Arcing fault at 480/277V panelboard, overcurrent protection, and grounding 1

[jmbelectrical]

I am investigating an incident in which a failure, resulting in a fire, occurred at a 480/277V, three-phase, four-wire panelboard. I do not know the causation of the failure at this time, however, I know that arcing occurred at some point due to the presence of irregularly-shaped openings along the panelboard's enclosure.

The panelboard was supplied by a service disconnect switch, fused at 400A with dual-element time delay fuses. None of the fuses blew. The panelboard was supplied by an insulated, #2 AWG (copper) equipment grounding conductor which originated at the service disconnect switch. The equipment grounding conductor lug located within the service disconnect switch arced off such that a portion of its fastener (that is, the fastener used to attach the lug to the enclosure) was missing. The lug was detached from the disconnect switch's metallic enclosure wall. Evidence of arcing (burn marks and an irregularly-shaped opening) was also observed adjacent to the disconnect switch's neutral conductor lugs.

The electrical utility transformer had a capacity of 150kVA. Based on its %Z, the maximum secondary-side fault current is just over 4,000A, assuming an infinite primary bus.

I know that arcing faults are associated with fault currents 30 to 50 less than that of a bolted three-phase fault, however, I still expected one or several of the fuses to blow at some point during the incident. Is this a reasonable expectation? And what could have caused arcing and detachment of the equipment grounding conductor lug within the service disconnect switch? Is this simply a matter of insufficient fault current for it to be cleared?

 

[wayne440]

What were the weather conditions at the time of the incident?

 

[jmbelectrical]

wayne440,

Clear. Weather data did not indicate the presence of lightning both on the date of the incident and several days prior.

 

[Heaviside1925]

Was the system still energized when the fire was noticed? Was the disconnect opened to make the system safe? Can you describe the equipment grounding system?

 

[wcaseyharman]

In my experience arcing around connectors suggests high resistance connections and intermittent contact, most likely due to being loose. With a high resistance in the ground connection at the disconnect there could be insufficient fault current to blow the fuses when a ground fault occurs.
Loose connectors can be caused by poor assembly, heating and cooling cycles and vibration, among other things.

That’s my $0.02 anyway.

 

[Kiribanda]

1) Is your 480V system is HRG or solidly grounded?
2) What is the Make & Model of the 400A fuse?
3) Could you please upload few photos?

 

[waross]

A large single phase load a and a poor connection may result in that damage if the neutral current is flowing in that conductor.
Inspect the entire installation to see what could have caused a current to ground.


--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[dpc]

These types of arcing ground faults and resulting fires were the driving reason behind the NEC requirement for ground fault protection for larger (1000 A) services back in the early 70s or so. It's unusual to see this for a smaller transformer. It is completely in character for the overcurrent protection to fail to operate for this type of fault. In many cases, the fire was there first external indication of a problem. 277 V is the "sweet spot" for these types of faults. The voltage drop in the arc limits the fault current to a level too low to clear but 277 V is sufficient to maintain the arc. As to the root cause, I can't say, but based on your description, this seems like a classic sustained arcing fault at 277 V.

 

[waross]

I am amazed by the number of sparkies who don't know how to properly tighten stranded cables into a lug.
Try this sometime;
Insert a stranded cable into a suitable lug with a pressure pad that tightens against the cable. (Lugs with pointed set screws that deform the cable are an exception)
Tighten the set screw hard.
Tighten it to several times the inch pound recommendation of the manufacturer.
Now twist the cable slightly clockwise and counterclockwise.
Now easily drop the cable out of the lug.
7 Strand cable is the worst, and the easiest to demonstrate.
It doesn't take much current to sustain a micro arc.
Walking by a 200 Amp service switch I could feel the heat radiating about 3 feet away.
I opened the cover and although the switch was stillin service with about 100 or more Amps flowing, I could see micro-arcs inside the switch blades. The switch blades were covered with heat corrosion.
The voltage drop across the arcs was about 1 Volt.(Measured on the milli-Volt scale.)
The point is that with a loose connection, it doesn't always take fault level current to do a lot of damage.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[che12345]

"... The equipment grounding conductor lug located within the service disconnect switch arced off such that a portion of its fastener ... was missing. The lug was detached from the disconnect switch's metallic enclosure wall. Evidence of arcing ... was also observed adjacent to the disconnect switch's neutral conductor lugs".
1. As per NEC, LV grounding resistance <25 Ohm, no further supplementary grounding electrode is required.. Ground fault protection is NOT mandatory for a 400A board.
2. Assuming 25 Ohm ground resistance, a dead short between the Line and the Ground Ilg =277V/25 Ohm = 11.08A. This Ilg is far too low to cause 400A fuse to blow, assuming there is no other loads.
3. Assuming a loose connection between the earthing conductor lug and the board say 2 Ohm. The localized heat generated would be I[sup]2[/sup] x R , = 11.08 [sup]2[/sup] x 2 = 245.52 W. An Arc could occurred which causes further burning. Attention: If Ilg is < 400A, the arc would persist burning.
4. Proposal: a) Check and torque to ensure low contact resistance (especially at painted connection points),
b) Add supplementary grounding electrode to say < 1.0 Ohm . This could be costly and may be not practical/achievable, see c) below.
c) Add RCD or Ground fault protection with set current at < 0.1 of system current rating etc..
Che Kuan Yau (Singapore)

 

[waross]

Once again you have demonstrated your unfamiliarity with North American practice and methods.
We must make a clear distinction between transmission and distribution systems and low Voltage utilization wiring.
Your calculations may be arithmetically correct for a fault that seldom occurs, and then only if there is some other unusual enabling condition.
Grounding serves two purposes;
1. To limit touch and step potentials.
2. To provide a low impedance path to facilitate the operation of protective devices.
The grounding electrode to which the 25 Ohm limit applies is for the purpose of limiting touch potentials. Step potentials are generally not an issue for 277/480 Voly systems.
The grounding electrode is rarely involved in faults.

The low impedance path to trip breakers is provided by the equipment grounding conductors.
Atypical ground fault path in a 277/480 Volt system will typically be from the transformer, through the breaker, through the feeders and circuit conductors, and returning via the equipment grounding conductor(s).
Typically, the 25 Ohm grounding electrode never becomes involved.

In the event that a ground fault occurs in a motor winding or transformer winding or in a heating element, at such a location that the winding limits the fault current, the fault current may be less than the trip setting of the breaker.
In the event that a heating element breaks and becomes grounded near the neutral point, close to full load current may flow in the grounding and bonding circuits.
This may persist until the arc at a loose connection burns clear.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[wcaseyharman]

That’s a good point about a fire caused by load current through a high resistant connection, that certainly would explain the fuses not operating. We’ve seen failures like that which caused quite a bit of damage and ended up leading to a fault.

 

[jraef]

I see a lot of people who don’t seem to understand the concept of a high resistance fault not being cleared by circuit protective devices. I think there has been some inadequate education on that front.

I just looked into a n incident where the load terminal of a 400A breaker completely melted down and the breaker never tripped, until the resulting arc finally burned away the plastic parts and made contact with the back panel. I determined the cause as being the use of “DLO cable” on the load side wiring, but terminated into the standard lugs of the breaker, which were not listed for high strand count wires. Yet the bigger argument I got from the owner was that the breaker should have tripped earlier. Unfortunately it was feeding a heating load, so the added resistance actually DECREASED the current during the event. A little basic Ohm’s law math finally convinced him.


" We are all here on earth to help others; what on earth the others are here for I don't know." -- W. H. Auden

 

[SilverfoxUK]

You say you have a 150kVA supply but don't give an impedance. So with a few assumptions in place I've got a fault current at the transformer of about 3500A. Travel even a short distance through the feeder cables and 3500A till drop quickly. I don't know what the fuses are but a BS88 takes 1 second to clear at 3850A. Surely max line current is less than 200A. I'm thinking at 400A these fuses are far too big to protect the installation. 1 sec is a long time for a fault.

 

[cranky108]

This might be why arc detecting technologies were developed, and I believe they are available in circuit breakers. But I don't know if those are available for 480/277 V.

 

[jmbelectrical]

My apologies for the delayed response. For those who asked:

1. The transformer's secondary is 480/277V, wye-connected, solidly grounded.
2. The transformer's impedance is 4.48%.
3. The fuses are 400A, dual-element time delay type, Trionic TRS400R.

 

[che12345]

"...1. The transformer's secondary is 480/277V, wye-connected, solidly grounded. 2. The transformer's impedance is 4.48%. 3. The fuses are 400A, dual-element time delay type, Trionic TRS400R".
I tried to visualize the set up.
1. 3-phase supply 480/277V to 400A fused Service disconnect sw. Its casing is used as the Ground G point. There are 2 conductors, a) The Equipment grounding conductor which is distributed to the load, and b) another Bonding conductor, bonding the G point to the Neutral point N , which is located near the Service disconnect sw. The Load is Grounded by the Equipment grounding conductor and other Bonding connections, if any.
2. When the connections are in order: any arc-fault, the fault current If flows from phase 277V to the load frame, to Equipment grounding conductor, to Service disconnect s/w point G. to Bonding conductor to Point point N, ends up at the trafo Neutral. Note: the 25 Ohm grounding resistor is NOT involved.
3. When the Equipment grounding conductor is with loose connection (high contact resistance) at the G point on the Service disconnect sw:
The arc-fault current If flows from the load frame to any other bonding to the Ground, to 25 Ohm? (or may be higher) grounding resistor , to the trafo Neutral. This low If would NOT blow the 400A fuse. Note: The trafo is deemed as solidly grounded but is actually grounded by the 25 Ohm? grounding electrode resistance, accepted by NEC.
Che Kuan Yau (Singapore)

 

[waross]

Solidly grounded refers to a solid connection between the transformer neutral and the equipment grounding bus.
This is to distinguish the connection from impedance grounding where an impedance is inserted between the transformer neutral and the equipment grounding bus to limit fault currents.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!