A solar panel cleaning robot in India typically pays back in 18 to 36 months on plants above 500 kW, where soiling losses run 15 to 30 percent in dry months. Below 100 kW, a manual crew with a water trolley is usually cheaper than importing a robot.
India is one of the hardest markets in the world to design a cleaning robot for, and most of the machines on the market were designed somewhere else. That mismatch is where real projects go wrong.
How much does soiling actually cost in India?
Published studies put soiling losses in northern and western India at 15 to 30 percent during pre-monsoon months, dropping to 5 to 10 percent after rain. On a 1 MW plant generating roughly 1.5 million units a year, a 20 percent soiling loss is 300,000 units. At 7 rupees a unit, that is around 2.1 million rupees a year of lost revenue.
That is the number a robot has to beat. A mid-tier imported system for 1 MW costs somewhere between 1.5 and 4 million rupees depending on configuration and whether it is a single unit or a tracked gantry system.
What breaks when you import a robot designed elsewhere?
Three things, repeatedly. Dust composition: desert dust in Rajasthan is fine and silica-heavy, which abrades pads fast and can scratch glass if a dry brush drags it. Water hardness: borewell water in Gujarat and Rajasthan often exceeds 500 mg per litre of dissolved solids, and drying that on a panel leaves scale that reduces transmittance permanently.
And spare parts. A robot needing a replacement brush every four months is fine if the brush ships in a week. If it ships in 10 weeks from overseas, your cleaning schedule collapses in month three. Ask for a local stock commitment in writing, with a named partner and a minimum quantity, before you sign.
Does a robot work without water?
Dry cleaning works on light dust and is popular because it avoids the water logistics and scale problem. It is also harsher on the panel surface. For heavy dust after a long dry spell, water or a semi-dry microfibre approach is safer, and it needs a supply you have to plan for.
On a 10 MW plant, water for cleaning can mean tankers. Figure out litres per square metre first, then whether a tanker route is realistic. Vendors quote water use per panel, which sounds small and becomes a logistics plan at scale.
What about locally made options in India?
There are credible domestic manufacturers and integrators now, particularly in Gujarat and Maharashtra. The trade is usually lower price and better service response against less mature engineering and shorter field history. For a plant where downtime is expensive, the local option with a two-day service visit often beats the imported option with a better spec sheet.
I would not buy on spec alone in this market. Ask for three site references above 500 kW, call them, and ask specifically how the units performed after the first monsoon and whether any brush or pad replacement lead time was a problem.
Is a roof-mounted skylight robot different from a panel robot?
Completely. Panel robots work on near-flat arrays in long lines. Skylight robots work on pitched glass with frames, and the geometry changes both grip and path planning. Our WINBOT mini 2 review is a useful example of the small-format end of that category, though it is a window machine, not a panel cleaner.
If your site has both an array and an atrium or factory skylight, they are two purchasing decisions, not one. Vendors that claim both usually do one of them badly. The review index splits them by application.
How do you size a system for a 1 MW plant?
Start with panel count and row spacing, not capacity. A 1 MW plant is roughly 1,800 to 2,000 modules across 30 to 40 rows. Tracked gantry systems span rows and cover in a few hours, single-unit robots cover one row at a time and need a crew to move them. The layout decides the system, the capacity does not.
What should you not assume?
How does the monsoon change the schedule?
It compresses it. In most of India, cleaning has to fit between the pre-monsoon dust season and the post-monsoon residue, which leaves a short intense window and then a long recovery period. A machine that needs a service call in that window is worse than useless, because the crew has already moved on.
Plan two cleaning cycles rather than a continuous schedule: one heavy pass at the end of the dry season, one lighter pass after the rains to clear the residue the rain brings down. Continuous daily cleaning on a small plant is a distraction, and it wears consumables for little gain.
What about rooftop security and theft?
A real concern on remote plants. A gantry system that sits on the array overnight is an asset someone can take. Ask the vendor what the fixed installation looks like, how it is anchored, and whether any part of it is removable by hand. Bolted frames and serial markings help with recovery; they do not prevent anything, but an insurer will ask.
How does the monsoon change the schedule?
It compresses it. In most of India, cleaning has to fit between the pre-monsoon dust season and the post-monsoon residue, which leaves a short intense window and then a long recovery period. A machine that needs a service call in that window is worse than useless, because the crew has already moved on.
Plan two cleaning cycles rather than a continuous schedule: one heavy pass at the end of the dry season, one lighter pass after the rains to clear the residue the rain brings down. Continuous daily cleaning on a small plant is a distraction, and it wears consumables for little gain.
What about rooftop security and theft?
A real concern on remote plants. A gantry system that sits on the array overnight is an asset someone can take. Ask the vendor what the fixed installation looks like, how it is anchored, and whether any part of it is removable by hand. Bolted frames and serial markings help with recovery; they do not prevent anything, but an insurer will ask.
- Soiling losses of 15 to 30 percent in dry months drive the payback case.
- Payback 18 to 36 months above 500 kW; below 100 kW manual usually wins.
- Borewell water above 500 mg per litre dissolved solids leaves scale.
- Insist on a local spare parts stock commitment with a named partner.
- Ask three sites above 500 kW about post-monsoon performance.