Ground Faults Only on Windy Days and After Rain
Infrastructure Application Domain 3 · Task 3.EA site reports intermittent ground-fault trips on one inverter. They cluster on windy days and in the hours after rain, and the inverter restarts cleanly every time. Up on the roof, the cable ties on one home run have perished, the conductor now rests across the edge of a rail, and there is a shiny worn patch on the jacket where it has been bearing on the metal. What explains the pattern?
Reveal answer and explanation
Correct answer: C) The worn jacket bearing on grounded rail is the fault: wind moves the conductor and water bridges the damage, so the path only exists in those conditions
The intermittency is the diagnosis rather than an obstacle to it. Something has to explain why the path exists on some days and not others, and the roof supplies it twice over: wind moves an unsupported conductor against the rail it is resting on, and rain provides the conduction that damaged insulation on its own does not. Put those together and you get a fault that is genuinely present in wind and wet and genuinely absent otherwise.
The chain that produced it started years earlier with the cable ties. Support is what keeps a conductor off the edges it would otherwise saw against, and when the ties perish the conductor does not fall to the roof dramatically, it settles onto whatever is beneath it and begins to move in the wind. Everything after that is a matter of time and jacket thickness. The shiny patch is the clock showing how far along it has got.
It is also worth seeing what this fault is on its way to becoming. An intermittent earth path through wet, damaged insulation is not a stable condition. It carries current when it closes, the damage grows, and a conductor worn far enough to reach copper against grounded steel is a ground-fault or an arc rather than a trip report. The inverter restarting cleanly each time is not reassurance, it is the system doing its job while the underlying condition gets worse.
The O&M point: cable management is the part of an installation that has a service life shorter than the system, and nobody schedules it. Modules last decades, racking lasts decades, and the ties holding the wiring up were chosen on the day and are the first thing to go. Walking a roof to look at what is holding the cables rather than at the cables themselves is unglamorous and it is the inspection that catches this class of fault while it is still a worn patch.
Why the other options are wrong
- A) The ground-fault device is faulty, because a genuine ground fault would trip consistently rather than come and go
This reads intermittency as evidence that the detector is wrong, when intermittency is a thing faults genuinely do. A protective device that operates only sometimes is reporting a condition that exists only sometimes, and the correct response to that is to find what changes. Declaring the instrument faulty is the conclusion that ends the investigation, and it ends it on the roof where the evidence was.
- B) Moisture in the combiner box, which is the usual cause of ground faults that track the weather
Water in an enclosure is a real cause of weather-correlated ground faults and it is a reasonable thing to check. What makes it wrong here is that the scenario contains the answer already: a worn jacket bearing on grounded metal, on the circuit that is tripping. Going to the usual suspect while ignoring the visible damage in front of you is how a straightforward fault turns into three visits. Check the combiner too, and check it second.
- D) Nuisance tripping from an over-sensitive inverter, which is best dealt with by raising the trip threshold
This is the answer that makes the symptom go away and leaves the fault in place, which is the worst outcome available here. The device is detecting current going somewhere it should not, and that current is real. Raising a protection threshold to stop a genuine detection is a decision to accept an unknown fault on a roof people walk on, and it will be found later by whoever is looking into why the array caught fire.
References
- NABCEP OMAT JTA v.2026.5 — Domain 3, Task 3.E: "Wiring best practices (e.g., PV connector compatibility, drip loops, service loops, minimum bend radius, support, exposure to damage)". Support and exposure to damage are the two named here that the question turns on.
- No standard is cited. The answer follows from what abraded insulation against grounded metal does when it is wet and moving, and from the fact that the scenario already contains the damage.