VFDs on same transformer as DC drives 5

[Skogsgurra]

In a recent thread ( thread237-178054 ) there was some concern about influence from PWM VFDs on the three-phase system supplying DC drives. I mentioned that I was going to measure on such a system and here are the results:

The top picture shows voltage from phase A (L1) to ground. The superimposed square wave is from the VFD. It has a 2500 Hz carrier frequency and the impedance in the transformer windings *plus* the fact that the neutral is isolated result in a very high common-mode carrier residual voltage.

The DC motors have a very tight coupling between armature winding and rotor (wound directly on rotor) so the HF components are coupled to the rotor very efficiently. Once there, the charge finds its way though bearings back to ground.

Typical discharge patterns were found between shaft and frame and there were also motors that had wash-board patterns in the bearings.

We also measured other transformers with and without VFD load. One transformer with about 50 constant drives (asynchronous induction motors) had a much lower HF common mode voltage. We think (are sure) that that is because there are 50 long cables shunting the HF to ground.

So, it is not only motors connected directly to a VFD that are damaged by EMD - also (DC) motors that have been running reliably for decades can get hurt if a VFD is installed on same transformer.

Comments invited.


Gunnar Englund

http://www.gke.org

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

 

[dpc]

Very interesting, Gunnar. Does the VFD have any type of line reactor or other filtering? A standard 6-pulse front end?

How large is the transformer and what % of the transformer load does the VFD make up?

 

[Skogsgurra]

Yes, the VFD has a line reactor and has a 6-pulse simple uncontrolled rectifier. Transformer is 2 MVA. Uk not known. Probably around 6 %. VFD is around 20 % of transformer.

Gunnar Englund

http://www.gke.org

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

 

[ScottyUK]

Thanks for posting this Gunnar. Intriguing.

From the look of the loads nothing actually needs the neutral to be grounded, but what benefit is brought by leaving it isolated? Ability to continue running with a single ground fault, or something else? 525V looks an 'odd' voltage for Europe - is it originally a specially constructed transformer dedicated to the DC drives which has been retrofitted with the AC drive?


----------------------------------

Sometimes I only open my mouth to swap feet...

 

[dpc]

Excuse the dumb question, but is the HF carrier being coupled to the dc drives via the common 575 V bus and wiring? I guess I'm used to seeing standard 5th and 7th harmonics from rectification at the drive input, but not normally the output carrier frequency at such a magnitude.

So what's the solution - tuned filter on the drive input?

 

[Skogsgurra]

The thinking behind the isolated neutral (IT system) was/is that a continous process should be able to continue running if/when an isolation fault happens. It is common in paper mills and steel works.

There are some drawbacks with the system (earth faults never seem to get located, so the system turns into a corner grounded system after some time). Some steel works and paper mills have started grounding the neutral, turning the IT system into a common TN-C system.

Yes, the transformer was installed for the DC drives back in 1985 or so. The VFD for the fan was installed on same transformer a couple of years ago. The fan was a constant speed drive with dampers before the VFD was installed.

Gunnar Englund

http://www.gke.org

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

 

[Skogsgurra]

dpc,

My "mental model" is that the PWM signal gets capacitively coupled from motor cable and also stator/fram capacitance of motor to ground. So, there is a rather low-impedance "connection" from inverter output to ground - at least for the 2500 Hz and its harmonics.

The resulting 2500 Hz square wave current wants to "go home" and the transformer impedance as well as the isolated neutral stops it from doing so. The result is the common mode square wave we see on three phases. We made quite a lot of measurements and transformers with more cables attached (se OP) had a lot less 2500 Hz.

We haven't decided what to do yet. Common mode filters, ground neutral, shaft grounding brushes, move VFD to another transformer etc etc.

Gunnar Englund

http://www.gke.org

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

 

[Laplacian]

Thanks for posting Gunnar. Very informative. Our plant has various PWM VFD's and DC drives all playing nicely together on common sources.

This is just FYI, but we use a solidly grounded transformer neutral with no ground fault protection; rely on phase overcurrent coordination for proper selectivity of ground faults (and phase faults) on this system. This is in a petrochemical environment where process interruptions result in PSV's bursting and flares flaming.

Consultants always recommend to convert to an ungrounded neutral, but proper maintenance and isolation of ground faults would not be performed.

 

[DickDV]

Skogs, I know you have much better test equipment and analysis skills than I do but, having seen this problem a few times, I suspect that you are exactly correct in stating that these pulses are coupled into the ground system which does not have a low impedance path back to the drive DC bus. As a result, it passes thru whatever route is available and, in this case, that route is thru the bearings of the DC motors.

One thing that jumps out at me is the lack of isolation transformers (delta-wye) feeding the DC drives. On floating or high resistance grounded systems, I will not commission any DC drive without an isolation transformer. The reason is simply that, with the power source corner grounded, the DC motor armature is cycled between ground and power supply voltage by SCR conduction. This cannot be good for the DC motor armature and its insulation. A reasonably good case could be made that that is hazardous.

Back to the PWM pulses, an isolation transformer on the AC drive with a wye secondary would provide a short path for the pulses back to the DC bus but this would be expensive due to the large size.

I have to wonder if, on the DC drives, all the grounds could be collected together on one system and routed thru a suitable reactor thus blocking the pulses in the ground connection. While this might be cheaper if the DC drives are small, it still isn't ideal because the pulses are going to be redirected somewhere else in their search for a path back to the AC drive.

I believe the best path here would be to fix the noisy ground problem at the source which would be the isolation transformer on the AC drive.

Not a pretty solution but one that has always fixed the common-mode ground noise problem for me.

Just one other thought, I suppose grounding rings on all the motor shafts would protect the bearings but, again, it simply moves the problem somewhere else. I would protect the AC motor bearings as well since it seems vulnerable along with the DC motor bearings. The Aegis rings made by ElectroStaticTechnology in Maine USA are good units for this.

 

[Skogsgurra]

Thanks for comments. Yes, the first thing we will try is grounding the transformer neutral. But we may have a problem with the transformer stray impedance. It will probably leave quite a lot of residual voltage at the higher frequencies.

I would love try ATP on a simplified circuit with transfomer, drives and motors. But, although I have got a license and a password, I haven't figured out how to get started yet. First problem is: from where do I download and then what? I am not unfamiliar with simulation as such, but I need a pointer here. Anyone?

BTW, I have tried the ETS AEGIS on heavy paper machine drives. We had a problem with grease and dust baking into the fibres. I have also tried it on fan motors and there we had a complete success. Took shaft voltages down to safe levels.

Gunnar Englund

http://www.gke.org

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

 

[ScottyUK]

Skogs,

These might be a little out of the size range you're normally looking at but as a possible alternative to the Aegis rings Sohre Turbomachinery has a range of shaft grounding brushes primarily designed for draining the windage-induced charge buildup on big rotating HV plant. I haven't tried them yet but I'm hoping to do so fairly soon. Anyone who has tried them, your experiences would be most welcome.

http://www.sohreturbo.com

----------------------------------

Sometimes I only open my mouth to swap feet...

 

[thinker]

For similar installations we use VFD with drive isolation transformers (not reactors), delta/wye with grounded neutral. This greatly reduces penetration of HF common mode voltage to the upstream (group) transformer feeding multiple drives (AC and DC). May be this is the case when transformer has more advantages in comparison with input reactor?

 

[electricuwe]

Really intresting discussion, let me just add my comments on the issues discussed:

- 500 V (or 525 V) has been used in european industy where a significant number of induction motors in the range of several 100 kW have been in use. Lower currents as for 400 V but no expensive MV motors and switchgear. Newer installations use 690 V.

- Ungrounded systems: Used to ensure continous operation even in case of ground faults. Equipment used has to be desinged to withstand higher voltage stress and isolation has to be monitored to avoid grounded corner operation for an extended period of time. There is special equipment available to search for ground faults in such system during operation.

- isolation transfomer:
an isolation transformer would solve that problem, but at siginificant cost. The root cause for the superimposed voltage is that the inverter "sees" significant winding cabling and winding capacitances against ground at it's output (or -maybe- even some kind of filter ?). Switching these capacitances by the inverter will move the dc-link and all other points of the inverter up and down in relation to ground. The use of an output (or input) filter specificly designed for this purpose would probably solve the problem with lower cost than using an isolation transformer.

 

[HCBFlash]

Gunnar, you never fail to impress me. Thanks for your work and these postings. At least to me these things are most fascinating and enlightening.

In my experience, limited though it is, we're seeing steady movement away from isolated ground systems here in the USA. -and I suspect some of it is from ignorance. Even airports are having more trouble with getting good engineering and maintenance on their ungrounded systems. In any case this situation has some benefits, and I'm looking at this post's topic in this light.

My own view from this distance is that you are also bumping more and more into a firmer and firmer reasoning to make a specific, planned change over to solidly grounded systems. Your 2MVA wye secondary as a substantial first step. An option everyone immediately recognizes as impracticable is that of an entirely ungrounded system, including motor frames, structures, metallic raceways.......and human life on the premises if need be! (Do you remember when "double-insulated" power tools were the latest thing?) But physics is simple enough for even me to understand that a ground loop is just another circuit.

Certainly isolation transformers will help and a different, but hopefully smaller and more acceptable set of problems will accompany that move. All of this says "VFDs NOT on same transformer as DC drives", which is to say "forget it, next problem - or next topic - please". Fine tuning of your bus capacitance(s), load and line side reactors, conductor lengths AND placement, conductor shielding, are all apparently the only other alternatives.

Is frequent maintenance and replacement of motors and / or bearings and all the labor and production impacts a pragmatic financially-based answer to this electrical question? I think the application of more and tightly specified transformers is a technical and electrical answer which needs an accurate, finely engineered financial solution to be properly applied.