VFD + Transformer Combo for MV motors 3

[edison123]

I would like to test MV motors up to 13.8 KV in my repair shop on no-load. My utility supply is at 415 V, 50 Hz.

Can I use a run-of-the-mill VFD with a step-up transformer to apply the required MV voltage to the motor ?

Is it possible to step-up the frequency to 60 Hz with the same VFD + Transformer combo ?

*Why a man thinks he outrun a chasing dog when it has twice as many legs?*

 

[Skogsgurra]

Hello ed!

You did read thread237-189702 ?

It is about that same problem. We used a standard VFD and a step-up transformer. We even did the preliminary tests in a shop like yours.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[edison123]

Thanks Gunnar. I just read the other thread and it seems possible.

Now other questions

To reduce the line current (due to supply restrictions), can I use shunt capacitors up-stream of the VFD ?

What will be the cost of a 500 V, 500 Amps standard VFD ? What brand would you recommend ? The duty cycle (consisting only of poor pf lagging current) would be around 8 hours in a day.

Also, can I get up to 60 Hz from a 50 Hz VFD ?

Does the step-up trafo has to be something special ?


*Why a man thinks he outrun a chasing dog when it has twice as many legs?*

 

[Skogsgurra]

Hi again,

I think that any modern VFD could be used. The vector control seems to be more difficult to run with synchronuous motors (at least in the MasterDrive) than other modes, like the current source mode. But there are some VFDs that have a sensorless synchronuous motor mode. I would look for one of those.

There are used VFDs available if you want a budget solution (and quick delivery). I have contacted a Dutch company that seems to specialize in used industrial equipment. Do not remember the name.

Yes, you can get up to quite high frequencies. Up to 300 Hz is not unheard of in standard VFDs. But, the output voltage will not be higher than your input voltage. So your transformer needs a few taps on the LV side to adapt for "difficult" voltage/frequency combinations.

The transformer does not have to be very special. BUT I would recommend an oil filled one. Dry resin transformers are risky because they tend to be PD prone and that degrades insulation if run from a PWM inverter.



Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[edison123]

Thanks again Gunnar.

Yes, I am looking at budget solution (because of the frequency of use) and if you can recommend a place / make / brand where I can buy one, it would be great.

I was also thinking oil-filled trafo given the voltage levels. Good point about the trafo V/F ratio (I didn't think of that)

How do you feel about the compensating caps up-stream of VFD ?

*Why a man thinks he outrun a chasing dog when it has twice as many legs?*

 

[Skogsgurra]

I would not just connect capacitors like that. VFDs produce harmonics (yes, I know you know) and, if the VFD is a large portion of the transformer load (some say 10%, others think 20 %), it may start a resonance between transformer impedance and capacitors.

Perhaps a VFD with AFE? They draw very near PF 1 and the current wavshape is very near sine.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[aolalde]

I think is a good idea low voltge 50 HZ>> VFD >>> Transformer for testing large motors no load at full speed, voltage and frequency . Even you will be able to adjust your voltage or frequency to match the motor, or analyse the current performance for voltage or frequency changes, or both.Try to get a VFD with the closer "sinus wave form" since deviation from that will affect the test results.
With that array you will avoid a very high cost 13.8 kV VFD.
As a plus, compared to other arrays, your starting current will be low and the acceleration smooth if you use the ramped function of the selected VFD. Only disadvantage that I see is that the low voltage side will require high current to match the required kVA. As Scogsgurra said be carfull adding capacitors and if so I will do it on the secondary side (high voltage) of the transformer.

 

[LionelHutz]

Capacitors upstream of the VFD will do nothing to help. The VFD will only draw real power from the line side and capacitors are used to compensate for inductive reactive power. If the motor is run unloaded then most of the current drawn from the VFD will be inductive reactive. The VFD capacitros and output section will source this current. The input current will actually only be the real power required to accelerate and maintain the motor speed. You'll likely find it's much lower than you thought it would be.

You can use a simple V/hz VFD for this application. Make sure you size it large enough to supply the required motor current to actually accelerate the motor plus the transformer losses and reactive current.

 

[edison123]

Sorry guys, I am getting a bit confused.

Gunnar - Active Front End Filter will compensate for the lagging pf

Lionel - VFD caps will take care of the pf

Aolalde - Put the caps on the trafo hv side (secondary)

How should I proceed ?


*Why a man thinks he outrun a chasing dog when it has twice as many legs?*

 

[Skogsgurra]

First. There is no lagging pf in the power going to the VFD. The low power factor is entirely due to current waveform distortion. Cos(phi1) is usually something like 0.98 - which is good.

Second. Distorted waveforms can not be corrected by ordinary phase compensation capacitors. You need either an active filter (aka waveform correction unit), a bank of tuned filters or an active front end.

Third. The lagging pf of the motor is still there. But it is an affair between VFD and motor. It will not be seen in the three-phase mains connection. Hint: the reactive power now circulates ("swings") between DC link in VFD and motor.

Fourth. It is not at all sure that you need to worry about the current waveform. It depends on your grid. If it is strong compared to your load, there is usually nothing to worry about. Line reactors will of course always be good.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[edison123]

Thanks Gunnar. I got it. (I am a bit slow but I get there eventually).

VFD will always present a good pf at the supply end regardless of the load pf. So, no-load run will show up as active power of the motor at the supply end meaning low current demand on the supply side.

Aolalde's idea of using caps on the trafo hv side will reduce the current drawn from the VFD (by compensating the motor lagging current) so that I don't have to use a large one.

*Why a man thinks he outrun a chasing dog when it has twice as many legs?*

 

[Skogsgurra]

Almost there now!

Capacitors on the output side of a VFD is not anything you should use. The PWM waveform has lots of medium and high frequency components that 1. will kill the capacitors and 2. trip the VFD because the capacitors represent a short circuit to it.

That is true also if you put the capacitors on the MV side. There will be a slight reduction of the current due to stray inductance in the transformer. But they will still be very close to a short for the medium frequencies.

The truth is that you do not need to worry about motor PF. The magnetizing current is supplied by the VFD. Capacitors would also be very tricky to use because of the frequency going from zero to rated frequency. You will probably be undercompensated most of the time. And overcompensated if you want to run over base speed.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[LionelHutz]

The output of the transformer should have fairly sinusoidal voltage and current waveforms. It will have some residual voltage and current distortion from the VFD pwm output but not enough distortion that capacitors could not be used if you do it carefully. I've never looked at the kVAR draw on the output of a VFD at different speeds so I can't comment on the motor reactance being a fairly linear function of freq allowing it to be corrected by capacitors.

The other problem I see with doing capacitors is that you need to change the capacitors for every different motor size you test.

The best solution is to just pick a transformer and VFD that are large enough to just work without any crutches.

 

[Skogsgurra]

Yes, if you have an elevated carrier frequency, then the output voltage may be sinusoidal - almost. But, that does heat a 50/60 Hz material core a lot. So you will have to use a rather low carrier frequency. The core will still get hot, so it is wise to use as low a carrier (switching) frequency as possible. I have used 1.6 kHz in one application. And still had to keep an eye on core temperature.

It might be a good idea to add a filter between VFD and transformer primary. But all these details are, well, details. Important details, though, when you decide to build the system.

The paradox is that, if you use a filter, it is usually better to run with an elevated carrier frequency. Simply because the filter components will be smaller then.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[rmgarnett]

Don't be drawn into the power factor on the supply side as issue. In addition to this, as stated earlier in the thread you don't have to consider capacitors on the output. What is important is the step-up transformer on the output, the transformer has to be suitably designed to take care of the DC offset from the VSD output waveform, so that it does not saturate the transformer. What you also need to consider is magnitude of the transient switching, from the VSD output. The VSD output waveform will be transformed together with its switching spike's, these spike's may be higher than that of the insulation rating of the motor and may cause some damage. To mitigate this isulation damage you could put a du/dt filter on the output, but you will have to factor in the additional voltdrop across the filter with the transformer ratio.

I hope that this helps!

 

[Skogsgurra]

There is no DC offset in a well designed VFD output. What make have you seen that in?

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...