Running Multiple VFDs on a single generator - re: thread237-451584 1

[kurt klingbeil]

re: thread237-451584

What are the implications and tricks to running multiple VFDs on a single generator ?
Apparently it can be problematic:

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

We have experienced high-kVAr trips when running a single VFD from a single generator and
typically bypass the input filter - in order to reduce the net source impedance seen by the drive

I imagine that running multiple drives, without filters, could possibly result in a cytokine-storm
of harmonics and reactive power possibly wreaking mutual havoc on the VFDs and on the generator.

I assume there is some load-factor below which there is no problem...
i.e. running 4 1HP drives on a 200kVA gen would likely be no problemo
however, as the aggregate load approaches some limit, instability results...

 

[jraef]

John Streicher at Rockwell wrote that White Paper in 1999, that is an EPOCH in terms of VFD technology. Much of what he was concerned with back then is no longer valid, both for VFDs and for Generators. Generator mfrs have, for the most part, updated their AVRs to be able to deal with non-linear loads, because so many loads have some sort of non-linear aspect to them now; HID, CFL and LED lighting, VFDs, Servos, Power Supplies for computers, PLCs, communications systems etc. etc. etc. Those that failed to adapt fell by the wayside. In addition, smaller VFDs have Diode front-ends, not SCRs, so the issue of them not tracking the distorted waveform is no longer valid until you get to larger (i.e. over 100HP) drives, in which case you can simply use Active Front End drives now.

An option / alternative if you have a lot of VFDs on a generator is to add a separate Active Harmonic Filter to the bus, it will measure and adapt to the harmonics to neutralize them. Several of them also will correct the Power Factor for you, which helps you to get more out of your generator.

If your generator is more than 10x the kVA of any VFD, you should add a Line Reactor ahead of an "un-buffered" VFD to protect it, but that rule applies whether you have a generator or utility source. Un-buffered means a VFD with no DC bus choke, which will be the case on 99% of the small Asian sourced VFDs under 30HP. Buffered VFDs designed for the North American and European market generally will have DC bus chokes inside, in which case that formula is 20x the kVA of the drive before needing a Line Reactor.


" We are all here on earth to help others; what on earth the others are here for I don't know." -- W. H. Auden

 

[itsmoked]

The problem is the that they ALL will be sucking power only at the generator's sinewave peaks. They will all do it in lock-step. This can be helped by running input chokes, including a DC choke. You can also use a transformer to change the phase angle from the generator source to spread some of the VFD current peaks to elsewhere.

Cross post LOL.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[kurt klingbeil]

Wow,
Thank you jraef and itsmoked for your super-speedy responses.

In the equipment we build we have been adding a breaker to bypass/short-out the input filter when running
on generators to prevent high kVAr trips. This isn't a problem when running on line power.

I am curious whether streicher's suggestions about 12-pulse transformers/VFDs are still relevant - which is basically what itsmoked is saying.

Our current proposed design is to run 4 20HP pumps via 4 30HP VFDs from a 180kW generator
with the view to expand to a total of 7 drives/pumps

I see there are even 18-pulse drives now

https://americas.hammondpowersolutions.com/en/resources/simulate-12pulse-vfd

so, it sounds like some combination of delta-Wye transformer, 12-pulse drive, AC-choke, DC-choke, and active filter well be optimal ?
I'll fly this by our supplier for quoting and see what they come up with...

We will have a 480-600 step-up xfrmr at the Mine-Compliant system (NGR and ground-fault/groundcheck) next to the generator.
I'm wondering whether it would be possible/plausible to get a Wye:Wye-Delta centralized transformer... but that would likely
result in phantom ground currents, plus one needs both delta/wye phases local to each drive which will be on separate skids