690 V arcing - signatures? Anyone?
690 V arcing - signatures? Anyone?
(OP)
Hello,
I am studying recordings from a fire that occured in a plant. Data: 6 MVA transformers fed from 6.6 kV 50 Hz. Secondaries are 690 V. Isc is around 32 kA.
I see initial arcing at around 10 kA. It stays for 400 ms. Then an 1800 ms period with no arcing (breaker has not yet tripped) followed by a 500 ms 10 kA arcing period with a 700 ms no-arc period. Then, there is a 1600 ms long arcing period with a rather high current, random variation between 20 and 30 kA and then breaker trip.
Protection is set at 1 kA/1500 ms and 5 kA no delay.
There was a water damage almost a year before this happened. the equipment had not yet been connected at that time. The equipment was dried and inspected and commissioned. There were some problems with electronic modules and cooling fans that had been damaged by water. And a contactor that had become rusty was changed.
The inverter is situated in a heavily ventilated electrical room. There is a raised floor with cables entering cabinets from below. Cabinets are welded steel with galvanized mounting plates.
There is extensive damage, but we have not seen any bad connections (torque controlled wrenches were used when building the equipment, no connections to these parts made on site). The fuses have tripped, of course, but no fuse explosion.
We are trying to understand what really happened. Dirt collected in voids in insulator system? Zink whiskers being transported by cooling air and settling on insulator surfaces? Anything else?
It would help very much if there were some similar recordings with a known failure cause where one could study/compare to see if we can have a clue from such a comparison.
I am studying recordings from a fire that occured in a plant. Data: 6 MVA transformers fed from 6.6 kV 50 Hz. Secondaries are 690 V. Isc is around 32 kA.
I see initial arcing at around 10 kA. It stays for 400 ms. Then an 1800 ms period with no arcing (breaker has not yet tripped) followed by a 500 ms 10 kA arcing period with a 700 ms no-arc period. Then, there is a 1600 ms long arcing period with a rather high current, random variation between 20 and 30 kA and then breaker trip.
Protection is set at 1 kA/1500 ms and 5 kA no delay.
There was a water damage almost a year before this happened. the equipment had not yet been connected at that time. The equipment was dried and inspected and commissioned. There were some problems with electronic modules and cooling fans that had been damaged by water. And a contactor that had become rusty was changed.
The inverter is situated in a heavily ventilated electrical room. There is a raised floor with cables entering cabinets from below. Cabinets are welded steel with galvanized mounting plates.
There is extensive damage, but we have not seen any bad connections (torque controlled wrenches were used when building the equipment, no connections to these parts made on site). The fuses have tripped, of course, but no fuse explosion.
We are trying to understand what really happened. Dirt collected in voids in insulator system? Zink whiskers being transported by cooling air and settling on insulator surfaces? Anything else?
It would help very much if there were some similar recordings with a known failure cause where one could study/compare to see if we can have a clue from such a comparison.
Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...






RE: 690 V arcing - signatures? Anyone?
Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...
RE: 690 V arcing - signatures? Anyone?
Are any of the connections burned off?
You can use all the torque controlled wrenches in the world but a year later the stuff can be loose again if there is heavy thermal transients. Maybe a cable came loose got high impedance and started cooking? That might explain on, and off, and on, behavior.
Keith Cress
Flamin Systems, Inc.- http://www.flaminsystems.com
RE: 690 V arcing - signatures? Anyone?
maybe there were reclosers in the upstream distribtuion system?
=====================================
Eng-tips forums: The best place on the web for engineering discussions.
RE: 690 V arcing - signatures? Anyone?
This, I was told, is so that the primary non-delay protection doesn't reach through transformer. An arc detector would have been just right here. But there wasn't any. Arcs are usually not expected here and arc protection had not been deemed necessary.
No, no reclosing. There are lots of data in the protection showing that the pick-up worked three times, but the trip worked only once, after the set 1.5 seconds delay. So, all these things worked as planned.
My problem is to understand what started the arcing. There was no overheating in the bus-bar connections (except from arcing), there are no cables connected to this part and there are no signs of undertorqued bolts and such things.
Anyone had any experience with water bringing debris and dirt into this kind of equipment? Or zink whiskers? Any recordings from similar "events"?
Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...
RE: 690 V arcing - signatures? Anyone?
Are you sure the upstream power from the distribution system did not clear and reclose several times? If yes, how do you know?.... Did there remain some lower level of current flowing during the in-between periods when the fault was not present?
=====================================
Eng-tips forums: The best place on the web for engineering discussions.
RE: 690 V arcing - signatures? Anyone?
But you also said "Protection is set at 1 kA/1500 ms and 5 kA no delay.". But the level went three times to at least twice this level, for durations of 0.4 sec to 1.6 seconds. On the surface, it sounds like the protection did not work properly, and that's why you had a fire instead of simply a trip with minor damage.
Am I missing something?
=====================================
Eng-tips forums: The best place on the web for engineering discussions.
RE: 690 V arcing - signatures? Anyone?
=====================================
Eng-tips forums: The best place on the web for engineering discussions.
RE: 690 V arcing - signatures? Anyone?
=====================================
Eng-tips forums: The best place on the web for engineering discussions.
RE: 690 V arcing - signatures? Anyone?
I know with a very high degree of certainty that the upstream breakers did not trip. This is a very modern plant with computers and digital protection. Everything is recorded. Also, tripping of upstream breakers would have stopped much of the plant. And that did not happen.
I am not so confused by the coming and going of the arcs. Arcs are usually unstable and restrike several times before finding a "nice place" to develop fully.
What I am after is recordings from similar events. Just to have something to compare with. I am also interested in possible root causes.
Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...
RE: 690 V arcing - signatures? Anyone?
You have addressed my question about whether the utility distribution system reclosed, if you are saying there are other plant loads fed from the same utility feed that didn't trip during this event.
My mistake on the voltage level of fault the fault current vs the protection.
Still, one would think that the investigation of this event would focus heavily on why the complete protection system allowed a fault continue for so long and progress to point of a fire. It would seem it was either improperly designed or it didn't work as designed. My vote is improperly designed if you're saying that the only instantaneous protection is set at 5KV primary, around 50KA secondary, and the available fault current is only 32KA secondary.
=====================================
Eng-tips forums: The best place on the web for engineering discussions.
RE: 690 V arcing - signatures? Anyone?
Yeah. Thanks for the bumps!
Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...
RE: 690 V arcing - signatures? Anyone?
Preliminary result.
I visited the site today. There is a very limited initial burn, that melted part of bus-bar. This was after the 800 A fuses and seems to have started it all. The 800 A fuse tripped and it seems that the arc spread and started a new arc from the fuses that hadn't tripped. The plasma then started a full shorting arc between the incoming bus-bars (no fuse any more) and started the real fireworks that burned through cabinet walls before the primary protection worked after 1500 milliseconds.
Comments invited. (Yes, they should have had arc guards.) But anything else?
Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...