Supply transformer resonance

[Marke]

I have recently looked at a number of installations involving VFDs on a stand alone transformer where I believe that there may be a resonance issue.

The transformers are rural, commonly fed by 10 + KW of overhead 11KV or 22KV lines and are feeding a VFD. Usually, there is no other load.
The VFDs have been installed in a manner to minimize the common mode leakage currents and voltages.
The VFDs are fed via a broadband passive filter.
Transformers are generally in the 100KVA - 500KVA sizing.
Supply is MEN.
On the supply side of the VFD, there is a ripple voltage appearing which is synchronized to the switching frequency of the drive. When the drive is stopped, the ripple disappears.
Altering the configuration in a manner that has a major impact on the common mode leakage, has no impact on the ripple voltage.
Adding capacitance (say 10KVAR) on the output of the transformer significantly increases the ripple voltage.

I recently came across one installation with an active front end drive where the ripple voltage phase to neutral (230V P-N 50Hz) was 128Vpk-pk and this burnt out two electricity meters and a PQ meter after a very short time.

In most cases, there is a length of underground cable on either the input or the output of the transformer.

It looks like the transformer is resonating with the cable capacitance and being stimulated by the common mode impulses from the drive. The magnitude of the resonance appears to be load limited rather than common mode amplitude related.

Can anybody suggest a practical solution to this? I know of one installation where they changed the transformer for a different brand and the problem was cured.

Best regards,
Mark

Mark Empson
Advanced Motor Control Ltd

 

[Skogsgurra]

The strive for lower losses and less harmonics sometimes result in high losses (financially) and extreme harmonics in the higher end of the spectrum.

You are quite right about the leakage inductance and the cable capacitance. The PWM edges in the AFE starts ringing in the weakly damped LC resonance circuit and, if the AFE carrier frequency is about the same as the ringing (resonance) frequency or a harmonic/subharmonic, the ringing will be a sustained and usually increasing phenomenon.

In some cases, it is possible to lower the Q of the system. But that seems to be impractical i this case (much heat and no discrete capacitor to connect in series with) so other ways have to be found.

One possible way is to change carrier frequency so it doesn't continuously add energy to the resonant circuit. That may be possible, but be very careul so you do not get closer to resonance instead! It is a bit risky, but it may be worth a try. It takes some time to "swing it up" and if you are observant - you may be able to swith off before it gets too bad.

If there is a way to introduce random variation of the carrier frequency (Siemens used to have the "Siemosound" option and other makes may or may not have that possibility) then the wider frequency range avoids "staying tuned" to the resonance.

And, as a final solution, get rid of that AFE. It often does more harm than it helps.





Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[sibeen]

What about throwing a choke at the front end of the VSD?

 

[mikekilroy]

I assume you are not concerned about the one case you once saw of an active front end drive? If you ARE, in addition to the other drives, then it may be worth looking at the AFE drive and make sure it was installed to mfgr specs. It is often required to have something like <3% Z feed them (Siemens), and a special input inductor. If either of THESE is missing on that installation, all bets off.

I agree trying different PWM freqs is good first try. Then adding series inductance (input choke) if application can handle the voltage drop, would be a potential 2nd try.

You mention dedicated xfmr: is it D-Y Y-Y? D-D? Perhaps if not D-Y then consider asking the drive mfgrs if changing to THAT would help with its inherent odd harmonic cancellation.

http://www.KilroyWasHere<dot>com

 

[Marke]

Hello Mike

The transformers are D-Y here.
Problems are not only with active filters and AFE, but also just standard drives. Moving the carrier frequency shifts the frequency of the ripple, but not necessarily the amplitude. I think that the transient edge of the common mode voltage creates a ringing voltage at the resonant frequency, but we are not at that frequency.
Next time I go to a site, I will have a good measure up of the actual ringing voltage and should be able to determine the frequency from that.
It does not seem to matter what we do to the drive installation, the ripple remains at close to constant amplitude until we stop the drive.
I am looking for ways of reducing the Q of the circuit without dissipating a lot of heat. Perhaps some form of snubber network that exhibits a low resistive impedance at around 5KHz but a high Z at 50Hz.
Might have to have a bit of a play when I get a chance.

Best regards,
Mark.

Mark Empson
Advanced Motor Control Ltd

 

[mikekilroy]

I assume you took the 5khz down to 2.5 and upto 7-8khz? Seems mighty strange, if it really is a resonant oscillation, that changing the alleged exciting freq up and down by factor of 2 each way let it still resonant.

I would be real interested in seeing what this oscillation looks like. Please post a scope pix if you can.

http://www.KilroyWasHere<dot>com