About 18 months ago, we had a large (1MW) solar system installed by a contractor on our commercial property; the power is being sold to a local utility.
Between each of the four 250kW, 480V 3-phase grid-tie inverters and the PV distribution panel, there are (3) 500 MCM CU-XHHW feeders installed in 4" conduit; most of this conduit is buried PVC. But there is one section of feeders (to Inverter #4) that is in EMT elevated along the back of some buildings. Two runs of buried PVC conduit are filled with standing water that entered through in-ground pull boxes. However, two of the conduit runs are continuous between the distribution panel and the inverters, and presumably are not filled with water. The elevated EMT is essentially dry inside.
Between the PV distribution panel and the PV main service gear, there are (18) 350 MCM CU-XHHW feeders installed in (3) parallel runs of 4" conduit. This conduit is buried PVC, but enters horizontally into an electrical utility room that is below grade. Because the un-sealed conduit openings in the utility room are at the same level as the 300-foot lateral run of buried conduit, it does not contain standing water (any water entering that conduit would drain out the openings in the utility room).
Problem 1: About a year ago, we experienced some faults: in the PV distribution panel, the 400A OCPD for Inverter #4 tripped several times; this is the only inverter with feeders in EMT conduit. It would trip either in the early morning when the system was coming online, or in the evening as solar production was tapering off. Each time after it tripped, we were able to successfully reset the breaker. We confirmed that Inverter #4 was operating correctly and was not exceeding its maximum rated output of 303A (additionally, the internal breaker in the inverter never tripped). We also swapped out the #4 OCPD in the distribution panel to ensure it was not defective, but the replacement also tripped. We had the feeders meggered, but each conductor showed no leakage. In all, the breaker tripped four times in early May 2013. We put a data logger on the feeders for a one-month period to try to capture the trip event data, but the breaker never tripped again. So we've been in a "holding pattern" to see if any other problems develop. My theory is that one of the feeder conductors has damaged insulation, exposing the copper wire. Most of the time, there is a sufficient air gap between this exposed wire and the EMT to prevent shorting (and to prevent the megger from seeing any leakage). But during early May 2013 (the first warm month since the system was put into operation), the wire was moving due to thermal expansion/contraction, just enough to cause a section of wire with damaged insulation to short out to the EMT on the four occassions where it tripped. After the breaker tripped, the wire would cool and contract, thus clearing the fault. Each time it shorted, some metal burned away to the point that now it's no longer able to close the air gap, explaining why it's no longer tripping the breaker.
Problem 2: As part of the installation of the solar system, the (16) 500 MCM AL-XHHW electrical service feeders bringing 480V 3-phase utility company power into our facility had to be removed from their four conduits and pulled back into two conduits (to free up conduits for the new solar feeders to the PV main service gear). The contractor who was installing the solar system also did this work. Last week, one of the conduits containing half our facility service feeders failed catastrophically: some of the phase conductors in the underground PVC conduit shorted out, destroying all eight wires within. Additionally, out of the remaining eight service feeders in the second conduit, we discovered leakage to ground on one phase conductor and one grounded conductor (neutral).
Problem 3: We also just discovered that another one of our inverter feeders, in buried PVC conduit, is leaking electricity into the water within that conduit.
In summary, of the six feeder runs that were installed during this project, we have three with confirmed faults. This does not give us confidence in the integrity of the remaining feeders. We will be bringing in a third-party electrical contractor to perform testing of all the feeders. We believe megger tests will be authoritative on all the sections of feeders that are submerged in water (which is the majority of the total length of wires). However, we are trying to determine how to conduct conclusive integrity tests on the hundreds of feet of wire in the conduit that is essentially dry. Obviously it would be difficult and costly to pull out all the feeders for a visual inspection.
Is there an established way to test for damaged wire insulation while it's still inside the conduit?
Between each of the four 250kW, 480V 3-phase grid-tie inverters and the PV distribution panel, there are (3) 500 MCM CU-XHHW feeders installed in 4" conduit; most of this conduit is buried PVC. But there is one section of feeders (to Inverter #4) that is in EMT elevated along the back of some buildings. Two runs of buried PVC conduit are filled with standing water that entered through in-ground pull boxes. However, two of the conduit runs are continuous between the distribution panel and the inverters, and presumably are not filled with water. The elevated EMT is essentially dry inside.
Between the PV distribution panel and the PV main service gear, there are (18) 350 MCM CU-XHHW feeders installed in (3) parallel runs of 4" conduit. This conduit is buried PVC, but enters horizontally into an electrical utility room that is below grade. Because the un-sealed conduit openings in the utility room are at the same level as the 300-foot lateral run of buried conduit, it does not contain standing water (any water entering that conduit would drain out the openings in the utility room).
Problem 1: About a year ago, we experienced some faults: in the PV distribution panel, the 400A OCPD for Inverter #4 tripped several times; this is the only inverter with feeders in EMT conduit. It would trip either in the early morning when the system was coming online, or in the evening as solar production was tapering off. Each time after it tripped, we were able to successfully reset the breaker. We confirmed that Inverter #4 was operating correctly and was not exceeding its maximum rated output of 303A (additionally, the internal breaker in the inverter never tripped). We also swapped out the #4 OCPD in the distribution panel to ensure it was not defective, but the replacement also tripped. We had the feeders meggered, but each conductor showed no leakage. In all, the breaker tripped four times in early May 2013. We put a data logger on the feeders for a one-month period to try to capture the trip event data, but the breaker never tripped again. So we've been in a "holding pattern" to see if any other problems develop. My theory is that one of the feeder conductors has damaged insulation, exposing the copper wire. Most of the time, there is a sufficient air gap between this exposed wire and the EMT to prevent shorting (and to prevent the megger from seeing any leakage). But during early May 2013 (the first warm month since the system was put into operation), the wire was moving due to thermal expansion/contraction, just enough to cause a section of wire with damaged insulation to short out to the EMT on the four occassions where it tripped. After the breaker tripped, the wire would cool and contract, thus clearing the fault. Each time it shorted, some metal burned away to the point that now it's no longer able to close the air gap, explaining why it's no longer tripping the breaker.
Problem 2: As part of the installation of the solar system, the (16) 500 MCM AL-XHHW electrical service feeders bringing 480V 3-phase utility company power into our facility had to be removed from their four conduits and pulled back into two conduits (to free up conduits for the new solar feeders to the PV main service gear). The contractor who was installing the solar system also did this work. Last week, one of the conduits containing half our facility service feeders failed catastrophically: some of the phase conductors in the underground PVC conduit shorted out, destroying all eight wires within. Additionally, out of the remaining eight service feeders in the second conduit, we discovered leakage to ground on one phase conductor and one grounded conductor (neutral).
Problem 3: We also just discovered that another one of our inverter feeders, in buried PVC conduit, is leaking electricity into the water within that conduit.
In summary, of the six feeder runs that were installed during this project, we have three with confirmed faults. This does not give us confidence in the integrity of the remaining feeders. We will be bringing in a third-party electrical contractor to perform testing of all the feeders. We believe megger tests will be authoritative on all the sections of feeders that are submerged in water (which is the majority of the total length of wires). However, we are trying to determine how to conduct conclusive integrity tests on the hundreds of feet of wire in the conduit that is essentially dry. Obviously it would be difficult and costly to pull out all the feeders for a visual inspection.
Is there an established way to test for damaged wire insulation while it's still inside the conduit?
