VFD: common DC bus & motor voltage 1

[thinker]

Gentlemen,
We are ordering a VFD system (PWM based, with active front end/regenerative) and need to ensure drive and motor compatibility, since they are provided by different vendors.
Drive vendor states that for the system with common DC bus and more than one inverter/motor, each of multiple motors will see on its terminals voltage spikes increase about 50% (peak value) more than it would have for single inverter configuration. In their opinion, the reason for this increase is a common mode voltage brought into drive via motor feeders and, related to that, the increase of voltage spikes level on DC common bus. Based on that, they require that the motor insulation must be upgraded to handle these additional spikes due to multi-motor configuration. This puts a burden on motor vendor. In numbers: drive input voltage is 480V. Motor insulation is designed per NEMA MG-1, part 30 to withstand 1600V spikes (for inverter duty motors). Now, drive vendor claims that each of motors must have insulation rated for 2400V if both motors will be fed from the same common DC bus drive (via separate inverters, of course).
Questions to VFD experts: have you ever heard about this phenomenon? Is this possible that these claims cover up drive design deficiencies and put unnecessary burden on motor designers? Any links to information to better understand the issue?

 

[Skogsgurra]

thinker,

I have been doing paper machine and printing press drives for decades. And PWM inverter drives the last twenty or so years.

Never heard of that problem before. Always use standard inverter duty motors in drives with a common DC bus. Not aware of any other motor class - is there one at all?

I think that it all stems from some myth that has thrived in an atmosphere of ignorance and superstition so long that it has evolved into some sort of a fact. Which it is not.

Gunnar Englund

http://www.gke.org

 

[jraef]

I agree with Gunnar. I have never heard that before, and if it is a truly regenerative drive, any excess energy would be put back to the line, not into the other connected motors. But be careful here. I have seen the term "regenerative" used to describe a system where motor energy is only fed to other drives on the same DC bus, not truly back into the line as it should mean. If that is the case in your vendor's system, then you have do deal with what happens to excessive energy as well, be it DB resistors or true line regeneration.

If for instance your vendor's drive is rated for a 500V supply (as many European designs are), it is capable of an 865V DC bus, so technically the peaks of the PWM output could be slightly higher if the DC bus is allowed to charge that high with regeneration from other motors. If you then also get capacitive coupling in the motor leads it will be higher because of that, but over 1600V? I don't think so, but the solution to that IMHO would be to deal with the motor lead and the excess energy issues individually, not try to build a special motor design.

"Our virtues and our failings are inseparable, like force and matter. When they separate, man is no more."
Nikola Tesla

 

[thinker]

Jraef,
This is a true regen drive with 4 quadrand operation. This is a dynamometer system with 2 absorbing AC dynamometers, and all regen energy is fed back to the grid via common front end converter (also PWM based). BTW, both drive and motor suppliers are European companies providing equipment for US market.

 

[Skogsgurra]

Sounds very much like a Siemens/Schorch setup. If that is so, you will not have to worry. Seen many of them (with AFE) and there are no problems at all. Except for one case where there were some rotor bar problems. But that is another story.

Gunnar Englund

http://www.gke.org

 

[thinker]

Gunnar,
This is not Siemens/Schorch, and I do not feel comfortable not understanding a technical background, also I have about 30 years of AC drives development and application experience.

 

[jraef]

Huh, I don't see where they get the overvoltage problem then. I am looking at a system like that right now from an EU supplier. I will check to see if they make any mention of that as an issue. I used to work with ABB SAMIStar drives years ago but they never once mentioned it and they have probably been doing this as long as anyone. I think that is what Skogsgurra has worked with as well. This could be something newly discovered however, we learn (or should) all the time.

Who's drives are you using?

"Our virtues and our failings are inseparable, like force and matter. When they separate, man is no more."
Nikola Tesla

 

[thinker]

These are ALSTOM drives and OSWALD motors.

 

[Skogsgurra]

Sorry, no experience with them.

Gunnar Englund

http://www.gke.org

 

[jraef]

Me either, but I will say that they are a large player eleswhere in the world so I would tend to think they would know what they are talking about. However, did this information come in the form of a white paper or some other printed document, or is it the verbal nattering of a salesman who is trying to impress you by waving a red herring?

"Our virtues and our failings are inseparable, like force and matter. When they separate, man is no more."
Nikola Tesla

 

[Skogsgurra]

OK, it could be the fact that an AFE can run with a slightly higher DC link voltage that is behind this reasoning. Since the AFE (active front end - an extra inverter working as a rectifier on the input) allows for higher DC link voltage than a "normal" inverter has.

But this has nothing to do with more motors being fed from a common DC bus. Not at all.

Read the

http://www.rema.uk.com/pdfs/vsd1.pdf

discussion. Snip: "For drives with PWM active front ends (regenerative and/or unity power factor), the effective supply voltage is increased by around 15% and this should be taken into consideration e.g. treat a 480V application as if it was supplied with 550V"

I wouldn't be too alarmed. A standard 550 V inverter duty motor should be OK. A du/dt filter is always nice to have if your cables are long. And then you do not have to worry at all.

Gunnar Englund

http://www.gke.org

 

[ozmosis]

A test-rig of this nature is quite an investment and if you're not sure of the compatibility of motor and drive in the US, I would double check. Oswald is very much an 'old-school' traditional German manufacture. This is not meant in a derogatory way, just that they often overlook issues such as high AC supply voltages (480V often becomes 500~510V I find) that are quite alien to German power grids. The AFE will increase the DC link voltage (it has to) and as Gunnar points out, will impact on the motor some way as the peak-peak volts rise. The insulation class of Oswald motors will be good. They sometimes use 70Hz wound motors specifically for inverter use and often these are also designed around 400V.
Maybe a storm in a teacup but I'd check.

 

[DickDV]

As for specifics, a DC bus on a non-regen drive will charge up to a few volts less than the peak of the incoming AC. For 460V, that's around 650VDC and for 490V, its about 700VDC.

When operating on a regen system (not necessarily the same thing as common DC Bus), the DC bus will typical run around the above-mentioned voltages when motoring, but when the motor is being overhauled as in an absorber or dyno, the DC bus will rise to a higher voltage from the absorbed energy until it reaches the trigger voltage for the regen inverter to begin working. This can be typically around 780-800VDC. This is normal and to be expected.

It follows, that under regen conditions the pulses going to the motor will be at the higher DC bus level. And the motor insulation will have to deal with these higher levels. However, in my view, as Gunnar has stated above, with reasonably short motor leads, a motor with an MG1-Part 31 endorsement will be just fine. We are, after all, talking about a motor tested to 1600V peak pulses, not 780-800V.

Of course, there are always some ringing and other pulses that appear on the motor leads but these should not make trouble. If motor leads are longer, then the usual suppression methods, motor lead reactors, dv/dt filters, or in extreme situations, sine filters may be needed.

Further, I have no idea where you would go to buy a motor tested to 2400V. Canada maybe?

 

[thinker]

Thank you for your inputs, colleagues. But I am still struggling to understand how the addition of another motor would result in so dramatic voltage increase (as suggested by our vendors) on the terminals of all motors.

 

[Skogsgurra]

I think that you can stop worrying about that. It is probably a mistake from the vendor's side and the reason for the mistake might be the higher DC voltage made possible by the AFE.

If there are any consequences w/ regard to guarantees etcetera to using a correctly sized standard inverter motor, then I think that you should ask for a solid technical explanation that you can understand* - or find another vendor. They do exist.


* We would love to hear the explanation!

Gunnar Englund

http://www.gke.org

 

[ozmosis]

I worked on a project about 4 years ago for a Landrover 4wd testrig using an AFE plus 4xinverters on the DC link. I got to meet up with a few highly technical gurus in our company (Siemens) and have asked them the question you pose. Just got an out-of-office reply from one of them so hopefully will shed light from another angle early next week.

 

[ozmosis]

thinker
just got a reply from one of the gurus in Siemens (even though he's on holiday-good chap!).
Basically, as Gunnar outlined originally, the AFE will increase the DC link volts by about 20% and this needs to be taken into account with the motor insulation. Assuming the motor has inverter rated insulation (curve "B" in the GAMBICA/REMA guide mentioned in the link Gunnar provided) then there should be no problem. The motor rated for 1600v should be ok. It makes little difference as to the number of inverters on the bus. If the engineering has been carried out properly and you have axis driving as well as being driven, then it should equal out on the DC link. One of the real benefits of an AFE is the fixed nature (albeit high) of the DC link. This is why sudden load changes show little drop and likewise regeneration is either fed into another inverter module that is being driven or goes back into the grid.
If you are concerned then you could fit dv/dt chokes per inverter/motor. As a footnote, this is advise from Siemens not Alstom. I think the 50% increase stated is a little steep myself.

 

[cswilson]

Thinker: I have a feeling that the reason for the caution you received is that the whole point of ganging multiple motors/drives together on a common DC bus is to be able to have short-term energy storage on the DC bus, where it can easily be used by the next motor requiring a lot of input power. This storage, of course, pumps up the bus.

Now, it is not the addition of other motors to the bus that causes the problem (unless you somehow did it so badly that you could get wild oscillations on the bus); but systems with multiple motors on the same bus are more likely to see higher bus voltages because of the way they are used.

My guess, anyway.