Severe Weather and Solar Panels

Severe weather and a solar array meet in three different ways: snow that covers the glass, wind and hail that load the structure, and the electrical equipment that still has to be shut down safely when the site is a mess. They are not one “storm mode.” Snow has a published energy model. Hail does not have a single diameter that this article can honestly print, because no test report with a diameter was opened. The useful habit is to separate the energy loss from the mechanical rating, and to leave the array alone while it is icy, live, and above you.

Snow is the severe-weather loss you can estimate

NREL’s snow model for the System Advisor Model treats sliding as the main way snow leaves a tilted module. New snowfall is assumed to cover the array. Across the weather files they ran, fixed-tilt annual snow losses were about 0 to 16 percent of annual energy in the continental United States for the tilts mapped, and some Alaska sites reached about 40 percent. That range is the estimate. It is not a reason to climb. The full account, including why a low-tilt roof sheds less and why brushing is not the model’s advice, is in snow on solar panels.

After the snow slides, look at where it lands. Federal installation guidance calls out gutters, icicles, and the path snow will take, including layouts meant to keep a walkway clear. A ground-mount aisle filled with the snow the modules just shed is not a maintenance path until it is safe to walk. Production that stays down after the glass is clear is no longer a snow loss. Then you are in ordinary troubleshooting, outside the roughly plus-or-minus 10 percent insolation band Pacific Northwest National Laboratory treats as normal weather variation.

Wind and hail are ratings, not slogans

Wind load and snow load belong to the rack, the clamps, the roof or the piers, and the module’s mechanical rating. A buyer’s specification used on one Xcel-related solar procurement says module framing should resist damage from snow, wind, hail, and blowing dust. That sentence does not state a hailstone size, a wind speed, or a ground-snow load. Those come from the module certificate and from the local structural code, applied by someone allowed to apply them. Do not copy a wind speed from a coastal project onto an inland roof. Do not invent a hail diameter because a brochure says “hail resistant.”

What you can check without inventing a number is whether the submittal contains a mechanical load, a mount engineering letter for this roof or this soil, and a module model that matches the letter. A datasheet that lists electrical values and is silent on load is unfinished. Xcel’s 2023 electrical requirements show how long the electrical list already is: power at standard test conditions, temperature coefficients, maximum system voltage. The structural list is additional. Both have to be in the package. If the roof structure cannot carry the combination of module, rack, wind, and snow, the right answer is a different mount, not a tighter clamp spacing someone guessed. Mount types and access are in roof-mounted versus ground-mounted solar.

Hail that cracks glass is a damage event, not an energy-model input. A cracked module can still be live in daylight. PNNL is explicit that modules produce whether or not you intended them to. Do not walk the array to count cracks. Do not hose a hot, cracked module. Photograph from a safe place, shut down by the posted method if a qualified person is doing it, and treat the module as damaged equipment. A warranty may or may not cover impact. Many contracts exclude hail above a stated size or shift the cost of labor onto the owner. Read the exclusion before you describe the storm as a warranty event. How those contracts are built is in what a solar panel warranty covers.

What to do with the electrical system in severe weather

The shutdown order does not change because the weather is bad. AC disconnect, then DC disconnect, by a person trained to use them, with the disconnects where the design put them in sight of the inverter. Monitoring that shows zero is not the disconnect. An inverter that has faulted on grid voltage during a storm may be doing what IEEE 1547-2018 requires: stop energizing an island within 2 seconds. It should come back when the grid is stable and the manual’s reconnect rules are met. Resetting it during the lightning is not the recovery procedure.

Surge protection is a design, not a reaction in the storm. A surge protective device limits surge voltage and diverts surge current. Adding modules to a building does not, by itself, change the lightning-strike risk, which is the point made in the IEC technical report on lightning and surge protection for photovoltaic systems. The protection has to be designed with the array, not screwed on afterward in the rain. That design conversation is in lightning protection for a solar array. It does not include a do-it-yourself ground rod layout.

After the weather, the inspection is visual and from a safe position: loose modules, a rack that moved, a combiner cover open, a ground wire pulled off, vegetation or debris on the glass. Soiling from a dust storm is an operations loss and a wash decision, not a structural failure. Washing methods that stay off the glass are in how to clean solar panels. If something electrical is open or burned, stop. The repair is not a homeowner reset.

Frequently asked questions

Will solar panels survive hail?

Some modules are tested for impact, and the certificate names the test. This article does not quote a diameter, because no such test report was used as a source. Ask for the certificate. A cracked module is damaged and still live. Do not walk on it to check.

Should I clear snow before the next storm?

The energy model is not a clearing order. Tilted modules are expected to shed by sliding. Clearing is a safety and warranty decision. An icy roof is a bad place to learn that.

Why did the inverter go offline in the storm?

Often because the grid voltage or frequency left the range it is required to tolerate, or because it detected an island and stopped, which is the correct grid-tied behavior. It is not, by itself, a failed array. Have it checked if it stays off after the utility is stable.

Sources and references

Snow-loss range and sliding: NREL snow model. Gutters, icicles, and snow paths: US Department of Energy, FEMP. Framing expected to resist snow, wind, hail, and dust, without a stated diameter: PSCo solar technical specifications. Live modules and the insolation band: Pacific Northwest National Laboratory. Islanding: IEEE 1547-2018 primer. Surge devices and the point that PV on a building does not by itself change strike risk: IEC TR 63227:2020, preview.