Transformer abnormal heating

Status
Not open for further replies.
patpappas

patpappas

Member
150kva 208 3PH dry transformer just had an infrared done on it, the temp rise was 10.3 degrees, the left and center coils show the temp. rise, aside from checking for a balanced load what other possibilities could there be with this transformer. The infrared showed no problem at the primary or secondary connections. Could there be an over abundance of harmonics on just these two coils.
 
080903-1621 EST

What kind of degrees? Even 10 deg C. (18 F) would not be an excessive rise. The ratio of F to C is 9/5. Also are you saying that one coil is cold.

Conversion equations are:

C = (F-32)*5/9
F = 32 + (C*9/5)

Test these at 32 and 112, or 0 and 100.

If you had 105 deg C insulation (221 F) and room was 120 deg F, then maximum hot spot should be less than 101 deg F rise.

.

.
 
patpappas said:
150kva 208 3PH dry transformer just had an infrared done on it, the temp rise was 10.3 degrees, the left and center coils show the temp. rise, aside from checking for a balanced load what other possibilities could there be with this transformer. The infrared showed no problem at the primary or secondary connections. Could there be an over abundance of harmonics on just these two coils.
Click to expand...

I have to assume you mean a 10C rise over the right coil. Was the scanning done without the covers on? Can you post a picture? You need to check loading on all 3 phases.
 
Can not post a picture, infrared results done by others, 10.3 rise is above the
right coil, the front cover removed, back cover still on. Planning on a load balance check today, if this is balanced I was wondering what else may cause this. Hoping the XFMR not breaking down.
 
10C is Small

10C is Small

A 10C differential is not much. It could be caused by unbalanced load. It could be cause by uneven air flow providing uneven coil cooling. Generally coils are not heated by harmonics, iron cores are, so harmonics is unlikely the issue. Are all the vents clear? Is the internal cabling blocking air flow?
 
Doing a spot check on the load, all legs were within 7 amps of each other, I was wondering about harmonics, air flow is good around the transformer, cabling is a bit messy on the bottom but no worse than the other seven just like it. The infrared Co. had this XFMR listed as "investigate" with a "minor problem" label in their classification category.
 
Other causes & Testing

Other causes & Testing

Is there a chance that the hot coils were getting a reflection from another heat source? IR light can bounce off an object. I have seen cases where the IR measurement was really reflected from another heat source.

If not, TTR and Megger of the transformer is in order. Shorted turns, ground fault, multiple core grounds, etc. need to be checked. With balanced load and voltage, the windings should be the same temperature.
 
beanland said:
Is there a chance that the hot coils were getting a reflection from another heat source? IR light can bounce off an object. I have seen cases where the IR measurement was really reflected from another heat source.
Click to expand...


A good IR tech should know how to avoid this, BUT as more companies hit the IR market and they fight to drive the prices down, you often get what you get.
 
a certified infrared thermographer can tell the difference in reflection if he knows anything. harmonics would show in the current reading of the neutral in relation to the highest unbalanced current between phases. this isn't normally a problem in overall heat produced by a transformer. we have seen where harmonics have caused the neutral bus to melt! if this transformer is loaded to 60 percent or higher constant loading---it's gonna be hot! i'm talking fry an egg hot! but this normal considering a 150kva especially in an enclosed room. i can think back at buildings with large transformers running at near full load---wondering when they were gonna burn up-?? never happened.................
 
A lot of good information here, thank you, we are currently recording data on the transformer,(Rustrak data logger) again thanks.
 
A transformer?s rated temperature rise (degrees Celsius) is the average temperature of the transformer?s windings over an ambient temperature of 40?C. In other words, the average winding temperature = ambient temperature rise.

Now, you have specified a 10 Deg temperature rise. You have not indicated if this rise was in excess of the rating of your transformer.

If it is in excess expect a shorter life for your most likely overloaded transformer. Consider a fan kit if your transformer supports one.

If it is not in excess, the slight difference in temperature is caused by load balance.
1. If you have a 3 Ph 208 transformer, if it's feeding a panel for distribution I'm guessing one phase is slightly underused, and this is normal. The installing EC most likely tried to balance the panel, but could not predict load demand at different times.
2. If the transformer is feeding a single 3 Ph distribution point such as a large drive proportional to its motor load. You may experience load unbalance sometimes up to 50% at the transformer
 
charlietuna said:
a certified infrared thermographer can tell the difference in reflection if he knows anything. harmonics would show in the current reading of the neutral in relation to the highest unbalanced current between phases.
Click to expand...
This normally a problem related to single phase non-linear loads. Third harmonic and multiples of it are the culprit.
Three-phase circuits don't generally produce triple-n harmonics.
Maybe patpappas could give some information on the loads being supplied.
 
The loads are gaming machines, the transformer was traced by a data loger, all three legs for five days averaged with in 6.2 amps of each other. the voltage was constant on all three also. The report simply specifies "Temp Rise 10.3 C" on center and left phase coils. The transformer is a 208V 3PH.
 
Gaming machines? As in arcade video games or casino machines?

As for arcade games, I can state that most use computer switching supplies of variable quality, so harmonics can be a definate issue.

I would imagine that casino style gaming would be the same.
 
Status
Not open for further replies.
Top