of turns in the main secondary windings vs. the "tails" to create the proper phase shift. More windings would probably increase the cost.

 

[electricuwe]

I agree to CJCPE. Such a transformer is also more likely to have identical impedances for both systems.

But there are also other solutions to that problem. Unfortunately I'm missing the corrct translation of the German vocabulary for that. I will be back if I have a correct translation available.

 

[GopelT]

Drivesrock,
I'm running a system that is almost identical as the one you are describing. Have you contacted REX transformer yet?
They built our 4MVA xfrmr which is feeding our 12 pulse system. No problems.

 

[teng22]

Again for your interest..
The suggestion to use zig-zag (Zn) LV's with HV's having phase-shifted windings would increase the complexity and cost of the transformer.

For the LV side you will still face the problem of a low number of turns, and you will introduce the additional problem of winding and connecting the 2 parts of a low voltage high current winding to make the zig-zag.

For the HV side introducing phase shifts again introduces unnecessary complication for a 12 pulse unit. If you have a requirement for HV taps, then you should be aware that the nominal 15° shift will be affected by the tap position. You could of course have an additional "regulating transformer", but again incurring more cost & complication

Phase shifted HV windings are usually reserved for transformers driving 24 (or more!) pulse drives.

Keep it as simple as possible.

As with my earlier posting, I will say that reputable transformer manufacturers know how to deal with these problems. If in doubt of their capability, ask for and cross check references.

 

[mc5w]

You could also put half of the variable frequency drives on 1 secondary and the other half of the VFDs on the other secondary. This can be done with standard drives using a separate DC bus and 6-pulse rectifier for each drive.

There is also a company named Mirus International in Canada that makes a number of power quality products for this type of application. One of them is universal harmonic filter for VFDs that reduces the harmonics of a 6-pulse drive to the equivalent of an 18-pulse drive. This would allow you to use a standard 4 MVA transformer and conventional AC distribution. Another one of them is a transformer with 4 secodaries for supply of 120Y208 for computer power supplies that for balanced load produces the equivalent of a 24-pulse rectifier even though the power supplies are single phase input. Their transformers usually have a 480 volt or 600 volt primary and they probably could not do a 13,800 volt primary or a 4 MVA but depending on the size of your drives they probably could help you.