How to prioritize which solar tables to clean first

Most sites still clean on a fixed rotation. Section A this month, Section B next month, whatever the calendar says, regardless of what the panels are doing out there. That works fine until a dust storm hits one access road and leaves the rest of the site untouched, or a bird decides one row of trackers is a good perch. Then you're washing tables that didn't need it and leaving the dirty ones for another six weeks.

Prioritizing by table instead of by schedule means you need a way to measure soiling loss row by row, not just estimate it from a weather station and a shrug.

Soiling loss by table varies more than you think

Soiling is not a site-wide number. A combiner box report might show a 4% output loss for a block, but that's an average hiding a lot of range underneath it. The table closest to the gravel haul road could be running 12% down while the table two rows over, shaded from prevailing wind, is barely touched. Low-tilt arrays catch and hold dust differently than steep-tilt. Tables under flight paths for birds pick up streaking that a table in the middle of the array never sees.

If you're cleaning on a fixed rotation, you're treating all of that as the same problem. It isn't. A crew that washes the whole block because the average looks bad is spending half its labor on tables that were already fine.

Rank tables by measured loss, not by their position on a map or their turn in the rotation. That means a flyover that can separate "this table is dirty" from "this table has a hot cell or a dead string," because those two problems get fixed by completely different crews. A wash crew can't do anything about a bypass diode failure, and sending a string repair tech out for a dust problem wastes a truck roll.

Turning a flyover into a ranked list

This is where pairing RGB with thermal actually earns its keep instead of just producing a prettier image. The RGB pass shows you the visible soiling pattern: streaking, bird droppings, dust accumulation along a specific row. The thermal pass shows you hot cells and string outages that look like nothing from the visible side. Overlay the two and you can tell the difference between a table that needs a wash and a table that needs an electrician, instead of guessing from a heat map and hoping you picked right.

Once you've got both layers for the whole site, the useful output isn't a report full of thumbnails. It's a ranked list: which tables are losing the most output right now, sorted so the crew knows where to start Monday morning instead of walking the rows with a handheld thermal camera to find out for themselves. That's the difference between a findings dossier and a stack of images somebody has to interpret.

A few things worth checking before you build your own priority list, flyover or not:

• Confirm the loss is soiling, not hardware. A dirty-looking table with a hot cell underneath needs a repair ticket, not a wash.
• Weight by string overlap. If three tables feed the same combiner, the worst of the three often explains most of the block-level loss.
• Don't ignore the clean-looking tables. Soiling isn't always visible. A light, even film can cut output without leaving a visible streak.
• Re-check after a cleaning pass. A table ranked worst in spring might not be the worst one in fall, especially if the access road or the planting cycle nearby changed.

None of this requires walking every row with a camera to build the list. It requires a method that measures the whole site in one pass and sorts the findings by how much output each table is losing. If your current process is a rotation schedule and a hunch, a paired RGB and thermal flyover delivered as a ranked findings dossier gives the crew an order to work from instead of a guess.

Next time the wash schedule comes up, ask which tables are losing the most output, not which ones are due.