Harmonics with VFDs and Wye-Wye Transformer

[irmalcol]

Hello,

I am looking for some opinions regarding a plant distribution system I am investigating with the following configuration:

There is a 1500kVA 25kV/480V wye-wye transformer feeding the plant. The bulk of the load is comprised of 2 600HP VFDs. The primary neutral of the TX is solidly grounded, and the secondary neutral has a 2A NGR. I am not able at this time to determine the exact construction of the transformer core, but I do know there is no tertiary delta winding. Large harmonic distortion has been measured on the utility side of the transformer, and 5th harmonic filters installed at the VFDs have failed due to overheating.

I am trying to determine whether the transformer is causing any problems or if the VFDs are solely to blame for the harmonic mess here. A couple of questions I have:
1. Could the transformer be contributing to or exacerbating the harmonic issues with the VFDs? Further, would the transformer only cause 3rd harmonic issues, or would it play a role in higher-order harmonics as well?
2. Am I right in thinking that harmonic distortion/zero sequence currents are passed easily to the utility with a wye-wye configuration?
3. Does the 2A NGR have any bearing on these harmonic issues?

Any thoughts anyone has on this problem would be appreciated.

Thanks.

 

[HornTootinEE]

I'll take a shot based on my limited experience with this type of situation:

1.) I'd say the transformer itself is not contributing (adding) to the harmonic content as it's a linear device-being it's not altering the 60Hz signal.

2.) Yes, a Y-Y will pass triplen harmonics (3,6,9,etc.) into the utility system. Put on a delta low side and you'd no longer see those on the utility side.

3.) Not sure on the NGR, but depending on configuration of the windings and such, it may cause higher harmonic voltages between neutral and ground which could be transferred to the utility side. There is probably somebody out here that knows the effects for a fact though, I'm just guessing

This customer probably needs an active harmonic filter or a better (18 pulse or higher) VFD. At the utility I was with, whenever we hooked up a VFD it was rare that we were able to allow anything under at minimum a 12 pulse VFD w/active filtering. If you have a very large amount of linear load compared to non-linear, you may be able to get away with a 6 pulse, or some minor filtering, etc., it all depends on how much fundamental is present in relation to harmonic currents and voltages to "water down" your harmonic distortion.

 

[irmalcol]

Thanks for the reply, HornTootinEE.

For any interested parties, the VFDs are 6 pulse. This is really just for your information, though, as I am mainly concerned with the transformer here (others are looking into problems with the VFDs).

 

[HornTootinEE]

I'd say (in general) that if you are seeing too many utility system harmonics, you should install a delta low side transformer (if allowed by customer) and/or require them to improve their vfds and filtering. 6 pulse vfds will definitely be spitting out alot of harmonic content. Remember though if your mitigation technique is a delta low side transformer, that winding will take the load and the heating effects of the harmonics directly, so instead of sizing that tub for 1200 HP @ 85% pf (just a rough stab) you'll be sizing it for 1200 HP @ approx. 50-60% true pf.

I don't think a delta low side will eliminate 5th, 7th, 11th, etc. harmonics from the utility system though.

Whenever I've connected similar customers (new connection), their VFD and/or THD requirements were stated in a service contract. It's a little late on this one, but for future similar loads its something to think about.

 

[irmalcol]

All good things, HornTootinEE. Thanks again.

One other important question I have is whether this configuration is likely or able to produce circulating currents in the transformer through the neutrals? I am fairly unfamiliar with wye-wye transformers, and the NGR is of unusual rating.

 

[HornTootinEE]

You shouldn't get transformer circulating currents through the neutral on this transformer. Using an NGR on the low side and solidly grounded on the high side would suggest that internally the neutral of both windings are independent of each other i.e HO and XO are truly independent. On some padmount transformers in Wye-Wye you'll see both HO and XO grounded together internally, but in this case that isn't likely.

Transformer connections can be tricky devils.

This case really sounds to me like the customer needs some pretty serious filtering or needs to upgrade their VFD to something like a 18 pulse. Remember also that you could likely be overloading your transformer too. If we make some assumptions, your two (2) 600 HP motors at say 90% efficiency with a 6 pulse VFD total pf of 60% you would see approx. 1650 kVA on that transformer, or 10 % overloaded.

Who called the transformer into question? Do you work for the plant or for the serving utility (either direct employed or contracted consultant)?

 

[irmalcol]

The engineering company I work for was hired by the plant to investigate these problems. I have little access (and no immediate access) to the utility side of things (i.e., current in the primary neutral, the core configuration or nameplate of the transformer). After speaking with my supervisor today, it seems the harmonic problems are being addressed. What I am really concerned with now is strange neutral currents in the transformer. It was found that only a small percentage of single-line ground fault current on the secondary side was coming back through the NGR. Is it possible (or impossible) for this current to be going to the primary solidly-grounded neutral, as it is a lower impedance path?

 

[sibeen]

Theoretically, with a balanced source voltage and ideal switch of the rectifier of the VFD, you shouldn't see any third (or other zero sequence) harmonics generated by the VFD.

In practice, there will be a small amount, but this will be swamped by the other harmonics - especially if the VFD is a 6 pulse without filtering.

 

[ausphil]

a couple of things:

Is this plant in an industrial area where other plants may be contributing harmonics? Your 5th harmonic filter may have been burnt out either from being underrated for your VFDs, or it was correctly rated and it is also acting as a sink for some other plant's 5th harmonic generation.

Is there a connection between the 2 neutrals within or external to the transformer? The design may have connected the 2 neutrals and given a single ground connection. This could cause a return path for your secondary fault current to flow and bypass the NGR. You could measure the resistance between the neutral and ground across the NGR and see if it is the rated resistance or a short circuit (off line of course!).

ausphil

 

[irmalcol]

Hello all,

I have an update to this problem. A solution to the harmonics issues has been found, so there should not be any further problems with harmonic filters melting or VFDs failing. There is still, however, an interesting phenomenon observed with the ground and neutral current at the MCC:

When the frequency output of the VFDs is ramped up (or down), the current observed in the ground bus of the MCC ramps linearly with the phase current, reaching up to 29A. However, the current in the NGR seems to spike with a change of frequency before leveling off to a small magnitude steady-state oscillation. The spikes reach no more than 0.7A.

I am trying to figure out (1) where all the ground current is coming from and (2) why it’s not coming back through the NGR. I remind you that the 25kV to 480V transformer is wye-wye, primary neutral solidly grounded, secondary neutral connected to a 2A (138.5 ohm) NGR, and the other side of the NGR is connected to the MCC ground bus. The ground bus of the MCC is solidly connected to the transformer primary neutral (grounded).

Any thoughts?

 

[HornTootinEE]

Is the X0 (secondary neutral bushing) running over to an isolated neutral bus on the MCC? that is, are there seperate ground and neutral conductors from the transformer to the MCC?

 

[irmalcol]

There is no neutral bus in the MCC. Running from the transformer to the MCC, there are the 3 phase conductors and the secondary neutral (goes to the NGR which is installed in the MCC building). The MCC ground bus is attached to the MCC building ground, which is attached to the transformer primary ground.

I hope that clarifies things.

Cheers.

 

[dpc]

Some VFD designs seem to have some issues with high-resistance grounding. You might want to browse this old thread: thread237-159625

 

[HornTootinEE]

Is the load a delta connected load?

 

[irmalcol]

Thanks for the link, dpc. I hadn't found that one yet.

HornTootinEE,

The input to the VFDs is a 6 pulse rectifier. I don't know whether you want to think of that as delta connected or not. In any case, the VFDs are 3-phase loads, and are presumably balanced, notwithstanding the power quality/harmonic issues they may produce.

 

[LiteYear]

A possibility:

(1) common mode voltage in the drive, conducted through the parasitic capacitance of the motor cables and motor windings
(2) because it returns to the drive via the harmonic filter instead

 

[irmalcol]

I've been wondering about capacitive effects as well. LiteYear, can you explain further how the current would return via the filter? I don't quite understand how that would work.

Thanks for the good suggestions.

 

[HornTootinEE]

Well, it would be "delta" in the sense that you would or wouldn't have a input neutral.

Dan J

 

[irmalcol]

There is no neutral running to the VFDs, just 3 phases.

 

[Kiribanda]

irmalcol,
Since your VFD is a 6-pulse unit it cannot generate any harmonics lesser than 5H.Theoritically it will generate 5H, 7H, 11H & 13H. Therefore, there cannot be multipliers of 3H.In my opinion it looks like there is some spurious ground path giving your reading when the VFD is ramping or 3H coming from external harmonic source.Why donot you connect a power demand anlyser and get a complete spectram with power flows at the 480 V PCC?