melted ground wire

[petersonra]

Recently I was up at a customer's plant. While I was there they had a minor explosion in their medium voltage switchgear.

On investigation, amongst other things, they found a ground wire that had melted from the current flowing through it. It sort of amazed me that a ground wire could melt and not manage to trip something to turn off the juice.

 

[charlie b]

Re: melted ground wire

Sounds to me like the ground wire had been undersized. This type of failure is the reason that the size of the ground wire is based (speaking loosely) on the rating of the load. Higher load means larger phase conductors, smaller resistance of the phase conductors, more fault current available, and therefore the need for the ground to be able to conduct away that higher fault current.

 

[Tom.Margillo]

Re: melted ground wire

I would hope that the "minor explosion" was the tripping of something if not a breaker. If not sounds like you have some serious problems with either the switch gear itself or the installation could be a combo I guess.

 

[petersonra]

Re: melted ground wire

Originally posted by Tom.Margillo:
I would hope that the "minor explosion" was the tripping of something if not a breaker. If not sounds like you have some serious problems with either the switch gear itself or the installation could be a combo I guess.

I do not have a problem at all. I stay away from medium voltage stuff.

 

[rcwilson]

Re: melted ground wire

I have seen several cases of melted ground wires on Medium Voltage (2.4 kV ? 35 kV) systems in industrial facilities. Sometimes it is the small ground wires from MV stress cones that melt when the cable faults. Other times it is a poorly installed crimp or bolted connection that develops high resistance and melts. And in a couple of cases it was because the ground/bonding cable was sized for the circuit full load amps and not the available short circuit.

A rule of thumb (from IEEE 80 ?Guide for Safety in Substation Grounding?) is for the ground conductor minimum size to be about 5 cmil/ amp for a 0.5 second trip time, 7 cmil/A for 1 second, 14 cmil/amp for a 4 sec trip and about 40 cmil/amp for a 30 sec trip time.

Example: A #10 wire is 10,380 cmils in cross section. That means a bare #10 can handle
2000 amps for 0.5 seconds,
1480 amps for 1 second,
740 amps for 4 seconds,
and 260 amps for 30 seconds before the copper melts.

Cables sized per the NEC Table 250.122 meet this criteria for most >600 V circuit breaker and fuse applications. For example, the #10 awg EGC can be used on circuits up to 60 amps. A typical 60 amp MCB trips on instantaneous at about 600 amps (10 x rating) and will clear a 600 amp fault in less than 0.5 seconds, well below the #10?s limit.

But in MV switchgear, the trips times can be longer and may not necessarily bear the same relationship between amp rating and time to trip. The designer must verify that the breaker or fuse clearing time is less than the EGC melt time.

MV systems can have overcurrent trip times of several seconds on the mains or large feeders. If the MV system has 10 kA fault current ( 5 MVA transformer @ 4.16 kV), a 4 second trip time requires 140,000 cmil (#3/0) and a 0.5 second trip time requires 5000 cmil (#2 awg) ground wire.

If the feeder is rated 200 amps, someone might put in a #6 awg EGC per the table. It will vaporize during a fault, probably making a load bang in the process.

Should something have tripped before the ground wire melted? Yes, but it may not have been designed that way.

Bob