Power Transformer Surge Arrester Connections 1

[jnims]

I am looking at the way power transformer surge arresters are connected to ground. Three methods I have seen used are (1) run bare conductors from the transformer's ground pad up the side of the tank to the surge arresters (2) use flat bar instead of bare conductors (3) use the transformer tank as the connection between arrester and ground grid connection. I anyone aware of any studies comparing these or other methods?

Thanks for any information you can provide.

 

[dpc]

You want as direct a path as possible from the arrester to earth. Use of the transformer tank seems like a bad idea to me. The grounding conductors should run in a straight path, with as few bends and connections as practical.

I'm sure there are some IEEE guidelines on this, perhaps in the Substation grounding standard, IEEE 80(?). Maybe others can find a specific reference.

 

[jwerthman]

I have also seen all three methods. I definitely don't like using the tank for the ground path. I generally make a loop through all of the arresters, connecting each end of the loop directly to the ground pad on opposite corners of the transformer. I usually try to keep this loop separate from the connection from the neutral bushing to the ground pad. It's probably overkill, but arrester discharges do unpredictable things sometimes, and I like a good solid connection to ground.

 

[busbar]

The connection that produces the lowest reactance is probably desirable. See IEEE Std C62.22-1997 Application of Metal-Oxide Surge Arresters for Alternating-Current Systems

thread242-19836 thread238-35386

 

[eduardom]

Surge arresters limits overvoltages produced by transient charge injection, in order to protect transformer insulation, for example, between windings and tank.
They should be connected to the tank. Grounding point is no important.
Grounding grid limits step and contact potential.
See ANSE/IEEE 80 (the green book).

 

[jghrist]

I agree with eduardom that the arresters should be connected to the tank at the base of the arrester. I would also provide a downlead to the grounding pad at the bottom which will be connected to the ground grid.

The voltage across the transformer insulation will be increased by the L di/dt in the downlead if it is isolated from the tank.

 

[MarkAthay]

The main thing you're trying to do is limit the overvoltage between the transformer windings, don't be fixated about keeping the absolute magnitude of the voltage down won't happen if you're hit by lightning....

The best performance is obtained by placing the arrestor under oil inside the transformer tank across the windings. This is done by the manufacturer. On pole-mopunted transformers a good way to do it is to mount the arrestor on he transformer tank, and connect the "ground" to whatever connection the transformer uses for the neutral. If it's a single-bushing transformer connect it to the tank. If it's a 2-bushing transformer use the neutral connection next to the transformer. Keep the connection distances short and straight.

I've seen studies that show close to 10kV generated in only 5' of wire during a lightning strike. That means that 5' of jumper will raise the effective protection level of the arrestor by 10 kV. Given that a 10 kV arrestor may clamp at 35 kV with fast rise-time surges, and you've thrown in another 10 or 20 kV or impedance with the jumpers, a 95 kV BIL transformer could quickly get into trouble....

Mark in Utah

 

[dpc]

to eduardom,

Grounding to the tank may help reduce voltage between the tank and the winding experiencing the surge, but it may actually increase the voltage between the tank and the other windings when the tank voltage rises well above "ground" potential.

 

[jghrist]

<<...it may actually increase the voltage between the tank and the other windings when the tank voltage rises well above &quot;ground&quot; potential.>>

The tank is connected to the ground grid no matter how the arrester ground is connected to ground. The tank will be at the potential of the ground grid where it is connected.

The bottom of the tank would have the same voltage rise to ungrounded windings as the station Ground Potential Rise (GPR).

If you connected the arrester ground only to the top of the tank at the arresters, you would get some L di/dt drop in the tank, depending on the inductance of the tank, so the top of the tank may have a higher potential than the bottom. This would be reduced by providing a separate downlead. I would expect the inductance of the transformer tank to be a lot less than the inductance of the ground lead. I think eliminating the connection at the top would put a lot more stress on the protected winding because of L di/dt in the downlead than you would have between the parts of the tank and ungrounded windings that may result from connecting the arrester ground at the top.

 

[cuky2000]

With the exception of arrester with surge counter, were insulation from the tank is required, I personally prefer the transformer thank as ground path for surge arrester.

Flat bars or regular conductors (bare and insulate from the tank) are similar method intended as separate ground path.

The surge arrester ground connection using separate ground leads are the most common practice. However, many utilities, Industrial and IPP’s installations had been using the transformer tank or metallic support structure for ground path for surge arresters. (Ref 9.6, IEEE Std 80-2000)

Both methods are technical sound and proven good practical experiences. For surge arrester with direct leads conductor connected to the ground, still larger amount of current flow via the transformer thanks since surge arrester are also direct bolted to the transformer under the following conditions:

a- Power transformers are typically grounded in two points to the ground grid.
b- Surge arresters are mounted in welded steel brackets.
c- The ratio of the copper conductivity is near 10 times the steel conductivity, but steel thank surface and cross section exceed 10 time the cooper leads equivalent surface and cross section.

Beware that if you work outside utility environment, electrical inspector may enforce the use of separate ground leads. With the pressure to get the job done, I not sure if worth fighting for approval with the local inspector.

 

[dpc]

To jghrist,

During a surge event, there will be many different voltages seen throughout the grounding system. The voltage at the top of the tank can be different than at the bottom. And the voltage of the tank will depend on how the arresters are grounded. Once the arrester starts conducting, the voltage at the arrester &quot;ground&quot; connection will be well above the voltage in substation ground grid.

Even a few tenths of an ohm of impedance can produce a considerable voltage difference if there are 100,000 amperes going through it.

 

[jbartos]

Suggestion: Visit

http://www.hipotronics.com/hvtest/resonant/hvtestbody19i.htm

http://www.usace.army.mil/inet/usace-docs/eng-manuals/em1110-2-3006/c-4.pdf

http://www.hubbellpowersystems.com/powertest/literature_library/pdfs4lib/EU1202.pdf

(for lead length calculation)
etc. for more info

 

[jghrist]

dpc,

All the more reason to ground the arresters at the top of the tank. This minimizes the voltage across the protected winding by eliminating the voltage drop in the arrester downlead (L di/dt). In the normal case, there will be arresters at the bushings of the other winding as well. These will protect the other windings if the tank voltage rises above the winding voltage. The other winding voltage will be at remote ground potential if it is ungrounded or at station ground potential if it is grounded.

If you isolate the arrester groundlead, the protected winding will be exposed to the arrester discharge voltage plus the L di/dt drop in the downlead. Other windings, if they are ungrounded, will be exposed to the station surge ground potential rise.

 

[dpc]

OK, I'm starting to come around I can see where grounding at the top of the tank could reduce voltage between tank and the protected winding, provided you can make a good connection. But it still seems to me that there is a trade-off regarding the other windings. If the arrester is directly grounded to the top of the tank, it would seem that a good portion of the tank would be at a higher potential relative to the other windings during the surge event. I guess as long as the othe windings are designed for this, it is not a problem. My concern would be for graded windings -- I would think a higher tank voltage to ground could be a problem.

As for cuky's comment on local inspectors, I agree that most would require separate grounding conductors, but I would think you could still ground to the tank in addition to this.

 

[jghrist]

Graded windings will be grounded to the tank at one end. True, other parts of the winding may have a voltage difference to other parts of the tank if some of the surge current flows through the tank. It would be very difficult to determine the voltage of different parts of the tank. I've never really worried about the winding insulation for a stroke to the other winding. I wonder if anyone has studied this in detail.

It is common to use a value of 1.6 kV/ft to estimate the L di/dt drop in distribution arrester leads. I would expect the inductance of a transformer tank to be less than that of an arrester lead, and certainly the parallel combination of the tank and a separate downlead would be less. So, let's assume a 12 ft high tank times 1.6 kV/ft would result in the top of the tank being no more than 20 kV higher than the bottom. The minimum insulation class for the neutral end of a 15 kV winding is 8.66 kV, with 95 kV BIL. Should be OK.

 

[dpc]

Thanks jghrist and cuky2000 - I learned something in this discussion.