For our solar generation plant, we have a run of feeder conductors from a 480VAC 3-phase distribution panel (with OCPD) to a 250kW central inverter. The 400A breaker in the distribution panel has been tripping intermittently. According to our inverter monitoring system, the faults have occurred once early in the morning, once late in the afternoon (both events during solar production, but at reduced capacity), and once late at night well after production has stopped and the inverter has disconnected itself from the grid power (with the exception of a small internal power tap for the inverter's monitoring & control circuitry).
To rule out the possibility that the breaker is faulty, we swapped it with the breaker from another inverter feed, but still got a "trip" on the same circuit as before. I don't think it's possible for the fault to be within the inverter because of the "trip" that occurred at night, after the internal contactor within the inverter would have disconnected the primary grid power. So that suggests we have a fault in our feeder conductors between the distribution panel and the inverter. I suspect that the feeders are moving as they heat up and cool down, possibly allowing a section with insulation damage to short either phase-to-phase or phase-to-ground. The feeders are 500mcm XHHW-2 600V copper. The run is ~700 ft: 400 ft below grade in 4" PVC and 300 ft above in 4" EMT.
The installation contractor did perform an insulation test on the feeders using a 500VDC megger, but did not detect a fault. However, the inverter had been shut off for a day prior to that test so perhaps the feeders had cooled and contracted to a position within the conduit where there was sufficient air gap to prevent leakage.
At this point, we have the inverter back online to see when it trips again while we try to determine the best way to diagnose and correct the problem.
Ideas:
Any advice on how we can test and locate this fault would be greatly appreciated.
To rule out the possibility that the breaker is faulty, we swapped it with the breaker from another inverter feed, but still got a "trip" on the same circuit as before. I don't think it's possible for the fault to be within the inverter because of the "trip" that occurred at night, after the internal contactor within the inverter would have disconnected the primary grid power. So that suggests we have a fault in our feeder conductors between the distribution panel and the inverter. I suspect that the feeders are moving as they heat up and cool down, possibly allowing a section with insulation damage to short either phase-to-phase or phase-to-ground. The feeders are 500mcm XHHW-2 600V copper. The run is ~700 ft: 400 ft below grade in 4" PVC and 300 ft above in 4" EMT.
The installation contractor did perform an insulation test on the feeders using a 500VDC megger, but did not detect a fault. However, the inverter had been shut off for a day prior to that test so perhaps the feeders had cooled and contracted to a position within the conduit where there was sufficient air gap to prevent leakage.
At this point, we have the inverter back online to see when it trips again while we try to determine the best way to diagnose and correct the problem.
Ideas:
- I think we need to megger the feeders again at a higher test voltage (although, not a hi-pot test that could be destructive to the insulation).
- I think it would be beneficial to do the insulation test after the inverter has been online for the day so that the conductors are heated, although I'm not sure how long we should wait after shutdown before meggering, or how long we should perform the megger test. (Hours while waiting for the conductors to cool?)
- Since we don't know which of the three phase conductors might be faulty, we might have to test all three simultaneously as they cool, but I'm not sure how we could do that.
- Since we have an intermittant fault, how can we help make it more consistent for troubleshooting? Should we fill the buried PVC conduit with water? Of course that won't help for the EMT that is above ground.
- I've also read about "thumping", but it may be premature to try that. I'm also concerned about the possibility of damage to the conductor insulation by thumping.
Any advice on how we can test and locate this fault would be greatly appreciated.
