Seperate Grounds for Metal-Oxide Surge Arrestors on Poles? 1

[majesus]

Hello,

We are installing a 69kV EPR incoming service cable from O/H lines to feed our plant. At the O/H power pole structure there is a 3 phase disconnect switch gang operated by a lever handle at ground level. The power pole is wood and sits in a steel pile. There are three ground rods, a #4/0 parameter grid and ground mat for the switch operator. The lever handle, system ground, ground mat are all bonded together and connected to the ground grid, grounds rods and pile. On the load side of the disconnect switch we've also
installed 72kV System (57kV MCOV) rated Metal Oxide Surge Arrestors to protect the 69kV cable. The question is what is best practice for the Arrestor's ground leads? Is it industrial standard to bring a separate #4/0 conductor from the grid and tie all three arrestor's ground leads or is a separate ground grid used? I'm worried about Step-touch potentials due to lightning or in case of a leaky arrestor.


Thanks,
Majesus

 

[stevenal]

The rule is to keep lead length as short as possible. In this case, the purpose of the arrester is to protect the EPR, so the lead length in question is the distance from the terminator to the arrester plus the distance from the arrester ground terminal to the cable shield. Your suggested solutions all apparently work to maximize lead length rather than to minimize it.

 

[majesus]

Thanks for responding Stevenal, but you misunderstood the question.

The arrestors are located on top of the power pole and the 69kV EPR cable's termination is connected to the arrestor's line side terminal as close as physically possible. The arrestor's ground lug is connected directly to a #4/0 AWG soft drawn bare ground conductor that runs directly straight down the pole and into the ground. (This is the shortest possible route.) Furthermore, the cable's shield is connected to the system ground.

My question has nothing to do with connecting the cable to the surge arrestor. The question is about step touch potentials.
What is best practice for grounding the arrestor's ground leads when you also have an operator's ground mat bonded to the same grid?

Stevenal, I am interested in the experience from the forum members about connecting the arrestor's ground lug to the ground grid that also bonds the ground mat and gang mounted switch or is it a better practice to build a separate and isolated ground grid solely for purpose of grounding the arrestors. I'm worried about Step-touch potentials due to lightning or in case of a leaky arrestor.

 

[stevenal]

Perhaps you misunderstood my answer. You now state that the arresters and the shield are independently connected to the grid? This ensures your lead length on the ground side is twice as long as it needs to be: arresters to grid plus grid to shield.

Isolated ground grids are never a good idea. Too easy to get a human in between them and your cable will be unprotected.

Connect arresters directly to the shields and ground the shields. Or connect the shields to the arresters and ground the arresters. Connect the switch handle to the grid. Also connect the switch handle to the mat. Do not connect the mat directly to the grid (I think Waross told me that few years back). Train your operators to stay on the mat when switching.

 

[majesus]

Thanks Stevenal,

My original post was a question about isolating the arrestor from a switch mat bonded to a ground grid to avoid step-touch potentials. I was focused on this topic that I did not realize you were discussing the fact that the arrestor's terminal leads also includes the ground lead between the arrestor and the equipment it is protecting. Prior to your comment, I thought the arrestor’s lead must only be minimized at the line side, but as you mentioned, it is actually both sides because the voltage rise the occurs across the arrestor and the leads is in parallel with the equipment. To minimize this voltage rise and thus enhance protection, you must minimize the leads on both ends. Thank you for pointing this out.

Stevenal I have a few more questions just for discussion and understanding:
You recommend to connect the mat directly to switch and not to the grid? ? Why is that?
You also mention that isolated grids are not a good idea, however if the isolate ground grid is buried and the area is covered in dielectric (substation grade) crushed rock, won’t this risk be minimized and the step touch hazard during a lighting surge be avoided?


Arrestors Connected to Cables
In this design, the engineers are using the cable’s shield as the fault/system ground return path back to the substation. The substation distribution includes a 200A neutral to ground resistor. It has been calculated that if a fault occurs, the shield would be able to handle the fault magnitude until the breaker trips. The shield is bonded to the overhead system ground conductor. The thinking is to isolate the system ground from the arrestor's ground in the event of lighting then the arrerstor’s surge current will not travel along the overhead system ground. This is similar to lighting rod conductors where the conductor is driven directly into the ground and it is not bonded to the system grounding. Therefore, why does this not apply to the arrestor bonding leads as well?

Lastly, in some circumstances, the shield at one end of a cable is left isolated to avoid circulating currents, if this was the case in this application, where the shield was isolated at the overhead pole, then I suppose the lightning arresting would provide poor protection.

Regards,
Majesus

 

[stevenal]

I found the discussion I had with Waross, see thread238-148268. Don't think I can explain it better.

The isolated grounding also recalls old discussions you might try searching. Usually it involves some communication equipment connecting to a "clean" and isolated ground. And just as in your scenario, the issue is not so much about step potential but about transfered potential. Consider for example the worker who plugs his case grounded electrical tool into an extension cord that extends from one ground to within reach of another, and then a system fault occurs.

When protecting humans or insulation from overvoltage, I find there are two mutually exclusive philosophies. 1) Divert the harmful potential before it gets to the thing you wish to protect. Farther away is best since you don't want any chance of this stuff getting anywhere near. 2) Short circuit the thing you wish to protect with a thick copper conductor as short as possible. Circuit analysis favors the second method.

Yes, if you don't connect the arresters to the shield; you may as well install them elsewhere where they might actually be protective. Or you could look at other means to minimize circulating current.

The IEEE surge protection standards are worth the study.