Soiling: How Cleaning Solar Panels Improves Efficiency

Soiling: Dirt Accumulation on PV Devices & Associated Performance Loss

The original source of the information below was prepared by SunPower Corporation with the goal of quantifying the impact of soiling on Photovoltaic (PV) arrays in arid climates. Soiling is the event in which solar panels become covered with a layer of dirt, especially in areas where rain doesn't fall for months at a time. The study was performed in Southern California. The conclusion for this study was that cleaning becomes economical at a energy value of about $0.25 / KWh. Cost of cleaning is more of an issue on commercial systems where the cleaning crew needs to be paid to perform the work. On residential systems, cost of cleaning is less of an issue when the home owner cleans the system. Residential systems on flat and low slope roofs, such as foam roofs, PV arrays are very accessible. Home owners can spray and squeegee and the system themselves. Sloped roofs that are too steep to walk on present more difficulty when it comes to cleaning. Ladders and squeegees and brushes attached to long poles are the tools of choice on these systems. Although PV systems on sloped roofs are less accessible, these systems hold significantly less dirt because of the steep angle. Thus, if you can't access your solar panels because of roof pitch your best bet may be to let the rain do most of the washing.

Expanding on Another Soiling Study

The phenomena of accumulation of dirt on solar panels (or photovoltaics), is called ‘soiling’; this blocks sunlight and prevents heat absorption. In other words, soiling can significantly reduce the performance of solar panels, reducing energy absorbed, and therefore reducing energy output. In 2006, Mitchell et al developed a theory to predict and quantify energy lost to soiling. This article expands on this previous work: A. Kimber, L. Mitchell, S. Nogradi, H. Wenger, “Proceedings of the 4th World Conference on Photovoltaic Energy Conversion”, Waikoloa, Hawaii, USA; May 7-12, 2006.

Below: This photo, taken in 2008, displays workers in china glue thin cell solar to a roof. This type of solar is advantageous because it requires no racks or roof penetrations which can void a roof warranty or compromise a roof. During extended periods without rain, the solar power generated by the solar film will be somewhat reduced unless cleaned. Once it rains again, or once the solar film is cleaned, power generation will be restored.

Soiling in Most Severe in Arid Climates

Soiling is more severe in climates where rainfall is infrequent or rare. Two examples of arid climates where PV systems are growing in popularity are the Southwestern USA and Southern Europe. Both these climates have infrequent rainfall, which leads to more severe soiling of PV systems. The increased severity of soiling degrades solar panel performance and on a macro level, has economic impacts as well. Studies, such as Mitchell et al (2006) attempt to quantify and measure the effects of soiling, to determine performance loss, which can then be translated to economic impact. The study of Mitchell et al (2006) also suggested optimal cleaning times for PV systems in arid climates to minimize performance loss.

Mitchell et al (2006) PV Soiling Study

The 2006, Mitchell et al study is based on data from fifty different power grid connected photovoltaic units in areas of the USA where rainfall is rare or infrequent. The 50 solar power systems included both flat, tilted, and tracking mounted units. Rainfall levels, and the frequency with which each system was cleaned, and power levels before and after soiling was tracked by the study. Systems were logically divided into sets, based on weather and location to determine soiling rates based on geographic location. The predictive model for power output loss due to soiling was based on raw data from two systems per area. Plot points were determined based on soiling levels and power output levels based on observational data, and a linear relationship was assumed between plot points. Based on the 2005 study data, performance loss due to soiling was determined to be about 0.0011 kWh/kWp/ per day without rainfall (in arid regions). Annual energy loss due to soiling varied between two and six percent based on the region and environment. Based on these observations, a theoretical model was developed which predicts energy loss due to soiling within a 3.5% degree of error. It was noted during this study that during seasons of consistent rainfall, energy output due to soiling was minimal.

Pictured below: A modern day solar power farm with wind turbines in the background. Cleaning these ground mounted solar panels is most likely much cheaper than cleaning roof-mounted panels. The upside for roof-mounted panels is more efficient use of available space.

SunPower Corporation Study

During 2006 to 2007 SunPower Corporation performed a study to further understand soiling, performance loss, and further enhance existing predictive models, using data from the work of Mitchell et al (2006) in addition to new data from PV systems in the Los Angeles area of Southern California.

Three Identical PV Systems Monitored

Three identical PV systems were installed and observed in an office park. The buildings for each of the three systems were all the same structurally. Here are the specifications of each of the three systems, which were installed all installed in 2003:

-Frameless 1,664 Siemens Sp-75 laminate panels were used
-Xantrex PV-100 AC-DC inverters connected the systems to the building electrical grid
-Systems were mounted on PowerGuard insulated roof system substrates
-Sunpower DAS (data acquisition system) was used to collect data
-Raw data collected includes: ambient temperature, wind speed, irradiance, and energy output levels
-Observations were made at 15 minute intervals and stored in an SQL database

The PV systems on each of these buildings were oriented such that wind patterns and sun exposure would be almost identical for each PV system. Raw data collected was validated against this assumption.

Methodology: Three Identical Systems

The study is based on the hypothesis that three identical solar PV systems with varying cleaning frequency (and in effect, varying levels of soiling) would produce correlated energy output levels. Every effort was made to use three completely identical systems to control the environment and to isolate and accurately measure the effects of soiling, while keeping other factors or statistical noise out of the data. Tests were done during a dry period in southern California with no rain. Data from previous studies indicated that energy output differences were negligible during two weeks immediately after rainfall. Hence, the study two weeks after rainfall in late May of 2006. Rainfall was not expected again until mid to late October, which would allow the study to observe significant levels of soiling. Substantial soiling was not expected until late June.

Cleaning Schedule

The total time from late May until mid October was split evenly into time periods based on the number of cleanings scheduled. All of the systems started out clean due to rain in late May (as noted above). As mentioned above, all three systems were identical and placed in identical circumstances on adjacent buildings. The only difference between the solar PV units was how often they were cleaned.

Unit A2: washed twice
Unit B1: washed once
Unit B3: never washed

A local window washing contractor cleaned the solar PV units using special brushes on long hollow poles. These poles called ‘tucker poles’ have water flowing through them to ensure a thorough and clean wash.

Results: Soiling Substantially Effects Performance

Based on data from passed studies, A2 is assumed to be the benchmark for small or negligible efficiency losses. The effectiveness of the cleanings was clear as power output increased with each cleaning. In fact, the power output of A2 was relatively constant due to the well timed cleanings. The effect of rain was also noticeable as power output increased for all systems after a rain on October 13, 2006. A2 did improve slightly after each cleaning, but this is difficult to quantify accurately, due to how minute these differences were.

The last cleaning was in October although the performance data was still recorded all the way through 2007. Performance during the Winter of December 2006 and into early 2007 dropped unusually low due to an unusually dry conditions (i.e., lack of rainfall). It can safely be deduced that energy output for all three systems was abnormally low (due to the extreme lack of rain during a longer than normal period of time). Rainfall patterns, notably, lack of rainfall, can explain variances in PV system performance from year to year. More studies and more raw data is needed to explore these conclusions further.

Cost vs Benefits of Scheduled Cleaning

The cost to clean the 100 kWp Solar PV Systems was $800 per cleaning. A2, the system that was washed two times produced an extra 8,000 kWh more energy than system B3 which was never washed. B2, which was washed only once, produced about 2,700 kWh more than B3. Unless the value of the energy produced is greater than the cost of cleaning, then the cleaning is not worthwhile. Any system installed under the California Solar Initiative would increase system revenue by $1,500 USD per 100 kWp capacity in Los Angeles, California. Also, European feed-in tariffs, are even more worthwhile, because a bi-annual cleaning would increase revenue by $3,000 per 100 kWp capacity.

Theoretical Model Comparison

The Mitchell et al model predicts annual soiling losses (in energy output) by 4.8% to 5.5%. This smaller, Southern California study measured an output loss of 5.1%, which validates the previously mentioned model. Conclusions found that a soiling rate vs rainfall model more accurately predicts system performance than a standard constant rate model.

Summary

Soiling is important to understand because it is one of the largest factors in system performance variances and because it is a controllable factor as well. Based on both studies soiling losses can be approximated around 5% in total reduced energy output. Based on the SunPower Southern California study, cleaning becomes economical at an energy value of about $0.25 per kWh. As demonstrated by the study performed by Mitchell et al (2006), efficiency losses due to soiling vary greatly by climate and geographic region. Because of varying climate conditions in different areas, optimal cleaning recommendations must also vary by region and clime to achieve the maximum economic benefit.

It appears marginal efficiency gains can be made for a solar power system by cleaning the panels a few times a year (about 5% efficiency is lost due to soiling, apparently much of this can be mitigated). However, this only appears to be an issue in arid climates.

So unless you live in a place where it doesn't rain at least once per season, you may benefit from cleaning your panels, whether by hiring a professional, or by doing it yourself.

Find a local Solar Power contractor.