The document summarizes research on the impact of soiling on PV module performance in various climates. Mini-module tests were conducted over 1-2 years in Tempe, AZ; Chennai, India; Ancona, Italy; Cologne, Germany; and Thuwal, Saudi Arabia. Results showed significant annual soiling losses of over 3% in Tempe and over 2% in Chennai due to dust accumulation between rainfall events. Frequent rainfall in Ancona and Cologne led to negligible losses. Periodic cleaning was found necessary to maintain performance in arid Thuwal. Soiling patterns varied substantially between climates and years.

1. Impact of Soiling on PV Module
Performance for Various Climates
4th PV Performance Modelling and Monitoring Workshop
22/23 October 2015, Cologne, Germany
Dr Werner Herrmann
TÜV Rheinland Energie und Umwelt GmbH
51101 Cologne, Germany
werner.herrmann@de.tuv.com
http://www.tuv.com/solarpower

2. Outline
 Introduction
 PV-KLIMA test locations
 Experimental approach
 Results:
 Angular impact on soiling loss
 Soiling patterns for different climates
 Impact of periodical cleaning
 Summary
26.10.20152 4th PV Performance Modelling and Monitoring Workshop
22/23 October 2015, Cologne, Germany

3. Introduction – PV power loss due to soiling
26.10.20153
 Dust deposition on the PV module surface is a complex phenomenon, which is
mainly influenced by the environmental/weather conditions, mounting principle
and glazing characteristics.
 Performance losses are site-specific and can strongly depend on O&M work.
Photo:TÜVRheinlandShanghai
4th PV Performance Modelling and Monitoring Workshop
22/23 October 2015, Cologne, Germany
Installation conditions
Non-uniform soiling,
glazing characteristics
Surrounding environment
Climatic impacts

4. PV-KLIMA Project
Comparative energy yield testing of PV modules
Location Country Operation since Köppen-Geiger climate classification
Ancona Italy 01 Nov 2013 Cfa (mediterranean)
Tempe Arizona/USA 15 Dec 2013 Bwh (hot dessert)
Chennai India 01 Feb 2014 Aw (tropical savanna, hot-humid/dry)
Cologne Germany 01 Mar 2014 Cfb (temperate)
Thuwal Saudi-Arabia 11 Mar 2015 Bwh (hot desert, sandstorm impact)
26.10.20154
ChennaiAnconaCologneTempe
Saudi
Arabia
4th PV Performance Modelling and Monitoring Workshop
22/23 October 2015, Cologne, Germany

5. PV-KLIMA Test Locations
:
26.10.20155
Tempe / Arizona Chennai / Southeast India
2)
http://www.chennai.climatemp
s.com/
Location TUV Rheinland PTL, LLC
Tempe, AZ 85282
TUV Rheinland (India) Pvt. Ltd.
Chennai 602117, India
Geographical position: 33.4°N / 111.9°W
358 m above sea level
13°N / 80°E
35 m above sea level
Inclination angle 33.5° 15°
Annual in-plane global
solar irradiation
2360 kWh/m² 1860 kWh/m²
Average annual rainfall 219 mm 1597 mm
Surrounding
environment
industrial area,
no vegetation
Rural environment,
Farm land
4th PV Performance Modelling and Monitoring Workshop
22/23 October 2015, Cologne, Germany

6. PV-KLIMA Test Locations
:
26.10.20156
Cologne / Germany Ancona / Italy
Location TUV Rheinland Group
51101 Cologne, Germany
Loccioni
60030 Angeli di Rosora, Italy
Geographical position: 50.6°N / 7.0°E
53 m above sea level
13°N / 80°E
35 m above sea level
Inclination angle 35° 35°
Annual in-plane global
solar irradiation
1195 kWh/m² 1556 kWh/m²
Average annual rainfall 774 mm 757 mm
Surrounding
environment
Urban environment,
Flat roof of 5 storey building
Hilly landscape with grassland
and forest, agricultural use
4th PV Performance Modelling and Monitoring Workshop
22/23 October 2015, Cologne, Germany

7. :
26.10.20157
Thuwal / Saudi-Arabia
Location KAUST, NEO
Thuwal, Saudi-Arabia
Geographical position: 22.3°N / 39.1°O
3 m above sea level
Inclination angle 25°
Annual in-plane global
solar irradiation
23861) kWh/m² 1) Extrapolation from 204 days
Average annual rainfall 70 mm
Surrounding
environment
University ground, coastal
environment
4th PV Performance Modelling and Monitoring Workshop
22/23 October 2015, Cologne, Germany
PV-KLIMA Test Locations

8. Experimental Approach
Measurement of soiling rate
26.10.20158
 Side-by-side irradiance measurement with two mini-modules:
 Standard PV glazing, center cell operated in short circuit (reference cell)
 Data recording interval: 30 seconds
 Soiling loss = Transmission loss  Lower effective irradiance reaching the cells
 Soiling loss factor (SLF) = Ratio of irradiances from “clean” and “dirty” cell
“Clean” mini-module “Dirty” mini-module
𝐺 𝐸𝐹𝐹 = 𝐼𝑆𝐶,𝑀𝐸𝐴𝑆/𝐼𝑆𝐶,𝑆𝑇𝐶 × 1 +  × (𝑇 𝑀𝑂𝐷 − 25°𝐶) × 1000 𝑊/𝑚²
𝑆𝐿𝐹 =
𝐺 𝐸𝐹𝐹,𝐷𝑖𝑟𝑡𝑦
𝐺 𝐸𝐹𝐹,𝐶𝑙𝑒𝑎𝑛
4th PV Performance Modelling and Monitoring Workshop
22/23 October 2015, Cologne, Germany

9. Results
Angular impacts on soiling loss
26.10.20159
 Dust accumulation on the glass surface changes its angular response
 Soiling loss depends on the angle of incidence (AoI) of direct sunlight
 Soiling loss due to AoI effects can be significant
 Tempe: -0.8% annual soiling loss is caused by angular transmission losses
in the range AoI >30°
W. Herrmann, M. Schweiger: Soiling and self-cleaning of PV modules under the weather conditions of two locations in
Arizona and South-East India, IEEE-PVSC42, New Orleans, USA, 2015
4th PV Performance Modelling and Monitoring Workshop
22/23 October 2015, Cologne, Germany

10. Results
Soiling pattern for Tempe/Arizona: Dec 2013 – Sep 2015
26.10.201510
Average daily
SLF decrease in
dry periods:
-0.08% to -0.22%
Annual soiling loss
(year 1):
-3.7% total loss
-0.8% angular loss
Annual soiling loss
(year 2*):
-1.4% total loss
 Single rainfall events  Effective cleaning: SLF recovery >99% observed
 Significant differences in soiling patterns for 1st and 2nd year  frequency of
rainfall, dust accumulation rate
4th PV Performance Modelling and Monitoring Workshop
22/23 October 2015, Cologne, Germany

11. Results
Soiling pattern for Chennai/India: Feb 2014 – Sep 2015
26.10.201511
Average daily
SLF decrease in
dry periods:
-0.12% to -0.39%
Annual soiling loss
(year 1):
-2.2% total loss
 3-months dry season  25% soiling loss observed
 Recovery in rainy season: SLF >99% in first year, SLF >98% in second year
 Significant differences in soiling patterns for 1st and 2nd year
4th PV Performance Modelling and Monitoring Workshop
22/23 October 2015, Cologne, Germany

12. Results
Soiling pattern for Ancona/Italy: Nov 2013 – Sep 2015
26.10.201512
Daily SLF decrease:
Not applicable
Annual soiling loss:
0,04% (Year 1)
0,02% (Year 2)
 High cleaning effectiveness due to rainfall during the whole year
 Comparable soiling patterns for 1st and 2nd year  ±1% SLF variation
4th PV Performance Modelling and Monitoring Workshop
22/23 October 2015, Cologne, Germany

13. Results
Soiling pattern for Cologne/Germany: Mar 2014 – Sep 2015
26.10.201513
Daily SLF decrease:
Not applicable
Annual soiling loss:
-0,16% (Year 1)
+0.17% (Year 2*)
 High cleaning effectiveness due to rainfall during the whole year
 Comparable soiling patterns for 1st and 2nd year  ±2% SLF variation
4th PV Performance Modelling and Monitoring Workshop
22/23 October 2015, Cologne, Germany

14. Results
Soiling pattern for Thuwal/Saudi-Arabia: Mar 2015 – Sep 2015
26.10.201514
 Corresponding wet cleaning of “dirty” mini-module  Reset SLF = 1
Experimental approach
for Thuwal test site:
Periodical mechanical
cleaning of the PV module
surface with rotating
brushes (Type NOMADD)
4th PV Performance Modelling and Monitoring Workshop
22/23 October 2015, Cologne, Germany
 PV installations in desert climate (i.e. MENA region) require periodical
cleaning of PV module glass surface to optimize the energy yield

15. Results
Soiling pattern for Thuwal/Saudi-Arabia: Mar 2015 – Sep 2015
26.10.201515
 High ambient dust concentration  Average daily percent decrease of SLF = -0.5%
 Dust storm  Max. SLF change per day = -7.7%
4th PV Performance Modelling and Monitoring Workshop
22/23 October 2015, Cologne, Germany

16. Summary 1/3
Measuring results
26.10.201516
Average
annual
rainfall1)
Annual transmission
loss due to soiling
(Year 1)
Average SLF
decrease
between rainfall
events
Max. daily
SLF
decrease
Tempe /
Arizona
219 mm Year 1: -3.7%
Year 2: -1.4% 2)
-0,15% per day -1.8%
Chennai /
India
1597 mm Year 1: -2.1%
Year 2: -7.5% 2)
-0,2% per day -2.2%
Ancona /
Italy
757 mm negligible N/A -0.4%
Cologne /
Germany
774 mm negligible N/A -1.1%
Thuwal /
Saudi-Arabia
70 mm N/A
Periodical cleaning
-0.5% per day -7.7%
4th PV Performance Modelling and Monitoring Workshop
22/23 October 2015, Cologne, Germany
1) http://de.climate-data.org
2) Measuring data of year incomplete

17. Summary 2/3
Principle effects
17 26.10.2015
 Self-cleaning effect:
Rainfall leads to nearly full recovery of PV module performance.
Permanent soiling can occur if humidity condensate sticks dust to the surface
or can be caused by algae growth.
 Soiling pattern:
Great variation in daily percent decrease of soiling loss factor (SLF) makes
soiling loss prediction difficult in tropical and arid climates
 Broader data basis required
4th PV Performance Modelling and Monitoring Workshop
22/23 October 2015, Cologne, Germany

18. Summary 3/3
Impact of soiling for various climates
18 26.10.2015
 Hot-dry climate:
Tempe: Moderate dust settlement between periodic rainfall events leads to
annual performance loss <4%
Thuwal: Considerable dust settlement and missing rainfall lead to substantial
annual performance loss >50%, which requires periodical cleaning for
economic operation of a PV power plant.
 Hot-humid/dry climate:
Chennai: Considerable dust settlement during 3-months dry season can lead to
substantial soling loss. Cleaning during dry season must be considered to
improve the performance of a PV power plant.
 Temperate and Mediterranean climate:
Ancona, Cologne: No measureable performance loss due to frequent rainfall.
4th PV Performance Modelling and Monitoring Workshop
22/23 October 2015, Cologne, Germany

19. Thank you for your attention!
werner.herrmann@de.tuv.com
http://www.tuv.com/solarpower
Work was partially funded by the German Federal Ministry for Economic
Affairs and Energy (BMWi) under contract No. 0325517B.