required bonding causes fire

[don_resqcapt19]

Last night a traffic accident took down a power pole and 2 of the 3 conductors of the service drop. One ungrounded conductor of the drop remained connected to the utility. The riser was damaged and the still connected ungrounded conductor was shorted to the grounded conductor. The water piping was bonded, but was served by a nonmetallic service. Current flowed on the water pipe to the gas water heater, and then on the flex gas line. I am not sure of the complete path on the gas line, but I think there was a path via the metal gas piping, the gas meters and the grounded conductor at the other unit of the duplex. The current either burnt a hole in the flex line or burnt it completely off and the gas was ignited making a torch in the utility closet. The flex line was completely seperated when I saw it, but it is possible that the current only made a hole and the burning gas did the rest. It is also possible that the current did the complete job. The home owner called the fire department and then put out the gas fire with a dry chem extinguisher. The FD shut off the outside gas valve and stood by until the power company arrived to cut the hot wire.
Don

 

[wbalsam1]

Hi Don: In your opinion, would it have made a substantial difference if the csst was bonded to the grounding electrode system by a #6 cu as per the manufacturer's tech bulletin?

 

[infinity]

Hi Don: In your opinion, would it have made a substantial difference if the csst was bonded to the grounding electrode system by a #6 cu as per the manufacturer's tech bulletin?

What tech bulletin are you speaking of? The one regarding flexible gas pipe? I think that Don's post had metal piping with a flex connection to the water heater.

 

[don_resqcapt19]

I think that the only prevention would be a bonding jumper run with the flex gas line. The current path was via the flex connection and any bonding other than a bonding jumper around the flex line would not have made any difference. Also I'm not sure this is the same product as the CSST that is used for the gas distribution in the house. This was only a 30" flex connection between the water heater and its gas valve that is installed on the end of the black rigid gas pipe.
Don

 

[wbalsam1]

What tech bulletin are you speaking of? The one regarding flexible gas pipe? I think that Don's post had metal piping with a flex connection to the water heater.

Oops! Now that I read his statement more closely, I think you're right.

 

[wbalsam1]

I think that the only prevention would be a bonding jumper run with the flex gas line. The current path was via the flex connection..........

Perhaps a proposed amendment/addition to the NEC reflecting this very same concern...?

 

[NoVA Comms Power]

Don:

Very Interesting!

It certainly lends some creedence to the folks who initiated (and won) the CSST lawsuit.

While I'm not at all doubting your paraphrasing abilities ... is there any more info re: this incident?

I realize that you actually "saw it" ... but is there any "write-up" or initial investigation / analysis available somewhere we could read for additional info?

Jim

 

[don_resqcapt19]

Jim,
No there will not be any real investigation or published report on this incident. One thing is that my heading is incorrect. If the gas piping would have been bonded as required by the current code, this may not have happened. With a bond to the gas pipe, both the water and the gas pipe would have been at the same potential. If the bond to the gas pipe was smaller than the one to the water pipe, then it is possible that the bonding conductor would have gotten hot enough to start a fire, but that bonding conductor would have limited the current through the flexible gas connection.
Don

 

[NoVA Comms Power]

Don:

Thanks for the quick reply; it clearly backs-up your earlier statement ...

"I think that the only prevention would be a bonding jumper run with the flex gas line."

It would seem from this incident that the thin-wall gas supply tubing (whether is was actually CSST or not is irrelevant) can't be trusted to carry "max-fault current" ... and that if the tubing's part of the fault path (due to bonding) ... then it should have a "suitably sized" electrical jumper connected in-parallel.

Talk about opening-up a can of worms !!!:smile:

Jim