The Weather Cleared and So Did the Fault: Decision Tree
Why this matters
The customer called during the storm, the freeze, or the heat wave, and by the time you arrive the weather has passed and the system runs fine. Every reading is normal because the condition that caused the fault is gone. Test it dry and you find nothing, close the ticket, and get the same call the next time the weather turns. A weather-event fault is diagnosed by recreating the condition, not by testing on the calm day the customer could finally get you out.
Start here: if the event pairs water with electricity, make it safe
The most dangerous weather faults mix water and power. Handle that before you recreate anything.
- Water intrusion into an electrical enclosure, panel, or fixture: de-energize the affected circuit and verify dead before you touch it. Do not power it up wet to "see if it faults." Dry it, find the entry, then energize.
- A breaker that trips only when it rains: treat it as a ground fault from moisture until proven otherwise. It is protecting against exactly the hazard you are hunting.
- Lightning-suspected damage: inspect for arc and heat damage with power off; storm surges can leave hidden failures that arc later.
Branch 1: name the weather variable
"Bad weather" is not a diagnosis. Pin the one variable that changed.
- Water - rain or wind-driven rain: intrusion, a failed seal, a flooded path, a blocked drain that only overflows in volume.
- Cold - freeze: frozen lines, contracted fittings, a component that only fails below a temperature, condensate freezing.
- Heat - heat wave: capacity ceilings, thermal expansion, a marginal component that only fails hot (see the peak-demand tree when heat and load combine).
- Humidity: condensation shorting a connection, a control misreading, a drain that cannot keep up.
- Wind: vent backdraft, a flame that will not stay lit, debris, movement that breaks a marginal connection.
Branch 2: recreate the condition on purpose
Once you know the variable, reproduce it in a controlled way instead of waiting weeks for the weather.
- Water: run a hose test, low pressure first, working from the bottom of a suspected intrusion upward, one zone at a time so you can see where it enters. Do not blast the whole wall at once.
- Cold or heat: where safe, apply localized cold or heat to a suspect component and watch for the fault to appear. Match the direction the weather pushed it.
- Humidity: introduce moisture near a suspect connection, or inspect for the condensation trail the last event left behind.
- Wind: check the vent or flame under an induced draft, or look for the movement path that only opens under load.
The point is to make it fail while you are standing there with a meter, not to guess from a dry system.
Branch 3: if you still cannot make it fail
Some events are hard to bench. Do not force a part swap on a dry system.
- Read the evidence the last event left: a water stain and its high point, corrosion at one terminal, a mineral trail, a scorch. The trail dates and locates the fault even when it is not active.
- Leave detection in place: a leak alarm, a logger, a moisture indicator, so the next event is captured rather than described.
- Correlate onset to the forecast with the customer, then schedule the recreate test for the condition you cannot fake.
When to pick which move
| Weather variable | Recreate method | Safety lead |
|---|---|---|
| Rain or wind-driven water | Staged hose test, bottom up | De-energize wet electrical first |
| Freeze | Localized cold on suspect part | Watch for burst lines under pressure |
| Heat wave | Localized heat plus real load | Confirm relief and limits work |
| Humidity | Introduce moisture, read the trail | Isolate before wetting a live circuit |
The recap
- If the event mixed water and power, de-energize and dry before you energize.
- Name the one weather variable that changed, not just "the weather."
- Recreate that condition on purpose - hose test, localized cold or heat - and measure during the failure.
- If you cannot reproduce it, read the trail the event left and leave detection for next time.
References
- NFPA 70 (National Electrical Code) on ground-fault protection and wet-location practice
- Trade-standard water-intrusion and hose-test methodology (staged, bottom-up)
- Manufacturer documentation on operating temperature and environmental limits
- See related: Temperature-Related Fault Cold vs Warm vs Humid; The Fault Only Appears at a Certain Time of Day