Damaged Transformers: What's the Real Story?

I don't know your location and it is impossible to diagnose the problem from a distance. Among the possibilities you mentioned, transformer quality and design may certainly be a contributor. Other possibilities include thermal overloading of the transformer which causes accelerated aging and deterioration of the internal insulation. Electrical disharges in the oil will cause outgassing, generate explosive gases, and increase internal pressure. Improper degassification of the oil. Improper coordination of the surge arresters and transformers. Review the entire insulation coordination curve of the transformers and compare that to the arrester characteristics curve from steep wave through 50 or 60 Hz to determine the overvoltage protective margins. Don't forget to add the line and ground lead voltage drops to the arrester discharge voltage. Those leads are in series with the arrester and must be included. Their voltage drop V = L di/dt. The inductance (L) of a straight lead wire is about 0.4 microhenries per foot. If the lead wires are coiled, their inductance can be much higher. That's the reason for keeping the arrester line and ground leads as straight and as short as possible.

You should also check the system and arrester grounding and make sure they are solidly grounded through low footing resistance.

Transformer failures due to overvoltage surges would also include switching surges and system faults.

It's quite likely that you will find your failures are due to a combination of factors rather than any single cause or event.

I would think the transformer manufacturer would be involved in the analysis of the failures. The number of failures over a relatively short time should certainly cause some concern on his part.

The location is in Democratic Republic of Congo. The Transformer was made year 2006 and 2007 in South Africa. The physical condition of this transformer are not very good compare to others. The designe are very poor, it is an Oil emmerse and totally sealed welded casing w/o any pressure relief valve, pressure, oil level, and temperature gauges.

For overloading reason definitely not, ground resistance very good, cable connection to arrester is very short and straight.

As you mention maybe the causes are for the quality, I will consult the manufacturer of this transformer soon. Or it might be the electrical discharge into oil that cause outgassing or generate pressure.

Thanks for your input, I will get you back once i get information from the manufacturer.

I would suggest sampling the oil in the unfailed transformers of the same vintage and design. Oil dielectric strength and dissolved gas analysis looking particularly for flammable gases, moisture, and other evidence of partial discharge and insulation breakdown.

Are the arrestors tested regularly? have they failed?

Beside lightning arrestors, have you considered surge suppressors?

There is also the possibility that solar flares and geomagnetic storms are causing the problems.

You might want to read this: http://www.hsb.com/thelocomotive/Story/FullStory/ST-FS-SOLAR.html

We did not test all our Arressters because we don't have test instrument onsite. Also we don't use surge suppressor.

The problem is only in our 33kV distribution line, but for 11kV line never happen this problem.

All the past topics are good topics, including oil test. You'll be able to figure out if there is somethings wrong with your transfo. Perhaps, you should look first if you aren't overloaded on your line or your transfo. I guess they aren't having all the same impedance. Take the measurement with standard equipment that you shall have like PT and CT and check this out if your having the same flow of power throught your transfo. Because there is no problem in both lines only one.

You could also take a thermographic image of all of your transfo it should tell you some answers.

Bonne chance, Luc

1. Pls. check insulation class and BIL rating of the equipment and perform insulation test on the rest of the units. Try to find out similarities in all failures.
2. Hire a consultant and vet the design criteria again.