Generator sizing with multiple large VFDs

[32Lateralus]

Sizing a generator for an existing building with multiple large VFD loads. I have seen multiple websites saying when sizing, take 2x your VFD load.

Are there other special considerations to take? With a lot of VFD load how big of an issue is harmonics on an existing site when adding in generator power? Does the 2x VFD load account for those harmonics or should a harmonic analysis be run in order to determine the effects on the generator and if harmonic mitigation should be put into place?

Appreciate any insight and advice on the subject.

 

[zbang]

Do the VFDs have line-side reactors (to mitigate harmonics)?
The generator companies should have guidance on this, too.

 

[texie]

All the big vendors of gensets have software and engineering help for this. You can't just go by size of machine and it will vary amongst brands. Details like VFD, elevator regen loads, etc have really complicated things and one needs to be careful.

 

[petersonra]

Do the VFDs have line-side reactors (to mitigate harmonics)?
The generator companies should have guidance on this, too.

Why would the generator care about harmonics?

 

[Besoeker3]

Why would the generator care about harmonics?

Non sinusoidal current will increase current.

 

[winnie]

Harmonics increase the KVA requirement, but shouldn't increase the size of the prime mover required.

-Jon

 

[Besoeker3]

Harmonics increase the KVA requirement, but shouldn't increase the size of the prime mover required.

-Jon

Yes, the increased kVA. That was my point.
But you are correct about the prime mover.

 

[zbang]

Possibly more accurate to say that mid-cycle turn-on (e.g. "phase control") can confuse some governors. Might not be a problem, but I'd at least consider the possibility before dismissing it.

 

[synchro]

The paper at the link below from Rockwell has some good information and discussion about using VFDs with generators.
One of the issues covered is the lowering of the peak voltage of the generator waveform by the high current drawn by VFD rectifiers at these peaks. This can occur because a generator will typically have a significantly higher output impedance than a POCO derived source. And since generators commonly regulate the level of the RMS output voltage and not the peak level, the peak ouput voltage will be reduced in this case. This would end up lowering the DC bus voltage on the VFDs.

Another issue mentioned is that VFD rectifiers using phase controlled SCRs can distort the voltage waveform and also modulate the instantaneous speed of the generator because of the limited mechanical inertia. As zbang mentioned this could cause issues on some governors.

Quite a few other practical considerations and ways of mitigating problems are mentioned in the paper.

https://literature.rockwellautomation.com/idc/groups/literature/documents/wp/drives-wp014_-en-p.pdf