BIL testing on already failed transformer

[Swgr-stuff]

Hi all,
I’m newly registered to this forum.
i have been consulting it from times to times in the past few years and i always found very interesting information.

The reason of my recent registration is that my company is facing a huge problem.
Let’s start by saying that we are a switchgear manufacturer for underground mining operations, and we sell a respectable amount of dry type transformer with our equipment.
Generally speaking, we sell about 75 dry type transformers per year, totalling about 700 of them in the last 20 years or so.

95% of these are typical mine duty 1000kva 4160V-600/347V 150C rise 30/10kV BIL Copper.
They are protected by fuses and switched by a typical air break load break switch (no vacuum).
Out of the 700 units, we have bought about 130 units of manufacturer "A", 100 units of manufacturer “B”, 400 units of manufacturer “C”, and about 70 units of manufacturer “D”.

We never had any problems and major failures with manufacturer B, C and D (except some clients who broke adjustment taps and stuff like this). All the transformers bought from manufacturer “A” have been bought between 2013 and early 2016.

Since then, we had 5 major failure (on units from manufacturer “A”) all at different mine site, with different site condition (some grid connected, some with genset, etc). All failed transformer are 1000kva, as I described above. 4 out of 5 failures occurred between 192 and 399 days after delivery to the client site. The 5th failure occurred 1070days after delivery. Today we learn that another transformer failed at a different client/site, but we do not have more details for now on this unit, except that it is also a 1000kva, from mfg “A”, again.

It is also very important to take into account that we have a lot of identical equipment at these same exact sites, with different transformer brand. And it is only the “A” that seems very problematic. No BIL test has been done on these (type test on similar unit, a while ago, we have no access to this report).

The failures always have the same signature (failure on HV winding, one phase only, exactly at the location where the voltages between 2 turns are at the maximum values).

Since the beginning of failure, we have sent transformer to analysis and for warranty repairs. But since the first event, they are not cooperating at all. They do not provide appropriate report, and keep blaming to site overvoltage. But we have very strong doubts since we have a lot of units from different manufacturer that are at these same exact mine sites and under the same conditions. The failure statistics speaks for themselves.

So, I am sure you understand that we are getting quite nervous about this since we have a lot of long terms clients that have these “A” installed. It goes without saying that we stopped buying from this manufacturer, as they are not willing to cooperate in deeper investigation. We think that there might be a flaw in the manufacturing process that is ending up with a winding not being able to withstand the 30kV BIL it is supposed to. 3 of the failed units have been manufactured in a 3 weeks production window.

We managed to take back one new unit of “A” from a client; it was never powered and has been in storage for about 3 months. We were thinking of doing some BIL testing on it to confirm if it is actually passing the 30kV BIL.

We could also potentially manage to get back 3 of the damaged transformer. Now comes the tricky question, we were wondering if there is any way of doing a BIL test on the 2 undamaged coils of each transformer? Would the test result be reliable? What would be the right method for this testing?
Also, can a BIL test be done to a transformer that is currently in service for about 1 or 2 year? Would the result be reliable?

I would be very grateful if someone can help us as this is truly starting to feel like a real nightmare.

Thank you!

 

[davidbeach]

Define "failure".

If it's a turn-turn fault it probably has little to do with BIL, the interturn insulation doesn't have nearly the withstand capability of the insulation to ground (the core).

 

[Swgr-stuff]

Davidbeach,
we do not even have this information (just to tell you the level of support we are getting).

one thing i forgot to mention, the 4.16kV network is resistance grounded (throu the NGR connected to the source transformer), limited to 15A or 25A (depending on mine site). So a phase to ground fault would always be limited.
600/347V secondary is also resistance grounded with a 5A NGR.

 

[davidbeach]

Turn-turn faults draw very little current from the terminals, but the current in the faulted turn will be very high.

 

[Swgr-stuff]

thanks for the info
all transformer were protected with 160A fuses (2 of them blown in each case), so i am guessing current was high, as these fuses have a high minimum breaking current.

 

[Swgr-stuff]

based on the little information we have, is there some test we could do the the new unit we have in hand? as well as one of the damaged units?

 

[cuky2000]

Before attend to answer your questions, here is some useful background information: All 600 volt and below transformers are related 10 KV BIL. Transformer is subject to overvoltage that creates dielectric stresses that induce the failure on the unit. Lightning surge are could strikes near the transformer or even at some distance from the unit. Voltage wave surges can travel down the line and into the transformer. High voltage switches and circuit breakers can also create similar voltage surges when they are opened and closed. Both types of surges have steep wave fronts and can be very damaging. To minimize the effects of these surges, the electrical system is protected by lighting arresters, surge capacitors and snubbers can be installed.

1. There is any way of doing a BIL test on the 2 undamaged coils of each transformer? Beware that the BIL test will impose dielectric stresses that might accelerate dielectric failure
2. Would the test result be reliable? This test might not be representative of the new unit because of the unknown operating conditions and stresses induced by the failed winding.
3. What would be the right method for this testing? Not sure
4. Can a BIL test be done on an existing a transformer that is currently in service:
o Operating time of 1 or 2 year: Yes, but impractical since the BIL test could degradation or induce failure on the existing unit
o Would the result be reliable? If the stress history is known (switching, overvoltage, etc.) some rough estimate could be projected. Accuracy still will be questionable since dielectric strength is not a deterministic phenomenon.

GENERAL COMMENTS:
a. Grounding resistor will impact the overvoltage imposed on the unit in a frequent line-to-ground fault or ferroresonance issues (See coefficient of grounding concept / overvoltage issues).
b. A combination of secondary surge suppression, snubber or surge capacitor could helpful to mitigate the overvoltage and reduce the risk of failure on the dry type unit.

 

[prc]

When you say BIL test, I take it as impulse test as per relevant standard. You are assuming that the cause of failure is lack of insulation strength to withstand surges. But chances for that seems remote as you are not using any VCB for switching.

If the failure is at some where near line end or the floating end of tap sections( you can easily see it on a visual inspection of the failed winding) BIL test may give some clue. Also check whether there is any failure to ground.

If you want to do BIL test it can be done on the failed unit with one limb in damaged condition. When you are repeating BIL test on a used transformer, the peak value should be reduced to 80 % of original test voltage. But at least on one limb do it at 100 % voltage to locate weak links if any.

 

[Swgr-stuff]

cuky2000: thanks for your comments. in short, what we need to do is to prove in come way that there are something wrong with these units (from mfg "A"). i am fully aware that there are some system criteria (transient, VCB, strike, etc) that can cause this type of failure. But, as i previously said, these failure only happens with this manufacturer. we never had any failure with the 3 other transformer manufacturer, and that is over 20 years (mfg "A" are about 130 units on a total of 700 units). This is why we need to find a way to prove there is a problem with the transformer itself.

prc: thanks for your comment. yes BIL as per relevant standard. The main reason we want to do some testing is that we have strong doubts that the transformer is not built to withstand the BIL it is supposed to.
The failure are always on the line end. We did not tried to do a hipot test as the HV coils does not even pass a megger test (5kV tester, only goes up to 1500V and 1.1Megohm reading).

 

[Swgr-stuff]

i managed to get a confirmation that all failure are turn-turn fault. in all cases, primary fusing blew (50% of the time 1 fuse, and 50% it is 2 fuses)

 

[VTer]

Was there any acceptance testing performed when these units were installed for the first time? Did you have certified test reports from the manufacturer? You can also request to witness testing of the unit from the manufacturer. For the spare unit, try doing some standard acceptance/maintenance testing first: Measure resistance, perform insulation resistance, PF tests, turns ratio... Also, if you have capability try applying rated voltage and perform an IR scan. By the way, are these cast coil dry type transformers?

"Throughout space there is energy. Is this energy static or kinetic! If static our hopes are in vain; if kinetic ù and this we know it is, for certain ù then it is a mere question of time when men will succeed in attaching their machinery to the very wheelwork of nature". û Nikola Tesla

 

[stevenal]

According to IEEE C57.12.00, The basic lightning impulse insulation level (BIL) test is not a routine test for class I transformers. It is in the "other" column of Table 18, so if it is desired it must be specified. Might want to specify front of wave and switching impulse testing as well.

 

[Swgr-stuff]

Vter: thanks for your comment! any help is very appreciated! no there was not any acceptance test. We have basic production test reports on some units.. i would say for maybe 10% of them. I know that we can request to witness type or production test, we did it for some very specific projects, but unfortunately not on these. For the spare unit, we will probably check with a test lab to do the evaluation and testing (as it will be required to be done by a third party, in case it have to go to court.. in fact, good chance we will). They are all dry type transformer, layer wound VPI.

stevenal: i am not sure i fully understand. the transformer i am mentionning are not class I transformer for electronics, they are 1MVA medium voltage power transformer. i agree that it is not a routine test, we should have asked for if first, but since we have a lot of transformer with the same and basic configuration, we truly expected no issues since it is for the same application as we were doing in the lase 20 years. it goes without saying that after these experiences, we now only work with well renowed manufacturer. About front of wave and switching impulse test, could they point out a weakness in the transformer that could lead to the type of repetitive failure we get with this manufacturer? As said before, we never had any issues with identical transformer from other manufacturer, installed at the same sites.

 

[stevenal]

I was referring to class of power transformer, but I messed up referenced the standard for liquid filled units. Try C57.12.01 instead. Per Table 17 greater than or equal to 501 kVA, applied and induced voltage test are routine, but impulse testing is not.