Ask whether cleaning solar panels increases their output, and the answer will always be yes: remove a layer that absorbs or scatters light, and more light reaches the cells. But that is not the right question. What matters is how much energy the dirt would have cost you before the next natural cleaning, and whether recovering that energy is worth more than the cleaning itself. Put that way, the question calls for very different answers from one end of Europe to the other. On a steeply pitched roof, in a rainy oceanic climate, a routine cleaning may recover less energy than a single cloudy week loses. On a low-pitched roof, in the middle of a dry summer in the south, the same cleaning can pay for itself in a few weeks. Drone solar panel cleaning is no exception: the answer depends on the climate and the roof.
On the programme: the mechanisms at work, three climate zones rather than countries, and a payback formula to fill in with your own figures.
Solar panel soiling: how it loses energy
Transmission loss and soiling ratio
Soiling reduces the light that passes through the front glass, by absorbing it and scattering it away from the cells. The industry measures it by means of the soiling ratio: the output of a soiled module divided by the output it would deliver if it were clean. The IEC 61724-1 standard, which governs photovoltaic performance monitoring, defines how to measure it, generally by comparing a regularly cleaned module with a module left to soil naturally.
For light dust loads, transmission loss rises roughly in proportion to the mass of dust per square metre, then levels off as particles pile on top of one another. The nature of the particles also matters: fine particles scatter more light per gram than coarse sand, and dark soot absorbs more than light-coloured mineral dust.
The scale of the problem is documented. A widely cited techno-economic study, published by Ilse and colleagues (Joule, 2019), estimated that soiling was then costing around 3–4% of global photovoltaic output. Most of that loss occurs in arid regions, not in temperate Europe. The IEA PV Power Systems Programme (Task 13) has published detailed reviews of soiling and ways to limit it.
The sawtooth profile: build-up and reset
At a given site, soiling loss follows a sawtooth profile. It rises at a roughly constant rate during dry periods, then drops back each time rain washes the panels. Two parameters therefore govern the loss:
- the soiling rate, in percentage points of output lost per dry day, and
- the length of dry periods, that is, the time during which the panels go without rain heavy enough to clean them.
How much rain is enough remains uncertain: field studies report thresholds ranging from about 1 mm to more than 5 mm per event, depending on tilt and type of dirt.
For a constant soiling rate r and a dry period of T days, the average loss over that period is approximately r × T ÷ 2. This simple result shapes the whole cleaning decision: the energy lost over the whole period, about r × T² ÷ 2, grows with the square of the length of the dry period, since both the peak loss and its duration increase.
What rain does not remove
Some soiling does not go away with rain and, in humid climates, it is often the bulk of the residual loss:
- Dirt bands along the bottom of the module. On low-tilt framed modules, water collects against the lower edge of the frame, evaporates and leaves a band of concentrated dirt. This band shades the bottom row of cells.
- Cementation. Repeated cycles of dew and drying dissolve salts and organic matter, then redeposit them as an adherent film. By the sea, hygroscopic salt aerosols speed up this process.
- Bird droppings and plant debris. They cast a dense local shadow. In a series string, a heavily shaded cell forces its bypass diode to conduct and can cut the output of a whole group of cells, generally a third of a module. A small dropping therefore costs disproportionately. It also generates heat: persistent partial shading creates hot spots, which is why module qualification standards (IEC 61215) include a hot-spot endurance test.
- Biological growth. In damp and shaded spots, algae, mosses and lichens can establish themselves along frame edges and in the gaps between modules.
Drone cleaning of photovoltaic panels: how it works and how effective it is
Spraying without contact
Most drone photovoltaic cleaning relies on the same basic set-up as low-pressure facade cleaning: a multirotor supplied with water by a feed hose connected to a ground unit, spraying at low pressure from a distance of about half a metre to one and a half metres.
Contactless spraying cleans through hydrodynamic shear and dissolution. The jet lifts and carries away particles held by weak forces, such as van der Waals attraction or electrostatic charges. It removes loose mineral dust and pollen well. It struggles more with dust bound by cementation or held by capillary bridges after dew, and it is least effective on bird droppings and biological films. These are precisely the deposits that brushing or wiping removes more reliably.
Rotor downwash, of several metres per second on a loaded cleaning drone, helps dislodge dry dust but can also throw debris onto neighbouring modules: passes should therefore progress downwind across the module array.
Water quality: why deionised water matters
In much of central and southern Europe, tap water is hard, rich in dissolved calcium and magnesium. As it evaporates on glass, it leaves carbonate deposits, which are themselves a soiling layer, harder to remove than the original dust. That is why good practice, independent of manufacturers, calls for deionised or demineralised water, which dries without leaving residues (produced for example by ion-exchange resin filters such as Unger’s HydroPower). Our article on drone window cleaning explains how such water is produced.
Detergents are generally advised against, unless approved by the module manufacturer. Many modules carry a porous silica anti-reflective coating that adds a few percentage points of transmission, and strong acids, alkaline products or abrasives can degrade it permanently.
Pressure, temperature and compliance with the warranty
Module installation manuals set cleaning limits that vary from one manufacturer to another. The most common are:
- a maximum water pressure, generally well below that of a pressure washer,
- no abrasive tools or harsh chemicals, and
- no cleaning when the glass is hot.
Thermal shock is a very real mechanism: cold water on glass at 60 °C or more creates steep temperature gradients in the glass and at the frame seal. Cleaning early in the morning or late in the day avoids it and also limits streaking.
For the warranty, the practical requirement is documentation: consult the manual for the installed modules and record the type of water, the pressure at the nozzle, the method and the date.
Drone solar panel cleaning: three climate zones, three answers
The figures below are indicative ranges, based on published field studies and on operating experience; they are not site-specific measurements. Local sources such as farmland, unpaved tracks, industry or bird colonies can weigh more heavily than climate.
Oceanic north-west
This zone covers Ireland, Great Britain, Benelux, Atlantic France, Denmark and coastal Scandinavia. It rains every few days in most months, and the typical dry period is short. On modules tilted at 20° or more, the sawtooth resets so often that the average loss from ordinary dust is generally small, often estimated at around 1–2% of annual expected yield, or even less. Treat this figure as approximate.
The remaining losses come from dirt bands along the bottom of modules, bird droppings and algae or lichens on shaded edges. These are exactly the types of soiling that rain does not remove and that drone spraying removes least reliably.
Frequency: 0–1 cleaning per year, triggered by an inspection or by monitoring data (in French) rather than by the calendar. Low-pitched commercial roofs, as well as sites near farms or seabird colonies, sit at the top of the range.
Continental interior
This zone covers Germany, Austria, Czechia, Poland, Hungary, the Swiss Plateau and northern Italy. Rain is regular but less frequent than in the oceanic zone, with marked dry periods in late summer and sometimes in spring. Spring pollen and agricultural dust build up quickly and, in the Po Valley as in industrial areas, particulates add a soot component.
Frequency: 0–1 per year on pitched residential installations and 1–2 on low-pitched flat roofs, ideally after the pollen season or during a long summer dry spell.
Hot, dry south
This zone covers most of the Iberian Peninsula, southern Italy, Greece, Cyprus, Malta and Mediterranean France. In many places, summer dry periods last 60–120 days. Soiling rates recorded during the dry months are often around 0.1–0.3% per day, and more near dust sources. Saharan dust events can deposit a heavy load in one go, and a light shower that follows leaves mud instead of washing the panels.
Here, the sawtooth arithmetic clearly argues in favour of cleaning. With r = 0.2% per day and a 90-day dry season, the average loss without cleaning reaches about 9%. A mid-season cleaning cuts it into two 45-day sawtooths and halves the average loss, to about 4.5%. That recovers nearly 4.5% of the season’s output, and summer output makes up the largest share of the year.
Frequency: 1–3 per year, scheduled during the dry season, plus one after any major dust event. Very dusty sites may justify more.
Cost of drone solar panel cleaning, relative to the reference unit
The reference unit corresponds to a standard drone cleaning pass on a reference house, roof and facades included. Photovoltaic cleaning is a lighter job, with no chemicals, a lower flow rate and a smaller area, but the fixed costs of mobilisation, site survey and flight preparation remain the same. Indicative multiples:
| Installation size | Typical setting | Cost per cleaning (× reference unit) |
|---|---|---|
| 10–30 modules | Residential roof | 0.3–0.5× |
| 30–100 modules | Small commercial or agricultural roof | 0.5–1.0× |
| 100–500 modules | Commercial flat roof | 1.0–2.5× |
| Per additional 100 modules beyond 500 | Large roof | +0.2–0.4× |
It is the trend that matters, more than the exact figures. Fixed costs dominate on small installations: a residential installation therefore pays much more per module than a large roof. In built-up areas, an operational authorisation issued by the national civil aviation authority (the DGAC, France’s civil aviation authority) is generally required, a burden that weighs much more heavily on a 12-module job than on a 400-module one.
Payback of photovoltaic cleaning: the return-on-investment formula
Use your own figures. Let:
- E = the expected output of your installation over the period covered by the cleaning, in kWh. This is generally the dry period until the next reliable rain, or until the next scheduled cleaning. Your monitoring data or a yield estimate for that period will give you this value.
- ΔL = the average loss avoided over that period, expressed as a fraction. For a cleaning in the middle of a dry period under a sawtooth regime, ΔL ≈ r × T ÷ 4, T being the total length of the dry period in days. This is the difference between the average without cleaning (r × T ÷ 2) and the average of two half-length sawtooths.
- P = what your electricity is worth to you, self-consumed or exported, per kWh.
- C = the cost of one cleaning.
Energy recovered: E × ΔL. Value recovered: E × ΔL × P.
A cleaning pays off when E × ΔL × P > C. Rearranging, the electricity value at which cleaning breaks even is P* = C ÷ (E × ΔL). If your electricity is worth more than P* to you, clean. If it is worth less, don’t.
Two corrections are worth applying before relying on the result.
- Inverter clipping. Many systems are sized with a panel capacity greater than the inverter’s. At midsummer peaks, the inverter is already limiting output, so part of the soiling loss at those hours costs nothing. Installations with a high panel-to-inverter ratio recover less than the formula indicates.
- Type of soiling. If a large part of the loss comes from droppings, dirt bands or cemented films, contactless spraying may recover only part of ΔL. If your monitoring can show the soiling ratio after a cleaning (for example with a soiling sensor such as the Kipp & Zonen DustIQ), use that measured gain instead of the estimate.
Drone cleaning: limits and trade-offs
- Data first. Without monitoring, cleaning decisions are guesswork. Cleaning a reference module and watching the difference costs little.
- Contact or contactless. Drone spraying avoids walking on modules, which matters, because foot traffic can cause microcracks in cells, a failure mode documented in the IEA PVPS Task 13 reviews of module degradation. But it removes adherent soiling less reliably than soft brushes with deionised water.
- Downwash and debris. On roofs ballasted with loose gravel or fitted with lightweight fixings, rotor downwash creates its own risk. On tightly packed module arrays, it may simply move dirt from one module to another.
- Risk to the coating. Any method that brings an anti-reflective coating into contact with chemicals or abrasion can cause permanent damage. The failure risk comes from the operator’s choices, not the drone.
Should you have your solar panels cleaned by drone? The verdict by zone
In the hot, dry south, drone cleaning of low-tilt or large installations is justified by the physics: long dry periods, high soiling rates and summer output combine so that a well-timed cleaning recovers several percent of the seasonal expected yield. The payback formula will generally be favourable, especially on large roofs.
In the continental interior zone, the answer depends on tilt and local sources. Commercial flat roofs near farmland or roads often justify one well-timed cleaning per year. Steeply pitched residential installations rarely do, unless monitoring reveals a persistent loss.
In the oceanic north-west, rain does most of the work. The remaining losses are mostly of the kind that drone spraying removes least reliably, and small installations bear the highest fixed costs per module. Targeted cleaning, triggered by monitoring, sometimes by hand with deionised water and a soft brush, generally beats a scheduled drone cleaning programme.
Whatever the zone, the right question is whether the energy recovered during the next dry period, valued at what your electricity is worth to you, exceeds the cost of a cleaning.
