Discharge following DC winding resistance testing on step up transformer?

[low1]

Good morning everyone.

I'm looking for help in understanding a recent incident that happened locally. During regular maintenance, a couple of technicians were performing testing on a 13.8kV/230kV D-Y generator step up transformer. One of their tests was winding resistance measurements using a Multi-Amp 830280 transformer ohmmeter. During their testing, they were running into issues with getting the readings to stabilize on the high side windings, so it was decided that they would go against the instructions in the manual for test lead set-up. The manual states that only one end of the winding under test is to be grounded, and all others are to be left ungrounded. Our standard practice for the high side is to leave H0 grounded, and to short and ground the low side windings with copper braid. I know from experience that when moving from one winding to the next (ie H2 to H3), it can sometimes take between 5 to 10 minutes for the resistance to "break" and start to fall to expected levels. The technicians performing the test were not seeing any drop after 2 minutes, so the decision was made to short and ground all windings that were not under test. I believe they were working on H2, so H1 and H3 were shorted to H0 (which was left grounded), X1-3 remained shorted and grounded. They performed the test again and were able to get the results they were expecting.

The test set was discharged and Technician 1 removed the current and potential leads from H2. Technician 2 was to remove the braid from H3 to move it to H2. As the 230kV bushings are quite tall, a fiberglass ladder modified to be rested against the bushings was used to access the top. Technician 2 was halfway down the ladder with the braid in one hand when a flash and discharge occurred. Technician 1 heard the "crack", Technician 2 described the flash as having someone flash a camera in his face. Technician 2 felt a slight shock through his arm.

No one is quite sure exactly why this happened. I'm curious if anyone would have any input into this occurrence. Some of the details are as follows: Test set was being operated at 5A (which is the correct, and maximum, output of this particular ohmmeter). Weather conditions were damp/misty. Test leads had been disconnected at least 10 seconds prior to the discharge. Fiberglass ladder was described as "in good condition". There was no apparent second point of contact. This was an older test instrument that had been replaced with a Tettex winding resistance instrument that operates at 50A, however it had been sent out for repairs. The multi-amp transformer ohmmeter was last calibrated in 2003.

My instant thought was that the test set had been left on, and Technician 2 was in contact with the winding just as Technician 1 lifted the leads, causing a dv/dt voltage spike, but we were assured the test set was "discharged" and the leads were removed quite some time before the discharge occurred. I'm not sure if the test set was turned off or unplugged, which is our standard practice prior to moving leads. I'm trying to understand the impact of grounding multiple windings while performing the testing, contrary to the instruction manual.

Does anyone have any thoughts as to why this happened?

Thank you for your input.

 

[Skogsgurra]

So, this happened around ten seconds after the di/dt peak?
And the transformer is oil insulated, OK?

I had a similar thing with a 130 kV transformer way back. But on a breaker top where we had been filling new oil. The discharge was violent and I almost fell from the ladder. So, charges stay for a long time. That is well known.

Your case seems to be a bit different. But the di/dt had induced a high secondary voltage and such transients can be quite high in amplitude. Possibly causing a flash-over somewhere. The "backswing" may not have caused a second flash-over and that left the winding with a residual charge.

But, you say that H0 was still connected to ground. If so, I can't see how the HV winding could could retain any charge.

So, no explanation so far. But you asked for thoughts, no explanation. And these are my thoughts.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[itsmoked]

I believe there have been several deaths among technicians testing spooled high voltage cable for insulation issues. The cable and shield represent a fairly big capacitance that charges up to the high voltage of the test and if not correctly discharged will sit there until a victim arrives.

With the transformer LV phases all tied to each other and grounded I would expect there to be some capacitance between the primary and the secondary that would be 'chargeable' and everything being HV insulated no reason for it to ever leak-down.

There is also a phenomenon where you can discharge a capacitor and given a few minutes it will appear to have recharged (recover), while not fully, still to a major percentage of the original charge level. This is probably why there are time periods involved in the test procedure and in regards to connection/disconnection procedures.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Compositepro]

I do not fully understand the sequence that was described but this may be a case of induced charge on the grounded windings. I expect that the low voltage winding is physically located between the high voltage winding and ground. If a metal object is in an electric field, the electrons in it will get pushed to one side, but there will not be any net charge on the object. If you ground the object while it is in the field then electrons will get pushed to ground. If you then disconnect the ground from the object and then remove the electric field the object will be charged.
With this sequence of events it is possible to put a high voltage charge on the low voltage winding. This could cause spark discharge of the low voltage winding, by insulation failure or by grounding the winding, and thus an induced voltage pulse in the high voltage winding.

 

[low1]

Thanks for the input so far.

A couple of points: I can understand a high voltage charge testing insulation with a high voltage megger. I've felt what a stator meggered at 5kV can feel like if you don't properly discharge it. The working voltage of the multiamp transformer ohmmeter is 30Vdc I believe, so I don't know how it could develop a high voltage charge, as long as the sequence of discharge and disconnection is proper. I could understand a high voltage spike if the circuit was broken while current was flowing, but we've been assured this wasn't the case. Assuming of course that the equipment was working properly. That said, I was personally using the same test set the week prior, and it worked fine for me.

I'm familiar with dielectric absorption, we test control circuit capacitors for it regularly while troubleshooting issues. I was/am trying to wrap my head around whether or not grounding all phases except the one being tested (while it's neutral end remains grounded) could result in some sort of "capacitor", if all connections inside the tank were ok. According to the rest of the test results, nothing out of the ordinary was found. All tested passed. I'm struggling to understand how a capacitor could be set up if both plates are grounded. I don't believe any of the grounds were removed/fell off while there was still a charge, at least that's what we've been told. I trust the information being given by Technician 2, he is a good friend and I'm not in a position of authority over him, and our conversations have been strictly trying to figure out what happened for our own personal interests. Of course our various management, engineering and safety groups are looking into this as well.

Thanks again for any insight.

edit: Re-reading my first post, I'd like to clarify one point. When the technicians were not seeing the resistance drop after 2 minutes, they stopped the test and discharged the test set prior to grounding H1 and H3. Grounds were not applied while the test was in progress.

 

[Skogsgurra]

If there was not any ground connection, even a weak leakage path, then it IS a mystery. Better forget about it. No possible explanation.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[low1]

So last night, I was speaking with Technician 2 again regarding this incident. He informed me that he was now questioning the integrity of the ground connection on the low side winding. X1,2,3 were positively bonded, but the connection to "ground" was simply a copper braid wrapped around a painted pipe. So it's entirely possible that the low side was bonded but not grounded, which makes the possibility of setting up a capacitive situation becomes more likely. Am I wrong to believe this?

So I suppose a follow up question is: IF the low side winding was shorted but not grounded, would there have been a different outcome with only a ground on H0, vs a ground on H0, H1, and H3, with 5Adc flowing from H2 to H0? My initial though is that there wouldn't be a second capacitive "plate" had the high side windings not been shorted together, or at least not as large of one. Am I correct in assuming this?

Again, thank you everyone for your input.

 

[waross]

So we have suspect grounds and tech's not following proper procedure ans remembering details after the event.
That makes it a guessing game.
One guess:
1. The low voltage winding was not effectively grounded.
2. The low voltage winding took on a small capacitive charge from the 30 Volts applied to the primary. (The primary side ground may have also been suspect.)
3. The ground wire to the painted pipe was disturbed and made contact with ground.
4. The capacitive charge was abruptly discharged to ground. The di:dt induced a monentary voltage of several times the 230 Volt rating of the winding.
5. This in turn induced a potential in the high voltage winding of several times the rating of the 13.8 kV winding.
Although the potential was very high the current was very low, hence no need to call 911.
Maybe next time.
A refresher on safety ground integrity and proper procedures may be in order to avoid the call to 911.
Incredibly lucky.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[low1]

Thank you, waross.

A couple of points of clarification: This is a generator step up transformer. Primary side would technically be the low voltage winding, which is 13.8 kV. The secondary side would be the high side, 230 kV. Winding resistance measurements were being performed on the high side 230 kV winding. H0 was grounded through it's regular connection, approx. 3" wide copper busbar with bolted flange to bushing. The flange had 2 bolts through it, finger-tight. The ground wire on the low voltage winding was not disturbed, but it's integrity was questionable since it was on a painted pipe. The bonding connection from H3 to H0 was removed as the technicians were preparing to move the test leads from H2 to H3. The discharge occurred while Technician 2 was halfway down the fiberglass ladder resting against the bushing, with the copper braid in hand. There was no obvious second point of contact.

I definitely agree that they were lucky this time, and absolutely the proper procedure was not followed. It's my understanding that the procedure had been incorrectly modified at some point in the past as a way to decrease the settling time. Regardless, I don't believe it was the decision of these technicians to vary from the procedure in the manual, rather it being a case of being told "we've done it this way in the past". Regardless...

I've been a strong advocate for better training for our technicians. The training program we receive now is very different from the one that I went through. It's much heavier on code and electrical construction techniques, and much much less on technical and theory.

Takeaways from this incident are as you mention, verify integrity of grounds and follow proper procedure. What I'm still curious about, though, is if shorting the high side bushings not under test to ground had an effect on the outcome? Assuming the low side (not under test) was shorted together and NOT grounded, would they have been in just as much trouble had they only had a single ground on the H0, versus grounding the other 2 windings not under test?

 

[Skogsgurra]

As you say, the transformer is D/Y, meaning that there s galvanic continuity between all terminals in both windings. Then, it doesn't matter if there are one or several ground connections on each winding. So, I think that the answer to your closing question is: "Probably". It seems to me that someone either doesn't understand or remember very well what happened. Or prefers not to tell the whole story.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[marks1080]

Sounds like a bad ground and workers not respecting the job/procedure.

This should be put into the category of human error (negligence), not 'un-explained'