Dust is the loss an Indian plant can see from the fence. It is also the loss most often guessed. Many cleaning schedules are set by habit, not by the soiling rate measured at the site.
Quick answer. Solar soiling loss in India depends on the local soiling rate, dust type, and rain. A field study at IIT Gandhinagar measured 0.37% energy loss per day without cleaning (Valerino et al., 2020). The monsoon washes panels, so dry-season loss sets the cleaning plan. Measure the rate on site, then set the interval with the cost formula below.
TL;DR
- Gujarat field data: 0.37 ± 0.09% loss per day of dry weather, with fine particles causing most of it.
- Dust type matters. At the same dust mass, Mumbai dust cut cell output about twice as much as Jodhpur dust.
- Rain resets soiling. A Chennai study saw loss turn negative after heavy rain.
- Optimal cleaning interval = square root of (2 × cleaning cost / (tariff × daily energy × soiling rate)).
- Measure soiling with a clean and soiled reference pair, as IEC 61724-1 describes.
This guide is for asset owners, O&M managers, and EPCs planning cleaning in India. It covers field evidence and the cleaning decision. For entering soiling into PVsyst, use our separate guide to PVsyst soiling loss modeling.
What is solar soiling loss?
Soiling loss is the energy lost when dust, salt, pollen, soot, or bird droppings block light reaching the cells. It builds during dry spells and drops after rain or cleaning.
Two terms help:
- Soiling ratio (SR): output of a soiled device divided by its expected clean output. An SR of 0.97 means 3% loss.
- Soiling rate: how fast SR falls per dry day, in % per day.
Uneven soiling, such as droppings or dust lines at the bottom frame, can also cause hot spots. That is a reliability issue, not just an energy one.
What have Indian field studies measured?
Published Indian soiling data is thin. These are the studies we could verify:
| Study | Location | Method | Key result |
|---|---|---|---|
| Valerino et al., Solar Energy, 2020 | IIT Gandhinagar, Gujarat | Outdoor soiling station and low-cost microscope | 0.37 ± 0.09% loss per day |
| John et al., IEEE Journal of Photovoltaics, 2016 | Dust from 6 Indian sites, lab test | Dust deposited at fixed density | At 3 g/m2: Mumbai dust 17.1% loss, Jodhpur dust 9.8% |
| Chennai field study, RE Grid Integration India, 2019 | Chennai, Tamil Nadu | One-year field monitoring | Soiling factor from -1.36% to 3.67% |
Sources: Valerino et al. (2020), John et al. (2016), and the Chennai study paper.
What the studies tell a designer
- Fine particles do the damage. Valerino et al. found particles under 5 micrometres caused more than half the loss. Fine dust sticks and scatters light well.
- Dust mass is not the whole story. In the John et al. tests, Mumbai dust caused nearly twice the loss of Jodhpur dust at the same mass. Composition and particle size matter.
- Rain cleans, sometimes better than new. In Chennai, loss turned negative after heavy rain, so panels beat the reference. Loss peaked at 3.6% in May and June, before the rains.
A loss of 0.37% per day adds up fast. After 30 dry days without cleaning, that is about 11% if the rate stays linear.
How does soiling differ by region in India?
Without more field data, a regional guide should list drivers, not invent rates. Use this matrix to decide what to measure:
| Region | Main soiling drivers | Dry season to watch | What to measure first |
|---|---|---|---|
| Thar desert, Kutch | Mineral dust, dust storms | Pre-monsoon months | Daily soiling rate, storm events |
| Western coast | Salt, humidity, fine urban dust | Post-monsoon to May | Dust stickiness, cementation |
| Indo-Gangetic plain | Haze, smoke, road and farm dust | Winter and pre-monsoon | Fine particle loss, winter smog weeks |
| Deccan plateau | Dust, long dry season | October to May | Rate across the dry season |
| Industrial zones | Fly ash, cement, soot | All year | Deposit type near the stack or road |
| Any region | Bird droppings | All year | Spot soiling and hot spots |
This matrix is an engineering screen built from the drivers above. It is not measured data. Rates at your site can sit well above or below the Gandhinagar figure.
What changes the rate at one site?
- Tilt. Flat arrays shed dust and water poorly. Our tilt angle guide covers the yield cost of going flatter.
- Nearby sources. Unpaved roads, quarries, farms, and stacks raise local deposits.
- Module height and row position. Edge rows and low modules often collect more dust.
- Dew. Morning dew can glue dust to glass, so a dry brush removes less.
How do you measure soiling on site?
Use a soiling station: two reference devices side by side. One is cleaned daily, one is left to soil. The ratio of their outputs is the soiling ratio.
IEC 61724-1:2021 covers soiling measurement for performance monitoring. Our IEC 61724 monitoring guide explains the monitoring classes.
Three rules keep the data honest:
- Clean the reference device on a fixed schedule. A dirty “clean” device makes soiling look smaller.
- Log rain events and cleaning dates next to SR.
- Calculate the rate over dry periods only.
Soiling also shows up in performance ratio. A slow PR drop between cleanings is a soiling signature. See our PR benchmark guide for how to separate it from heat.
How often should you clean? The cost formula
Treat cleaning as a trade between cleaning cost and lost energy. If soiling builds linearly at rate r, the lost energy over an interval of T days is E × r × T² / 2. E is the plant’s clean daily energy.
Minimizing the total cost per day gives:
T optimal = √(2 × C / (p × E × r))
- C = cost of one full cleaning (₹)
- p = value of energy (₹/kWh)
- E = clean daily energy (kWh/day)
- r = soiling rate (fraction per day)
Worked example
These inputs are illustrative. Replace them with your own.
| Input | Case A | Case B |
|---|---|---|
| Plant size | 1 MWp | 1 MWp |
| Clean daily energy E | 4,500 kWh/day | 4,500 kWh/day |
| Energy value p | ₹4/kWh | ₹4/kWh |
| Soiling rate r | 0.37%/day | 0.10%/day |
| Cost per cleaning C | ₹3,000 | ₹3,000 |
| Optimal interval | about 9.5 days | about 18 days |
| Average soiling loss at that interval | about 1.8% | about 0.9% |
Doubling the cleaning cost to ₹6,000 stretches Case A to about 13 days. The interval moves with the square root of the inputs, so it is forgiving of small errors.
Limits of the formula
- Soiling is not always linear. A dust storm can do a week’s damage in a day.
- Rain resets the clock. In monsoon months, skip scheduled cleans after heavy rain.
- Water is a real cost. In dry districts, water supply can cap the interval.
- Cleaning method matters. Abrasive brushing can damage anti-reflective coating. Follow the module maker’s cleaning instructions to protect the warranty.
How soiling enters the energy model
The energy model should use monthly soiling values, not one annual number. Use dry-season rates and your planned cleaning interval to set them. The PVsyst method is covered in the soiling modeling guide linked at the top.
Soiling and irradiance interact. A high-diffuse, hazy site can lose a lot to fine dust. Our irradiance data guide for India shows the diffuse share by city.
FAQ
What is a typical soiling loss in India?
It varies too much by site for one number. The verified Gujarat field rate is 0.37% per day of dry weather. Annual loss then depends on rain and cleaning frequency.
Does the monsoon fully clean panels?
Heavy rain removes most loose dust. Sticky deposits, droppings, and bottom-edge dust lines can remain. Inspect after the first heavy rains.
Are robotic cleaners worth it in India?
Run the cost formula with the robot’s cost per cleaning and its frequency. Robots can make short intervals affordable on large, dusty plants.
Is soiling worse on bifacial modules?
The front side soils as usual. The rear side usually soils less, but still sees dust from below. Our bifacial module design guide covers rear-side gain assumptions. Measure both if the plant relies on rear-side gain.
Get a soiling assumption you can defend
A defensible soiling input combines site data, the right monthly profile, and a costed cleaning plan. Heaven Designs prepares PVsyst yield reports with monthly soiling assumptions and documented sources. For new sites, we model soiling during 3D pre-design.
To scope a yield study, request a project quote. We reply within 1 business day.