underground distribution system grounding

[tem1234]

Hi

On a MV overhead distribution system, 4 wire solidly grounded. The neutral is grounded every xx pole or at every pole mounted transformer or other equipments.

We have to make a little section of underground on this line with 3 shielded cable. The shield act as the neutral wire.

So, is it better to connect the both side of the shield to the overhead neutral wire (and the ground), or it's best to leave one side of the shield not connected to avoid current loop in the underground circuit?

Thanks

 

[jghrist]

If the only metallic connections between neutrals at the cable ends are the cable shields, then the shields should be grounded at both ends to provide a low impedance ground return path.

If there is a separate ground wire, then you need to decide based on:

1. If both sides are grounded, there will be circulating currents induced in the shields which will cause heating and reduce the current carrying capacity of the cable.
2. If one side is ungrounded, there will be a voltage induced in the ungrounded end of the shield that may be hazardous.

Either of these problems can be dealt with. Most US utilities will ground both ends of MV cable shields and derate them accordingly.

 

[tem1234]

Thanks jghrist, that help me a lot

There's no ground wire, so the shield should be grounded at both end.

by looking at a manufacturer data for a similar cable, i saw that the ampacity of the cable are calculated with 1/3 of nominal current in the shield and the shield are bonded at both end. So the cable are already derate for circulating current.

 

[apowerengr]

This is a 4 wire circuit. You need to maintain continuity of the neutral conductor throughout the circuit. The "shield" as you called it is really a neutral and needs to be continuous and sized accordingly. Whether it's grounded at the riser poles or not is secondary to maintaining continuity of the neutral conductor. I'll assume you're installing lightning arresters at the riser poles so pole grounds would be required at these locations and the grounding conductor would be connected to the neutral conductor at these points.

 

[tem1234]

Hi apowerengr,

That's what we were thinking, but we look at an other similar circuit on the same location that we didn't do and the shield isn't connected on one side, so that made us confused.

I begin to think that we should tell the client his installation isn't correct and he must connect the shield. I know that in Low Voltage it's a code violation, but in distribution i'm not sure.

 

[rockman7892]

I have always wondered the same question for MV shielded cables for a motor feed for example. If the motor leads are going through a zero sequence CT, then should the shields be landed to ground at both ends, or only the source end. If both ends are grounded and charging current drains through the shield to ground at the motor end, wont this cause false trips?

 

[tem1234]

HI rockman

Since the motor don't need neutral, and the shield doesn't have to carry current, i would ground only the source end of the shield. This will prevent circulating current.

You can ground both end, you will have circulating current and have to derate the cable and don't pass the sheild in the zero sequence CT (or pass it 2 time so it will cancel)

 

[rockman7892]

tem1234

Thanks for the response. I understand your comment about passing x2 through the CT

What about for a feeder circuit from a breaker to a switchgear lineup? Same thing with only grounding at one end?

So you are saying in this case to land the EGC to the ground bus bus, and tie all of the cable shields together but do no ground them at the switchgear lineup ground?

 

[jghrist]

I'd be concerned about grounding only one end of a shielded cable (single core shielding not overall shielding) regardless of the voltage if the cable can carry a large current. Voltage induced in the shield is a function of the conductor current, not the conductor voltage.

 

[rockman7892]

jghrist

I agree with your point about induced voltage. I was thinking that leakage current onto the sheild draining at a load ground could cause issues with the zero sequence CT's.

Maybe this is why we have to set a time delay on our ground settings?