How often should solar panels be cleaned, and with what water?

There is no correct cleaning interval, only a correct way of working one out for a particular roof. Soiling sources by site type, how to set the interval from measurement, why groundwater is the wrong wash water, and the damage bad cleaning does.

There is no correct cleaning interval for a rooftop plant, only a correct way of working one out for a particular roof. A plant beside a brick kiln in Savar and a plant on a warehouse in a residential part of Dhaka do not have the same answer, and neither has the same answer in February that it has in July.

Soiling depends on what is upwind

• Brick kilns, cement works and unsurfaced roads. Heavy mineral dust that binds to overnight dew and does not wash off in light rain. The worst category here, and the one that most often justifies a short interval.
• Rice mills and boiler houses. Husk ash and fly ash, fine, alkaline and sticky once it gets wet.
• Jute and textile mills. Airborne lint, not opaque by itself, but it traps dust and builds a mat along the lower edge of every module.
• Poultry and feed plants. Fine organic dust plus ammonia, which attacks the anodising on the frames as well as cutting transmission through the glass.
• Coastal and riverside sites. Salt aerosol, corrosive as much as it is opaque, and it needs rinsing rather than scrubbing.
• Every site. Bird droppings, which are the real problem because they are opaque and localised. A dropping covering part of one cell forces that cell into reverse bias, creates a hot spot and puts the bypass diode to work. That damages the module as well as costing you the day's yield.

Set the interval from measurement, not from a calendar

The method that works is to keep one string clean as a reference and compare it with the rest of the array. The difference between the reference string's specific yield and the array's is your soiling loss, and you can watch it accumulate week by week. When the value of the energy being lost passes the cost of a cleaning cycle, clean. That is the whole calculation, and it gives a different answer for every site.

Two seasonal facts shape it here. From roughly November to March the air is dry and dusty and soiling accumulates without interruption, so this is when intervals are shortest. From June through September the monsoon does a large part of the work, and a plant needing a wash every few weeks in winter may need very little through the rains. The trap is light rain: enough to wet the dust and not enough to carry it away leaves streaks that are worse than the dry layer was. That is the one time an unscheduled clean is justified.

If your monitoring cannot show string-level yield, fix that before arguing about intervals. Washing to a fixed calendar on a site that does not need it is money spent scratching glass.

The water matters more than the brush

Most rooftop plants here are washed with whatever water is on site, which on a factory usually means deep tubewell water. Two properties of it cause trouble.

Hardness. Calcium and magnesium left behind as water evaporates form a scale film on the glass. The next wash does not remove it, it adds to it, and it scatters light in exactly the place you least want light scattered. On a hot roof it evaporates before anyone can squeegee it off, so the film forms faster than on a cool morning.

Iron. Groundwater across much of Bangladesh carries high dissolved iron. It oxidises on contact with air and leaves an orange-brown stain on glass and on anodised frames. Once set, it needs an acid clean to shift, which is not a job anyone wants on a roof full of live modules.

What to specify: filtered water, softened if hardness is high, with an iron filter if the source is a tubewell. Deionised water in a tank on a trolley is the cleanest answer and needs no detergent at all, because pure water is itself the solvent. Have the total dissolved solids of your wash water measured once. It costs almost nothing and it tells you whether you have a problem before the stain appears.

What bad cleaning does

• Cold water on hot glass. Washing at midday in May puts a thermal shock through the glass and the cells behind it. Clean in the early morning or after sunset, which also stops the water flashing off and leaving scale.
• Pressure washers. High pressure drives water past the frame seal into the laminate edge and finds the junction box. Moisture inside a laminate shows up as delamination years later. Use flow, not pressure.
• The wrong brush. Anything harder than soft bristle or a rubber blade will microscratch the anti-reflective coating, and it does not grow back. Grit trapped in a brush head does the same. Rinse the brush between rows.
• Walking on modules. The fastest way to put invisible microcracks into cells. They do not appear as a fault on the day, they appear as a string that quietly declines over two seasons. If the array has no walkway, that is a design fault to correct.
• Detergent left on the glass. Surfactant residue is a dust magnet, so the module gets dirty faster than before. If detergent is used at all it gets a full rinse.
• Cleaning a hot spot away without investigating it. A dropping baked onto a cell for weeks may already have damaged the module. Clean it, then put a thermal camera on that module.

Back to all Insights