Field Notes
5 min read

Soiling Detection for Utility-Scale Solar

How drone-based soiling detection maps quiet, site-wide production loss across utility-scale solar arrays and turns wash timing into a data-driven decision.

On a utility-scale solar site, the losses that hurt you most are rarely dramatic. A blown fuse or a dead string shows up in your monitoring data and gets a truck rolled. Soiling does the opposite — it bleeds production quietly, evenly, across thousands of modules, and your SCADA system reads it as a slightly off day. By the time anyone notices the trend, you have already given away megawatt-hours you will never bill for. Drone-based soiling detection exists to make that invisible loss visible, map it across the array, and tell you whether a wash is worth paying for yet.

What Soiling Actually Costs a Utility-Scale Array

Soiling is the accumulation of dust, pollen, agricultural drift, bird droppings, and mineral residue on the glass surface of a module. It is not a defect. It is physics — anything sitting on the glass scatters or blocks light before it reaches the cell. On a clean desert site a few tenths of a percent per day adds up fast, and in agricultural or industrial corridors the rate climbs higher.

The problem with soiling is that it is uniform enough to hide. A single dirty module stands out. An entire array that is two percent down looks identical to an array having a slightly cloudy week. Most operators discover their real soiling rate only when a rain event resets the system and production jumps — at which point you know exactly how much you were losing, after the fact. Recurring drone inspection turns that guesswork into a measured number you can act on before the loss compounds.

How Drones Map Soiling Across the Whole Site

A drone carrying a calibrated thermal and RGB payload flies the array at a fixed altitude and overlap, capturing every module under consistent lighting. Soiling shows up two ways. In the visual imagery you see the pattern directly — the dust band along a row edge, the streaking below a perch point, the heavier accumulation on the downwind side of the site. In the thermal imagery, heavily soiled cells run hotter because the obstructed light becomes heat instead of current, so the worst patches glow even when the dirt is too thin to read by eye.

The value is not any single image. It is the map. A ground crew with an irradiance meter can spot-check a handful of modules and extrapolate. A drone documents the actual distribution — which blocks are clean, which inverters are dragging, whether the soiling is site-wide or concentrated near a dirt road, a feedlot, or a quarry. That spatial picture is what tells you whether you have a washing problem or a localized one.

Turning Soiling Data Into a Wash Decision

The whole point of measuring soiling is to answer one question: is it cheaper to clean the array now, or to keep absorbing the loss? Washing utility-scale solar is not free — water, labor, equipment, and the production you lose during the wash all carry cost. Cleaning too early wastes money. Cleaning too late wastes energy.

Recurring drone inspections give you the soiling rate, not just a snapshot, so you can find the break-even point. When the measured loss across the array exceeds the cost of a wash cycle, you clean — and you clean the blocks that actually need it rather than the whole site. Operators running this discipline stop washing on the calendar and start washing on the data, which usually means fewer wash cycles and higher net production at the same time.

Why Recurring Beats One-Off

A single soiling inspection tells you where you are today. It does not tell you how fast you are getting there. The operators who get the most out of aerial soiling work fly the same site on a regular cadence — often paired with the thermal defect inspection they are already paying for — so the soiling rate becomes a tracked metric alongside string-level faults and module degradation. Over a season you learn your site's real soiling curve, you can predict the next wash window before production sags, and you can defend O&M spend with numbers instead of intuition.

If you operate or maintain commercial or utility-scale solar and you are washing on a fixed schedule without knowing your actual soiling rate, that is worth a conversation. Corvus runs recurring thermal and RGB drone inspections built to surface exactly this kind of quiet, site-wide loss — and to tell you when a wash earns its cost. You can reach us at corvusrecon.io.