Performance of PV modules in arid areas - effects of soiling and module technology on yearly yield

Solar power is a rapidly growing source of electricity. The use of solar power is now spreading throughout the world, also into sunny, but warm, sandy and dusty areas. A solar park consists of individual solar modules, which will stand in harsh outdoor environments and transform solar energy into electricity for at least 25 years, possibly much, much longer. The modules consist of individual solar cells, connected in series and protected behind the module glass. The modules will unavoidably collect some amount of dust and soil from the environment, which will block some of the solar irradiation. For a solar module this means that less electricity will be produced. It has therefore been assumed that soiling of solar modules will give significant efficiency losses if the dust is not removed, particularly in areas with a dry climate and little rain. Which solution is best suited to minimize the efficiency losses due to soiling depends on both the type and amount of dust, the solar cell technology deployed, the weather, and access to cleaning water. In addition, the cost of the cleaning solutions is critical. There is a need for increased knowledge, and great possibilities for new innovative products specially adapted to solar plants around the globe. Through this project we have build a field test station tailored to study the effect of soiling on the solar modules and how the loss in power production can be minimized. The field station includes a well equipped weather station with measurements of diffuse and direct global irradiation, rain, ambient temperature, module temperature and wind. With tailor made equipment we are also able to measure the full current-voltage characteristics of the modules. Both crystalline silicon modules from ReneSola, and thin film Cadmium Telluride modules from First Solar are installed at the field station. In addition, the effect module tracking has on soiling is currently being monitored. All weather and production data from the field station is transferred wireless in order for the researchers both in Stellenbosch in South Africa and in Norway to have continuous access to the data. Analysis of the production- and weather data are carried out by the two research partners in the project: IFE and Stellenbosch University. Dust samples collected at the field station have been physically and chemically characterized at IFE, and optical measurements of glass samples with soiling have been carried out at IFE. These measurements will be used to expand existing optical models to account for soiling losses at all angles of incidence. Detailed loss analysis have been carried out to quantify the effect of soiling on the PV performance throughout the year. Comparative analysis of the different PV technologies and different cleaning routines show that the impact of soiling at this semi-arid site is very small. Soiling is only detectable during the winter months when there is little rain. Even for these months the losses in production due to soiling are less than 2%. It will very seldom be cost effective to clean PV parks with this level of soiling. Machine learning is a useful tool to study the correlation between different weather parameters and soiling. Rain results in natural cleaning of the modules and is hence the most important weather parameter for predicting soiling losses. However, also wind and humidity appears to correlate with the amount of soiling on PV modules.

This project aims to develop methodology that allows prediction and optimization of the energy performance of photovoltaic systems in arid areas. The project will be based around measurements from a field test station that will be created in South Africa in the vicinity of Scatec Solar's Kalkbult plant. The field test station will monitor power output and module soiling for various PV module technologies, and supply data and dust samples to Stellenbosch University and to IFE. The field test station will be the platform for investigation of various cleaning products and procedures with regards to cleaning efficiency, time and water consumption, safety and reliability and long-term degradation. Moreover, investigation of tracking movements to reduce soilin g effects will be investigated here. Dust samples collected at the field station will be physically and chemically characterized at IFE. Moreover, the spectral and angular transmission properties of clean and dirty samples will be measured at the IFE soll ab. These measurements will be used to expand existing optical models to account for soiling losses at all angles of incidence, and will be used as input in annual yield calculations. The aim is to develop a model to improve design and operation of PV par ks that includes the effect of soiling and cleaning requirements. This model together with field measurements will form the basis of a methodology for prediction of soiling losses at new sites. The project includes exchange of personnel and building of in frastructure in South Africa. A workshop with international partners is planned in 2016 with emphasis on future expansion of the competence and infrastructure developed in this project.