When Does Washing Your Panels Actually Pay for Itself?

Soiling is one of the most widespread losses in solar. A study published in Joule in 2019 estimated that, even with optimised cleaning, soiling reduced global solar power production by at least 3–4%, costing at least €3–5 billion a year in lost revenue. Yet the decision that controls most of that loss on any one site — when to send the cleaning crew — is usually made by habit. Clean too often and you pay for washes that recover little. Clean too rarely and you lose more to dirt than a wash would have cost. Both mistakes look the same on a maintenance schedule.

Who this matters to

O&M managers who set cleaning schedules and budgets, asset managers who approve them, and owners of plants in dry, dusty or agricultural regions, where soiling is a large and recurring line in the loss budget.

How soiling actually builds

Between cleaning events, dust, pollen, bird droppings and agricultural or industrial particles settle on the glass and steadily reduce the light reaching the cells. The rate depends on the site, the season, the module tilt and what is happening nearby — a harvest, a construction site or an unpaved road can change it in a week. Rain cleans, but not reliably. A heavy shower can wash an array almost back to clean. Light rain or morning dew can do the opposite: NREL researchers have described dust that “cements” onto panels after dew, becoming hard to remove even with a strong rain. Research on soiling in humid climates has found that rainfall needs to pass an intensity threshold to clean at all, and that the threshold varies by site.

The break-even rule

Every day after a clean, the soiling loss grows a little, and so does the revenue lost that day. The cumulative lost revenue since the last clean therefore grows faster and faster. A wash resets it to zero for the price of one cleaning. For soiling that builds steadily, the cleaning interval that minimises total cost — washes plus lost revenue — falls at the point where the revenue lost since the last clean equals the cost of one wash. Before that point, a wash costs more than it recovers. Well after it, every day of waiting costs more than the wash would have. Rain changes the arithmetic: if a cleaning rain is likely before the break-even point, the right decision is to wait.

Three cleaning strategies: an illustrative case

Numbers illustrative. A 50 MW plant in a dry season produces 250 MWh a day at $50 per MWh. Soiling builds by 0.15% a day, with no rain to help, and a full wash costs $25,000. On those assumptions, the revenue lost since the last clean reaches the cost of a wash after about 52 days.

Monthly and quarterly schedules fail in opposite directions — one overspends on washing, the other loses more to dirt — and land at almost the same cost. Cleaning on the economics saves $53,000–59,000 a year on this site before rain is considered at all. Skipping a wash that a heavy rain has made unnecessary saves more. The catch is that every input has to be real. The soiling rate must be measured at this site, this season. The tariff must be the one the plant actually earns. And the loss must be soiling, not an electrical fault or a drifting sensor that looks like dirt.

How Ellume Vector makes the call

• •Soiling measured, in kilowatt-hours. Vector measures dust and dirt accumulation as it happens, rather than guessing it from a calendar, and turns it into lost kilowatt-hours so “the panels look dirty” becomes a number you can weigh against the cost of a clean.
• •Rain recognised as free cleaning. Vector recognises rain events and their cleaning effect. If rain washed the array for free, you should not be paying for a clean the following week. Snow and sand-storm checks are applied only in climates where they are possible.
• •When a wash will pay for itself. Vector weighs soiling loss against cleaning cost, so cleaning is scheduled on economics rather than habit. Every loss carries a money figure at your tariff, and if the tariff is assumed rather than configured, Vector says so.
• •Dirt kept apart from everything else. Soiling is separated from passing cloud, high cell temperature and electrical faults, so the cleaning budget is not spent on a problem a wash cannot fix.

A check you can run this season

Look at production for a few clear days immediately before and after your last wash, adjusted for irradiance. The step up is what that wash recovered. Multiply it by your tariff and the days until the next wash or heavy rain, and compare it with what the wash cost. If the recovered energy is small, you are washing too often. If the step is large, you waited too long.

Frequently asked questions

How do I know when to clean solar panels?

When the revenue lost to soiling since the last clean has grown to roughly the cost of a wash, and no cleaning rain is likely before the loss grows further. That requires a measured soiling loss and a real tariff, not a fixed calendar.

Does rain clean solar panels?

Heavy rain can clean an array substantially. Light rain and dew often do not, and can even make dust harder to remove. The amount of rain needed varies from site to site.

How can I tell soiling from an electrical fault?

Soiling losses build gradually across the site and recover after rain or a wash. Electrical faults tend to affect specific strings or inverters and do not recover after cleaning. Our post on soiling versus electrical faults covers the diagnostics in detail.

Is a fixed cleaning schedule ever right?

It can be a reasonable starting point where soiling is steady and predictable. It should still be checked against measured soiling and rainfall, because a site's soiling rate changes with the season and with what happens around it.

The honest close

Sources & Further Reading