Delta-Wye VFD Power Problem

[wzrdstrm]

Dear Readers~

We have a VFD rectifier that has two bridges on it. One bridge is fed from a delta connection and the other from a Wye connection. The idea is to provide harmonic cancelation for a VFD system. We did not originally have input line reactors and the fuses to the rectifier Y bridge were blowing when we tried to start the VFD (just powering up the rectifier)(sized appropriately). We then put input line reactors in, which stopped the fuses from blowing when putting power on the rectifier, but after about 3/4 speed to the VFD the currents between Y bridge and D bridge (up to now about the same current) shifted, whereas the Y takes about 60% more amps then D. Even after we reduce the speed of the VFD, the imbalance in current stays about the same. Can anyone explain the imbalance and why the imbalance stays after reducing the speed?
Please and thank you,

 

[krsh]

What do you mean by "we reduce the speed of the VFD"?

 

[MAGTiger]

This kind of front end system is sensitive to impedance balance and transformer output voltage. It is unlikely first that the delta and wye transformer outputs are exactykl the same voltage, depending on transformer size Larger transformers have fewer turns and so voltage error can be significant. Also, winding impedance will be different for the delta and wye windings. It is for this reason that a line rector should be in the output of each transformer winding but should have complimentary impedance with the respective transformer windings. Such a setup will eb challenging to make work at better that 60/40 sharing. There are other transformer circuits with much better balance. Alternatively a scr bridge could be used which would allow the currents to be actively balanced as in the case with older large hp DC drives.

Neil

 

[Drivesrock]

Hi wzrdstrm

The transformer secondary voltages must be within a few percent (1-2%? but I'd have to check some specs) of each other for adequate current sharing. Some drives will then have reactors (or interbridge reactor) after the rectifiers on the dc side to further make the currents as close as possible, to compensate for impedance differences in the rectifiers.

I have seen just a 10V difference (on a nominal 690V Y D fed drive in that case) cause 70% current on the higher voltage transformer secondary and 30% on the other.

Presuming that the transformer secondary voltages are outside of the tolerance, I think your solutions will be either a new transformer but that will probably be the most expensive option (and therefore not an option!) or inline reactors to match impedances and voltage drops. This will depend on yr drive rating and installation, I suppose, amongst other.

I am intrigued as to why the Y side incoming fuses were blowing at power-up. Maybe the pre-charge circuit was not working correctly?

 

[wzrdstrm]

KRSH,
The drive is oerating in encoderless vector, but what I mean is when we reduce the frequency to some lower level. The motors are 40HZ 788rpm 600V. We drive these to a about 150% of rated Hz, it appears that the problem arises after 40hz, and this imbalance stays when we lower the Hz back down to any value (except 0 of course)

MAGTiger,
We have a 7500KVA xformer, 11KV fed, to a fixed secondary of 640V, unloaded. Supposedly the transformer has an additional phse offset of 7.5 degress, to add to the already 30 degress natural offsetting of a delta/wye system. I am aware that if they delta wye voltages are different, we not only lose harmonic balance, but causes severe power problems as a result. What do yo mean by a complimentary impedance?

 

[itsmoked]

He means the reactor's impedance can be adjusted (by model selected) to make the transformer/reactor impedance of the D bridge supply, match the transformer/reactor impedance of the Y bridge supply.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[sibeen]

Supposedly the transformer has an additional phase offset of 7.5 degress, to add to the already 30 degress natural offsetting of a delta/wye system.

Why is an additional 7.5 degrees added. Surely this is going to force an unbalance in the currents drawn between the two bridges and would also cause some fairly atypical harmonic currents to be drawn as well.

Isn't it?

 

[MAGTiger]

Right, complimentary impedance to make the 2 sides equal, like Keith said.
Do you mean an additional 7.5% between the delta and wye outputs or are they still 30 degrees but both are 7.5% from the primary. To 12 pulse properly, you must have 30degrees between the 2 sets of 3 phase output. Anything else will result in further error in harmonic cancellation.

Neil

 

[wzrdstrm]

Thank you all for you comments, good stuff here. The additional 7.5 degree is added in to make it a 24 pulse harmonic cancellation. I do not know how the 7.5 degrees is implemented at the HV transformer. Our company is asking the manufacturer for information (which I probably will not get in the field). I beleive this is suppose to give is harmonic cancellation to the 21st harmonic (I'd have to read up on it again, I'm not the designer).
We have established that the imbalance does not occur until we reach approx 500amps of load (speed independent), when it hits, it comes on fast and doesn't leave us much time until we blow fuses. If we never reach this load, we never get an imbalance, but whenever we reach it, the imbalance stays even after we reduce the load.
Today we noticed that the voltage for the generator (3Mwatt Marine CAT)droops when we bring the VFD system online.

 

[rhatcher]

I have to wonder...is the problem on the front end at the paralleled transformers, in the middle at the rectifiers and DC bus, or in the back end at the paralleled inverters?

Note that I am not a drive expert and my actual experience with 12 pulse systems is for 12 pulse DC drives having an active front end (SCR rectifiers) instead of the passive diode bridge that I think you are referring to.

I suggest that the problem may not be the transformers because theoretically you cannot circulate AC current between the two (imbalanced ??) supply tranformers. This is because the current would have to pass through two opposing (back to back) bridge diode rectifiers.

A problem in the wye portion of the VFD (rectifier or inverter) seems more likely.

I will finish by saying that the transformer and line reactors should have impedance values (at the power frequency of 50/60hz) listed on the nameplates. If you provide this information then perhaps someone in the forum can specifically determine if impedance match is a problem (in case my suggestion is wrong).

 

[sibeen]

wzrdstrm, moving the HV side 7.5 degrees doesn't make this a 24 pulse system. It will remain a 12 pulse at this VSD. The VSD can't possibly know, or care, that you've moved the upstream HV.

What may be happening on your site is that there are multiple systems and the designers have decided to move these in relation to each other.

The 7.5 degrees also doesn't make sense for a 24 pulse system. If you had two VSDs on your site, then phase shifting one by 15 degrees will give an effective 24 pulse system upstream. In this way with two VSDs with 12 pulse bridges the phase displacements are 0⁰, 15⁰, 30⁰ and 45⁰ on the four rectifier bridges. A nicely balanced system.

With a 7.5 degrees shift and just two VSDs he phase displacements are 0⁰, 7.5⁰, 30⁰ and 37.5⁰, then a big gap until the sequence repeats. Quite asymmetrical.


7.5 degrees shift are used when even higher pulse systems are required. I worked on a system, back in the early 90s, where there were numerous large 12 pulse (dual bridge) UPS systems on the one site causing harmonic problems. A series of phase shifting auto-transformers were placed in front of the UPSs. These were 7.5⁰, -7.5⁰, 15⁰, -15⁰ phase shifting transformers.

We had phase displacements of 0⁰, 7.5⁰, 15⁰, 22.5⁰, 30⁰, 37.5⁰, 45⁰ and 52.5⁰. A nice symmetrical 48 pulse system seen upstream.

 

[rhatcher]

sibeen has made a good point. This may be a faulty design problem.

If wzrdstrm can provide more specifics about the situation then perhaps he can get better help.

 

[wzrdstrm]

For the record so far. It was determined that if we only ran 1 generator (remember this is an offshore rig), there was a KVAR spike on the generator the same time we went into imbalance, and KW dropped at the same time. It was also determined that if we had two generators online, an imbalance situation did not occur. This was also recently tried swtiching bewteen one and two generators to verify that this was the cause, which it was. This may not be the whole solution yet, as some of the information brought up here and some lingering problems are yet to be solved.

 

[DickDV]

Sounds to me like the drive front end should be returned to a conventional 12 pulse configuration with one bridge fed from the wye winding and the other from the delta winding.

What goes on upstream from the transformer is a separate issue and I don't see where it would affect the drive one way or the other. As long as three phase AC power with the phases 120degrees apart is provided on the transformer primary, you should be ok.

 

[ozmosis]

I think Inter Bridge Reactors (IBR) are probably needed for this drive system as you will be seeing a different instantaneous voltage from each rectifier. The IBR usually works to balance the currents.
It does depend on the type of VFD though. Some already have DC reactors built in but others don't and some have such a compact rectifier design (I recall Yaskawa had one) that installing IBR is practically impossible and if not installed by the factory, proves difficult to do yourself.

 

[sibeen]

So the problem occurs when supplied by one generator and not when supplied by two.

I'd suspect a harmonic or notching issue. The distorted waveform throws out the switching of the input rectifier. OK, I'm assuming that it is a controlled thyristor rectifier.

Check your voltage waveform for excessive notching and/or high harmonic content.

 

[Drivesrock]

It's a small world. Little did I know when I replied on 9th March that I would actually get to this site!

I think some problems with the rectifier pre-charge circuits are now sorted.

But a voltage imbalance occurs with the 640V supply

Some data:

Passive 12 pulse rectifiers on the drives
Generator: 11kV, 60Hz, 5500kVA - each Wye with neutral grounded via resistor
Transformer: 8800kVA, 11% impedance (according to nameplate), 11kV extended-delta primary, hence a -7.5 degrees shift (the other transformers are +7.5); 642V nominal secondaries Delta and Wye ungrounded. Under low load Wye = 636 and Delta 634V so normally only 2V difference so well within tolerance.
At a load of about 350kW on each secondary and one generator on the 11kV bus after a few moments the delta secondary voltages drop - at the transformer terminals - to 620V. The transformer growls and a spike in kVArs are seen on the generator. Sometimes 1000kVAr spike! If at this point a second generator is connected to the 11kV bus the voltages across the 2 secondaries normalize.

So it's a system issue - any thoughts?
Why should the Delta secondary winding interact with one generator and it must be circulating a lot of current internally for the terminal voltage to drop?

Of course with the voltage imbalance comes a great current imbalance at the 12 pulse rectifiers with the Wye higher voltage bus supplying nearly all the current which was blowing I/P fuses as well as a pre-charge problem.

 

[Marke]

I would suggest that you look at the voltage across the line reactors in series with each rectifier.
If the voltage across the two reactors is no equal at low - medium current, that would suggest to me that you do not have sufficient inductance in each reactor or they are not matched.
If the dominant reactance is the leakage reactance in the transformer, there will be an imbalance in the currents. If the dominant impedance is the line reactor, then the current should be equal provided that the voltages are equal.
If the current through the reactors is two great, they can tend to saturate at which time, the series reactance will drop and the current will rise. I suspect that when you have the problem, the voltage across the reactor with the high current will be less than the voltage across the ow current reactor.
The reactors may need to have an air gap, or higher rated current.

Best regards,

Mark Empson
Advanced Motor Control Ltd