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!

 

[electricpete]

"I think that the labelling is correct. There are two sets of vertical scales,volts to the left and amperes to the right."
But that particular graph only shows current, not voltage.


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

 

[electricpete]

OK, I see the colors are supposed to tell us that we're looking at current. And the graph a few above has the voltages.

I also see what you are saying the current peaks do not appear in the same order as the voltage peaks which seems like just a measurement polarity.

You are right - this looks like a transformer input waveforms resulting from add together two phase currents of the FWR input such as would appear on the input of a delta-wye transformer feeding a FWR. See last waveform on page 2 here

http://home.houston.rr.com/electricpete/FWR.pdf

Why do you think it has to be a thyristor-commutated rather than a line commutated (diode bridge) rectifier?

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

 

[Chapmeister]

Thanks for your input everyone!

I'll try to collect a few more load specifications which might help us find a solution.

The puzzling thing is that we have 8 other similar extruders, a few with dedicated 175kVA tranformers, and they don't seem to make nearly as much noise. Of course, those ones aren't likely K13 rated. Is it possible the higher K-rating is trying to 'fight' (damp) the harmonics generated by the DC motor and is therefore working harder?

Addressing the graphical issues, I don't see the problem. It goes Ch.2-Ch.1-Ch.3 (Yellow-Red-Blue) in both cases; voltage and current. Do you think Ch.1 and Ch.2 should have been reversed? Does it really make a difference?

 

[electricpete]

You're right: 2, 1, 3 for both current and voltage.

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

 

[ScottyUK]

Re. voltages and currents:

The vector sum of the three phases should be zero. The red phase signal is definitely reversed on both the voltage and current traces. It is inverted in polarity on the image. The other two phases look normal. If the transformer really is wired up like this, and it isn't a measuring instrument error, then expect wierd things to happen. You can verify this by looking at the time axis on the images: there should be 5.56ms between each successive crest for a 60Hz system, where your sequence shows periods of 2.78 ms / 2.78 ms / 11.11 ms between crests.

Transformer must be a delta-star (or star-delta) type as noted above to produce the current waveform displayed. Not much inductance in the load either.


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

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

 

[Chapmeister]

I will contact the person that performed the testing and point out the waveform issues mentioned. I agree that the vector sum of the waves should be zero, and clearly this is not the case.

 

[Skogsgurra]

Sorry to be late with these comments to the posts above. Tried earlier, but the "<." stopped me.


First, if it is a DC motor for an extruder, it has to be a speed controlled one. And that means thyristor control. The commutation is when phase B takes over from phase A etc. That means a momentary short between the phases and that's what normally produces the sound, either in the line reactors or in the transformer - as seems to be the case here.

Second, the angle between the phases shall be 120 degrees. You have 60, 60, 240 degrees in the recordings. That's why I think there is a polarity error in phase A (the one that sits between B and C).

Gunnar Englund

http://www.gke.org

 

[Chapmeister]

I just collected the info off the DC motor:

Brand: Thrige-Scott Ltd.
Type: Shunt Motor
Power: 150kW
RPM: 2000
Armiture: 460V - 350A
Exciter: 360V - 6.1/4.85A

Again, the control cabinet also powers some small melt-pumps and heats.

I checked the power factor stated by the switchgear and it said 0.77 (with the extruder off-line). How is this related to the ~0.5pf displayed in the graphs? I didn't think the pf on the switchgear bus and the transformer primary would be different considering they're directly connected.

 

[electricpete]

The phase error would be in the voltages (you said current before). I agree with you know one phase voltage is flip-flopped ( I guess I've flip-floppwed a few times on that subject... better keep my mouth shut).

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

 

[jghrist]

What do you mean by "power factor stated by the switchgear"? Is this power factor measured on the 600V side or the 416V side? If it is with the extruder off-line and the pf with the extruder on-line is 0.5, then the extruder has a lower power factor than the rest of the load. How much depends on how much of the total load is made up by the extruder.

 

[Chapmeister]

We have a piece of metering equipment installed in the switchgear which shows the instantaneous kVA, kVAr, kW, pf, etc. of the SG bus. It claims a pf of 0.77, which is okay. Why would the pf on the transformer primary be so bad (0.5pf)? I'm starting to get the feeling that the test equipment must have been set up incorrectly.

 

[stevenal]

"I didn't think the pf on the switchgear bus and the transformer primary would be different considering they're directly connected."

I'm not surprised by the difference even if the two metering points are being monitored simultaneously. Your PQ metering is looking at true rms values that include higher order harmonics. Your switchgear metering is probably only looking at fundamental quantities.

 

[ScottyUK]

If your instrument had the red phase inputs reversed, you could not really expect to get any meaningful pf information from it. Was it a different instrument to the one which captured the waveforms?


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

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

 

[Chapmeister]

Well the fact that the waveforms are not as expected, and the powerfactor is significantly lower than I think they should be, I'd say the leads were hooked up wrong.

The thing is, I remember when he hooked them up, he said "the tester won't allow them to be backwards, it'll beep or something". Haha, as a skeptic I believe everything can be hooked up wrong!

I've already emailed the guy, so I'm hoping he'll get back to me soon, maybe tomorrow.

 

[jghrist]

Calculating the power factor from the measured kW and kVAR gives around 0.47 pf. At about noon, total kW is around 90, total kvar is around 167, giving total kVA of around 190 and pf = 90/190 = 0.47. This would include the distortion factor.

Calculating the power factor from the phase displacement also gives around 0.42 pf. Yellow phase voltage peaks at about 5.6 ms, yellow phase current peaks at about 8.6 ms, 3 ms lag or 65°, cos(65°) = 0.42.

These two measurements are within the margin of error of picking up points from the graphs. It wouldn't appear that the distortion contributes significantly to the low power factor. This is to be expected with very low voltage distortion.

The transformer var loss would not account for the difference. It is only 60% loaded and the impedance is only 3.3%. I still think the difference can be explained by the fact that the switchgear bus measurement was with the extruder off-line.

 

[stevenal]

I agree the red phase leads, both V and I are reversed. As Skogurra noted, this placed the red phase pf right in line with the other two phases. Watts, VARS, rms values, THD, pf; all measures the same. Just the two waveforms are inverted.

 

[Chapmeister]

OKAY, here's what the "technical representative" said in his reply:

"Yes, it looks like the PT may have been reversed. Hard to say for sure, but it is either that or you have a more serious power problem that we first thought! <G>"

I'm not sure what the "<G>" means.. but it looks like I might not get much help out of this guy.

So you think besides the waveforms that the other data is correct? I find this a bit hard to believe.

 

[Skogsgurra]

Hello Chap,

The <G> probably means "Grin" - a way of saying "Sorry, but what can I do about it?"

I have studied your waveforms in detail. Except the phase error in A voltage and current, I think that the data are correct. The low PF is typical for a thyristor controlled DC motor running at low speed. In fact, there is an almost linear connection between percent rated speed and power factor.

As I said earlier, the commutation between phases seem to take about 3 electrical degrees. That is a rather typical value. Perhaps on the low side. And that indicates that the transformer with its low impedance is the only impedance in the thyristor/motor path. That makes the transformer noisy - nothing to do about it. It won't fall apart.



Gunnar Englund

http://www.gke.org

 

[dewone]

Well alot good info has been discussed and without an electrical solution I suggest. Placing rubber under the transformer mounts to reduce the noise transfer to the floor that will help reduce the total amount of noise in the room.

 

[Chapmeister]

That's a good idea. I'll discuss it with my superior. I'm going to recommend more testing (ensuring the leads are attached properly), and maybe an analysis of the motor control cabinet.

Thanks again to everyone who offered info and advice.

If I see any further conclusion to these issues I'll be sure to post back with the results. THANKS!