Soiling is the single largest O&M variable in solar system yield — yet it is often dismissed as a minor nuisance. In arid markets like the Middle East, Central Asia, and Northern Africa, soiling losses of 20–30% are documented at unmanaged utility-scale sites. Even a temperate-climate rooftop in a dusty industrial zone can lose 5–8% annually with quarterly cleaning.
What causes soiling and how it reduces yield
Soiling refers to any material deposited on a module's front glass that reduces light transmission. Sources include:
- Dust and sand: the dominant source at most desert and semi-arid sites
- Bird droppings: highly concentrated, cause severe local shading and hotspots — 1% area coverage can cause 10–25% string output loss due to bypass diode activation
- Industrial particulates: cement, carbon black, and chemical residues near factories are adhesive and resist rain cleaning
- Pollen: seasonal, tends to wash away with moderate rain
- Moss and algae: common in humid climates on low-tilt modules; requires mechanical cleaning
Soiling reduces yield through two mechanisms. First, it directly attenuates incident irradiance — a 10% reduction in light reaching the cells reduces output by approximately 10%. Second, non-uniform soiling (such as bird droppings or heavy soiling on one end of a string) causes mismatch losses within a string, activating bypass diodes and further amplifying the output reduction beyond the directly shaded area.
Soiling loss by climate and site type
| Climate / site | Typical annual soiling loss | Cleaning frequency needed |
|---|---|---|
| Humid temperate (UK, NW Europe) | 1–3% | 1–2×/year or rain-dependent |
| Mediterranean (South Europe, East Med) | 3–6% | 3–4×/year |
| Tropical (SE Asia, sub-Saharan Africa) | 2–5% (heavy rain season) / 10–15% (dry season) | Seasonal, 4–6×/year |
| Arid / semi-arid (MENA, Central Asia) | 15–30% or more without cleaning | Monthly or automated |
| Industrial rooftop (any climate) | Add 3–8% to climate baseline | Increase frequency by 2× |
How to measure your system's soiling loss
There are three practical approaches for commercial systems:
1. Reference module comparison
Keep one module permanently clean. Compare its irradiance-normalised output to the average of adjacent uncleaned modules. The difference is the soiling factor. This method is low-cost and continuous, but requires a dedicated clean module and a monitoring system with module-level data.
2. Before/after cleaning comparison
Record the performance ratio (PR) of the system immediately before and after a cleaning event under similar irradiance. The PR improvement multiplied by annual generation gives the annual soiling loss in kWh. This is the most practical method for existing systems without reference modules.
3. Monitoring-based anomaly detection
Most modern monitoring platforms (iSolarCloud for Sungrow, SolarEdge Monitoring, Enphase) flag strings with persistent output below expected based on irradiance. A string that shows gradually declining PR over weeks is almost always experiencing progressive soiling. A string that drops suddenly is more likely a fault.
Cleaning ROI calculation
Before specifying a cleaning programme, calculate the breakeven cleaning cost:
- Determine your soiling loss rate in % per month between cleanings
- Multiply monthly loss by system generation × electricity value (or feed-in tariff) to get monthly revenue loss in USD
- If cleaning cost < recoverable revenue, cleaning pays back
Example: A 500 kWp system generating 700,000 kWh/year sells electricity at $0.10/kWh. At 1.5% soiling loss per month (arid climate), the monthly revenue loss is 700,000 × 0.015 ÷ 12 × $0.10 = $87.50 lost per month. If cleaning costs $300 for the whole system, cleaning every 3 months costs $100/month and recovers $262.50/month in avoided losses — a clear win.
Cleaning methods
- Manual dry brushing: low cost, no water needed. Risk of micro-scratches on cheaper glass. Suitable for lightly soiled modules.
- Manual wet cleaning: deionised water and soft cloth or brush removes stubborn soiling. Most effective. Cost: $0.002–0.005/Wp per cleaning.
- Automated robotic cleaning: brushes traverse rows on rails, triggered by soiling sensors or schedule. Suitable for large ground-mount (≥ 5 MWp). CAPEX $0.01–0.02/Wp; payback 2–4 years in MENA/CIS.
- Drone-based inspection + targeted cleaning: IR drone identifies hotspot modules (likely bird droppings), crew cleans only affected rows. Efficient for large sites where whole-system cleaning is expensive.
Design choices that reduce soiling
Soiling loss is partly a design problem, not just an O&M problem:
- Tilt angle: modules at ≥ 15° shed dust with rainfall far better than near-flat installations. In arid markets, accept a small yield penalty from off-optimum tilt to gain self-cleaning.
- Anti-soiling glass coating: hydrophobic nano-coatings (available on premium modules from LONGi, JA Solar) reduce adhesion and rain-wash soiling effectively. Add-on cost $0.005–0.01/Wp, typical yield benefit 1–3% in medium-soiling sites.
- Row spacing: modules at the downwind edge of dense rows accumulate more dust. Slightly wider row spacing improves airflow and reduces edge soiling.
- Avoid low points: water pooling at module edges re-deposits dissolved dust as a stripe when it evaporates. Tilt frames to ensure full drainage.
Including soiling loss in yield models
When submitting a bankable energy assessment (PVsyst, Helioscope, SAM), soiling loss should be input as a monthly value based on site data, not a single annual average. Using an annual average underestimates peak-season soiling in arid markets. For sites without measured soiling data, the NREL global soiling database provides country-level defaults.