TASK — 13

Soiling Losses – Impact on the Performance of Photovoltaic Power Plants

On a global scale, the soiling of solar photovoltaic (PV) systems from dust and snow, and subsequent loss of energy yield, is the single most influential factor impacting system yield after irradiance. Especially in arid regions, soiling may affect large utility-scale PV plants to a significant extent – making it necessary to mitigate these effects by cleaning whole systems – and thus leading to a reduction of revenues, caused by higher operating and/or capital expenditures (e.g., for investments in anti-soiling coatings [ASC] or cleaning robots and their maintenance).

This report therefore summarizes aspects of soiling from different perspectives including particle types and global distributions (Chapter 1), mechanisms and contributing factors (Chapter 2), sensors and measurement techniques (Chapter 3), modelling approaches and results (Chapter 4), economic impacts (Chapter 5), mitigation strategies (Chapter 6), and special installation and operation considerations for snow shading as solar arrays increasingly proliferate into higher latitudes (Chapter 7). The report is intended to serve the communities of PV customers, PV industry, O&M companies, investors, asset managers, testing equipment developers, testing companies, standardization authorities and research institutions alike.

The main authors have identified the following 3 Key Takeaways from the report:

– After irradiance, soiling is the single most influential factor impacting solar photovoltaic (PV) system yield and is estimated to cause a loss of annual PV energy production of 3-5%, corresponding to an economic loss on the order of 3-5 billion euros from higher operating, cleaning, and/or capital expenditures.
– Soiling is a factor not only in equatorial regions around Africa and Asia with high atmospheric suspended particle densities but also in high latitude regions due to snow cover as installations proliferate in these locations.
– PV soiling is highly heterogeneous, both at the module and plant level, and requires multi-sensor networks to accurately assess soiling rates and cleaning decision timelines.  Modelling approaches are still being refined and require additional data for validation.