Recycling Power Through a Wye-Wye Transformer

[JohnSmith1]

Good afternoon,

I work in a company that specializes in 3-phase power converters. Our assembly line electrical test stations need to be able to test our 480V and 208V units, both at 100 A. However, the current reduction at the end of the circuit is only about 5 A. In order to try to save money and electricity, we reroute the remaining 95A back into the test circuit. This means that our power usage for one of these test stations should average maybe just above 5 A multiplied by the test voltage.

Okay. So, the main voltage we run throughout our facility for tests is 480V. We really do not want to run an additional line in order to test our 208V units, or have to run even more lines to test 380V or 400V units as they are developed. Recycling the 480V power is very easy because it can junction directly into the main line. But, we need a transformer to test our 208V units, and as a result we must develop a custom solution to recycle this power while still achieving isolation of the unit under test. That is something I forgot to mention - we need a transformer at each test station in order to ensure all units being tested are isolated from the rest of the building.

Anyway, we have developed a plan to recycle the 208V power in a very unconventional way. We desire to have a custom transformer produced with 480V primary and 480V secondary, as well as 208V taps on both the primary and secondary windings. What we plan on doing for our 208V tests is having 480V power initially enter with relatively high current into the transformer. It would then be transformed into 208V, 100A, and pass through our test unit and test equipment. Once it has passed through all of these, it returns to the input side of the transformer enters the primary winding via the 208V tap. Once the current is steady at 95A entering the primary winding (recall 5A current loss in system), then only a few Amps will be required to sustain the system, entering in the 480V input.

So basically, at t=0:
I_480_in=43A
I_480_out=0A
I_208_in=0A
I_208_out=100A

Then, at t=inf:
I_480_in=3A
I_480_out=0A
I_208_in=95A
I_208_out=100A

Now, the only way this could work is by using Wye connection types on both the primary and the secondary. This would permit two inputs for the transformer, because current and voltage vectors align. This would not be possible for a delta-wye or delta-delta transformer.

So, specs look something like this:

kVA: 75
Qty: 7
Primary: 480V Wye
Secondary: 480V Wye
Taps: 208V Wye on both primary and secondary
Material: Aluminum
Temp Rise: 150

The main issue that we have is that no one is willing to make it for us. We have tried many custom transformer companies, and received answers ranging from, "it's not safe," to, "you don't have enough volume."

Is this in fact an unsafe design? If it is unsafe, why? If it's not unsafe, why won't anyone produce it?

Thanks for the help in advance!!!

 

[Ingenieur]

What is the conversion?
ac/dc?
ac/ac voltage level, freq?

so you put power into the converter
100 A and you lose 5 A of power in the process?
and you take the converter output back into the supply grid?

 

[Ingenieur]

Single ph equivilent
is this what you want to do?
do you not see a few issues?

 

[JohnSmith1]

Ingenieur,

We have a number of different converters that we test from DC/DC to AC/AC to grid tied inverters. These don't matter so much because our test filtering equipment converts the electricity back to 3-phase AC 60 Hz 208V for return to the grid or the transformer in this case.

I do not see an issue with your diagram, although the block you have drawn and labeled "power converter" might be better named simply as "load". Power converter misleads one to think there is DC or something else coming out of the test apparatus system. The electricity exiting the test system and returning to the transformer is of the same phase, voltage, and frequency as that entering the system (60 Hz 208V).

 

[Ingenieur]

Assuming you can keep the 480 and 208 in phase
if the primary has 1000 turns total
how many volt/turn above the 208 tap?
and how many below?
how many amps in the winding above the 208 tap?
how many below?
translated to the secondary with a fixed turns ratio?

have you shown this to your plant power ee?
how many xfmr mfgs said no go due to safety?

 

[junkhound]

Good afternoon,The main issue that we have is that no one is willing to make it for us. We have tried many custom transformer companies, and received answers ranging from, "it's not safe," to, "you don't have enough volume."ce!!!

Call and tlak to Tierney in Seattle, they have made anything I've ever asked them to make, and were low bidder besides on some 150each qty builds of 120/480 delta wye xfmrs.

 

[steve66]

Recycling the 480V power is very easy because it can junction directly into the main line.

Can I stop you right there? When you say it needs to draw 100 amps and return 95 amps, I assume you mean it needs to draw about 50KW, and return 47500 watts to the source. (Even more power if we are talking about 3 phase.)

It is possible to get that with electronics. But it doesn't usually just happen - it takes a lot of design. Power needs a reason to flow. In general, it's not going to flow from a source to a load and back to the source. For the converter to draw 100 amps (and 50KW), it needs to see a low impedance on its output.

In general, by tying the output back to the input, you are going to make the converter see a high impedance on the output. That's going to reduce the current it draws on the input.

So can you be a little more specific about how you plan to recycle the 480V power before we get into the various voltages and transformers?

 

[ggunn]

I do not understand what you mean by "recycle" power. Power (energy, actually) is consumed by loads and converted to heat, light, mechanical energy, whatever, and you aren't getting it back. Sure, you could use electrical energy to drive a motor that spins a generator that powers an inverter that returns power to the grid, but it's a less than zero sum game; you'll return less to the grid than you used in the motor. The Laws of Thermodynamics have not been repealed.

 

[Ingenieur]

I think what he is proposing
power up an unloaded 'converter' (undefined) with 5% inherent losses
feed the output back into the supply
but what in the supply will make the inverter produce full load?
the power flow does not make sense
no 'load'?
on top of that he wants 2 xfmr inputs on a common winding/core of different voltages

 

[Besoeker]

Can I stop you right there? When you say it needs to draw 100 amps and return 95 amps, I assume you mean it needs to draw about 50KW, and return 47500 watts to the source. (Even more power if we are talking about 3 phase.)

It is possible to get that with electronics. But it doesn't usually just happen - it takes a lot of design. Power needs a reason to flow. In general, it's not going to flow from a source to a load and back to the source. For the converter to draw 100 amps (and 50KW), it needs to see a low impedance on its output.

In general, by tying the output back to the input, you are going to make the converter see a high impedance on the output. That's going to reduce the current it draws on the input.

So can you be a little more specific about how you plan to recycle the 480V power before we get into the various voltages and transformers?

I think he means it is used to test the converter (of whatever type) at its rating and sends the converter output back to the supply. This is not altogether uncommon on test set ups, particularly if there is a lot of power involved that would otherwise be wasted.

 

[steve66]

I think he means it is used to test the converter (of whatever type) at its rating and sends the converter output back to the supply. This is not altogether uncommon on test set ups, particularly if there is a lot of power involved that would otherwise be wasted.

Yes, I'm sure that's what he means, and that can easily be accomplished with any of these:

http://www.bing.com/search?q=electronic+load&src=IE-SearchBox&FORM=IENTSR

But 50KW is a whole different ball game.

I'm not sure it can be accomplished with just a transformer, but a motor-generator with some paralleling gear to sync it to the utility would easily work at very high power levels.

Edit: Ops, maybe all of those electronic loads don't return power back to the source, but the one we had in my college power electronics lab did. So they are available.

 

[winnie]

JohnSmith1,

The basic concept that you propose is reasonable and safe.

As I understand it, you wish to test a device which converts input electrical power to output electrical power, without wasting the output electrical power. Instead you want to 'recycle' the output electrical power back into your supply network, or at least into the supply of the device being tested.

We do this all the time with electric motor dynamometers. We have two motors with their shafts tied together, run by VSDs. One motor is driven as a motor, the other is in regen. The mechanical power going down the shaft is 'recirculated' on the DC bus between the two drives, to run a test we need only supply electrical loss.

The devil, however, is always in the details. For example, how will your power conversion device tolerate voltage at both input and output when the device is off. How will you control the system to actually get power to circulate? Will your converter device adjust its output to force power to circulate, or do you need to adjust the '208V' of your simulated load to get power to circulate? Do you need some way to turn your simulated load on and off?

In your plan, why do you need taps on both primary and secondary? Primary is connected to your 480V supply; secondary is _either_ connected to your 480V converter input or your 208V converter output. IMHO even though power would be flowing from 208V to 480V, the 480V side remains your primary since it is connected to your steady 480V bus.

If you always connect 'primary' to 480V and 'secondary' to your converters, then you don't need the wye to wye phase matching. A conventional delta-wye connected transformer would be suitable.

As far as actually sourcing the transformers that you need, they should be easy to build. I have been very happy with drive isolation transformers from Fargo Electric (www.fargoelectric.com). Their existing designs include devices with have 480V, 240V, and 208V taps on the primary, as well as 2.5% taps on the secondary. They are probably quite capable of building designs with 480V primary and multiply tapped secondaries.

You could also simply 'bank' single phase transformers to create three phase transformers.

Finally, I think that both your power converter input and output could be connected to the secondary of the _same_ transformer; I don't think you would need two transformers for each test station. You might need to add three phase variacs in 'buck/boost' configuration to adjust voltage and control your load, but that goes back to your design of your control system.

Best of luck,
Jon

 

[Besoeker]

Yes, I'm sure that's what he means, and that can easily be accomplished with any of these:

http://www.bing.com/search?q=electronic+load&src=IE-SearchBox&FORM=IENTSR

But 50KW is a whole different ball game.

I'm not sure it can be accomplished with just a transformer, but a motor-generator with some paralleling gear to sync it to the utility would easily work at very high power levels.

I've done it quite a few times for various projects.
One of the more interesting was for testing energy storage flywheels. The problem is what to do with all the stored energy.

A few numbers. The source was three phase mains transformed up to about 1200V with a Delta-delta wye (Ddyn11) transformer, rectified and smoothed to get an ultra low ripple 1500V, 600A supply (900kW).
That powered up a whole bunch of flywheel motors. Initially, the stored power was just dumped in a load bank. Expensive waste of energy.
The next phase of the project was to recover that energy - recyle it if you like. We built a step down chopper to get from 1500Vdc to 400Vdc and followed it with a mains commutated inverter to stick the energy back into the 400Vac utility supply.

Interesting as I said, all topology we already employed, but the scale made a little more of a challenge.

 

[Ingenieur]

What he proposes is not safe
that is why some mfgs refuse to do it
he wants to tie 480 and 208 together on the same xfmr primary
he will have short ckts and most likely destroy the converter
many shops will build one off's
http://linepower.com/
http://www.pemco.net/expertise/
but not this design

 

[GoldDigger]

The transformer has a fixed turn ratio and there is no guarantee that the service taps and load-influenced changes will be compatible.
Basically the 480 and 240 need to be on different transformers. The converter can bridge from one to the other, but a direct grid connection may not make a suitable load for all tests. You still need a resistive or motor/brake load at some point.

mobile

 

[Ingenieur]

The transformer has a fixed turn ratio and there is no guarantee that the service taps and load-influenced changes will be compatible.
Basically the 480 and 240 need to be on different transformers. The converter can bridge from one to the other, but a direct grid connection may not make a suitable load for all tests. You still need a resistive or motor/brake load at some point.

mobile

Bingo

he needs electrical isolation
motor/gen
charge batteries, invert back to grid
caps and invert back
Maybe use one of his 208:480 inverters

I would never hook it up as proposed
it would not last very long
he's applying power to both sides of the converter

 

[mike_kilroy]

John smith u appear to want to keep the details of ur black box secret. Ok. I sent u pm with my email to reply to. We will be happy to quote and build u 7 custom xfmrs. But we will also request details on what u are trying to do in order to help u do it right. Yes, lots of ways to screw this up.

Sent from my SM-G900V using Tapatalk

 

[mike_kilroy]

If isolation is indeed significant we can simply add a second 208v secondary to each phase instead of going back into prim winding. Same result... or perhaps no isolation is required and we just make it an auto xfmr....

Sent from my SM-G900V using Tapatalk

 

[Sahib]

I also think OP proposal to connect convertor output back to transformer primary risky as any phase reversal (say by POCO by mistake) could cause short circuit. Silly, my earlier post saying it got deleted.....
I want to add a reverse power flow relay may solve the above issue.

 

[Besoeker]

What he proposes is not safe
that is why some mfgs refuse to do it
he wants to tie 480 and 208 together on the same xfmr primary
he will have short ckts and most likely destroy the converter
many shops will build one off's
http://linepower.com/
http://www.pemco.net/expertise/
but not this design

Is that his diagram or yours?