Noisy Tranformer - Bad core laminations? 4

[Chapmeister]

Hey all,
I have a 300kVA Marcus brand transformer with an excessively high sound output. The sound output has been metered at ~85dB, which is well over the rated 65dB. The transformer was only purchased and install a little over 1 year ago. It was physically inspected and is clean as a whistle. Runs very cool.

Here are the tranformer specs:
Brand Marcus
Model MK300A7-13-F
Power Rating 300kVA
High Voltage 600?347V
Low Voltage 416Y240V
Impedance 3.3%
Temp. Rise 115ºC
K-Factor K13

We had some test equipment install for approx 24 hours, which showed low power factor ~0.5, and elevated current harmonics %THD consistantly over 25% (between 27.5% and 30%). What should I make of these results?? I was told it could be bad core lamination. Any advice/comments are appreciated!

 

[Skogsgurra]

It sounds like you have an elevated primary voltage. Not much that can be done (add more turns to primary winding or lower primary voltage is all that I can think of). I do not think that the laminations are loose.

K13 is not a very high distortion factor. Do you have much non-linear loads? (Like VFDs and SCRs).

Gunnar Englund

http://www.gke.org

 

[Chapmeister]

The primary voltage readings stayed below 347V for the duration of the 24 hour test. At some points the voltage dipped to as low as 338V, but only for short time periods. The voltage %THD stayed between 1.0% and 1.7%. So I don't believe the voltage is the problem.

The loading on the transformer is part of a plastic extrusion line: an extruder motor (150kW DC), a melt pump (small), and heating elements.

 

[electricpete]

I assume those voltage THD numbers came from the line side.

Were the current harmonics on the load side or the line side? Did you check the opposite side? If current THD higher on the load side than I tend to think the load created the harmonics, if higher on the line side I would think lean more towards overexcitation.

Don’t rule out overexcitation based on a voltage measurement without looking at the tap settings. With the fairly low transformation ratio, is this possibly an autotrasnformer?


=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[Chapmeister]

The testing was performed on the primary side only.

It's not an autotransformer. It's a K-Factor 13 as described at

http://www.marcustransformer.com/vw/fs/p017.htm

The Marcus site claims its design is aimed at non-linear loads. And it's 'features' are listed as:

§ Specifically designed copper windings with multiple conductors reduce harmonic high frequency 'skin effect' and increase efficiency
§ Low flux density wound cores reduce saturation
§ Oversized neutral conductors eliminate overheating
§ Double grounded electrostatic shielding minimizes electrical noise

I just don't know what to think about all this.

 

[davidbeach]

If it has been noisy its entire life, the first thing I would check are the internal bolts. Usually the core is firmly bolted into place for shipment and then once installed the bolts are supposed to be backed off so that there is some give in the rubber mounts.

 

[Chapmeister]

Hmm that's an interesting thought. We had some professional maintenance guys do a visual inspection when the plant was shut down last month. They said everything was tight (probably referring to the connections), but as you say, perhaps something is too tight!?

I still think it must be something electrically related because it buzzes loudly when under load, but not when the machinery is off-line.

 

[Skogsgurra]

Your 150 kW DC motor is probably (I'm sure) thyristor controlled. SCRs produce lots of harmonics.

Is the sound level the same if the extruder is lightly or heavily loaded?

Is the sound a typical "electric hum" or is it a "yelling" sound?

Are there line reactors for the thyristors? Or does the transformer serve as a combined line reactor/transformer?

In the latter case, sound is inevitable.

Gunnar Englund

http://www.gke.org

 

[Chapmeister]

The sound level isn't really variable. It's either on or off: loud or quiet. If's definitely more than a hum, like a loud buzzing. We have a larger 750kVA in the same electrical room which makes half the noise but runs way hotter.

As for reactors or thyristors, my knowledge of the internals of the control cabinet is limited. I'll check in to it if I can.

 

[wwright]

Loud sound can be relative. If you can somehow check the frequency of the sound and distinguish it from 60 cycle hum, you may find that it is indeed a result of nonlinear loads. Sounding louder when the load is on is a sign it may be caused by the nonlinear loads.

 

[Chapmeister]

It was proposed by the company that did the testing that the source of the bad current harmonic distortion could be a result of bad iron core laminations. From the email I recieved:
-------
It is possible that these harmonics are at least partly responsible for the noise level in the transformer. However, I have spoken to several people about your problem and the general consenus is that the most likely culprit is the laminations in the transformer. Although there is likely nothing "wrong" with the transformer, our suspicion is that the iron in the laminations is simply not top quality - which is relatively indicative of a low cost Marcus transformer.
-------
It kind of makes sense that bad iron would result in greater hysteresis losses which in turn increases the THD.
Would anyone agree with this evaluation?

[I'll post the graphics they sent me so you can take a closer look]

 

[edison123]

Core lamination noise (called magnetostriction) is not load dependent since the flux remains almost constant from no-load to full load.

Did you check whether the frequency is as per name-plate (50 or 60 Hz) ?

* Algebra - The weapon of math destruction *

 

[wwright]

The most common thought is that the harmonics results from the load connected to the transformer. However, I can agree that the iron lamination condition can contribute to excessive noise.

 

[Chapmeister]

http://myweb.dal.ca/~mr663256/IPG_Xfmr/Xfmr_Graphics.html

Here is a link to the data graphics. I welcome any and all interpretations.

 

[edison123]

of course, my comment is valid only for linear loads. Non-linear loads may id=nduce higher harmonic noise from the laminations.



* Algebra - The weapon of math destruction *

 

[jghrist]

The current waveform (3rd graph from end - I assume that the graph Y axis is labeled incorrectly) is typical of 3 phase bridge rectifier current upstream of a delta-wye transformer. This is apparently the source of the current harmonics. The source must be pretty stiff or you would have voltage notching at the current discontinuities.

 

[Chapmeister]

The transformer is fed from a large piece of switchgear (in the same room). The fuses are Ferraz Shawmut A6D300R. And the switchgear is fed from a 5MVA (25kV/600V) transformer. I'm not sure what you mean by an upstream rectifier? There's nothing else inline with the tranformer.

 

[jghrist]

What I meant was that the current was measured on the source side of a delta-wye transformer. By stiff source, I meant that the source impedance is low. If it weren't, then the step increases in current (high di/dt) would result in sharp voltage increases or decreases (notching) at those points in the waveform. V=L·di/dt

Considering that the 300 kVA transformer has only about 70 kVA of load and that it has a K-factor of 13, there does not appear to be a loading problem with the transformer. There is no voltage harmonic problem. The only problem appears to be the noise.

 

[electricpete]

I agree with jghrist's observations. What you labeled as V is actually current drawn by a three phase full wave rectifier.

In most cases I believe transformer noise is electromagnetic in origin and unrelated to any looseness. Certainly it applies to motors (like fractional slot motors which are rich in harmonics and very loud).

It was suggested this is not electromagnetic since the flux does not change with load much. That may be true for sinusoidal current but I don't think it's true with so much harmonic content.

My vote (guess) is that it is most likely electromagnetic noise related to the load harmonic current. There isn't much you can do about that... unless maybe someone sells a filter for this situation?

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[Skogsgurra]

I think that the labelling is correct. There are two sets of vertical scales,volts to the left and amperes to the right.

The current waveform is not typical for a bridge rectifier input current, there are usually two equally high current pulses. But given that there are also linear loads and that the transformer winding also influences the total curve form - current pulses added - I think that this is a very normal look for a transformer primary when a thyristor controlled DC motor is connected.

What I do not like at all is the phase relationship between the currents. There must bee a mistake when connecting the voltage CTs. Phase A current should be 180 degrees shifted.

The interesting thing is that the voltage waveforms are also shifted. Again, the phase A is 180 degrees wrong. I really hope that it is a measurement connection error and not a condition that the transformer is working under. That would certainly produce lots of interesting effects.

It is interesting that PF seems to be about right (assuming that the extruder was run at low speed). So the connection U vs I is probably right.

Anyhow. With that primary current, there is bound to be a lot of noise from the transformer. The thyristor commutation seems to take place in about 3 degrees electric so it looks like the transformer is also doing duty as a commutation reactor. That is probably why you get so high a noise level.

Gunnar Englund

http://www.gke.org