Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Master → Norvege → Cleaning Schedule based on Soiling Effects on Photovoltaics in Kalkbult, South Africa

Norwegian University of Life Sciences, Ås (2017)

Cleaning Schedule based on Soiling Effects on Photovoltaics in Kalkbult, South Africa

Solend, Tor Atle

Titre : Cleaning Schedule based on Soiling Effects on Photovoltaics in Kalkbult, South Africa

Auteur : Solend, Tor Atle

Université de soutenance : Norwegian University of Life Sciences, Ås

Grade : Master Thesis 2017

Scatec Solar’s solar photovoltaic power plant in Kalkbult, South Africa, is in an area with excellent properties for harvesting solar energy. Vast areas with little or no conflict with agriculture and high irradiance. This semi-arid area has low precipitation and dust and soiling can be a problem when accumulating on solar module surfaces. The purpose of this thesis is to investigate the power loss from soiling on the solar panels, and to create a model to investigate if, and when, the modules should be cleaned. The soiling analysis will be conducted by analyzing data from a test facility inside the Kalkbult solar plant. The test facility consists of 16 crystalline Silicon modules, and 8 Cadmium-Telluride thin film modules. Current-Voltage data from these modules is used to calculate losses due to soiling by comparing uncleaned reference modules with clean modules. The results from the soiling analysis will be used to calculate a cleaning schedule. In addition, an experimental analysis of a dust sample from the area is analyzed. The period considered in this thesis is November 2016 to April 2017. The soiling analysis showed no soiling losses of the silicon modules, and just 1% power loss for the thin film modules in early November 2016. This resulted in no cleaning action in this period, as the marginal cost of cleaning per module was much higher than the marginal cost of soiling from the thin film panels. There was quite frequent rainfall in the period, and the cleaning effect of the rain seemed to eliminate power loss from soiling. Data from May to October 2016 was available, and in July there was enough soiling to trigger the cleaning schedule. However, the model assumes that all soiling is eliminated for a whole month, and the power plant is cleaned over a period of 47 days, there is reason to believe that soiling will occur simultaneously as the cleaning, and the economically viability of cleaning at all is highly questionable. A dust sample was analyzed in a scanning electron microscope to determine the composition and size distribution of the local dust in Kalkbult. The analysis showed that the dust was mostly organic, accompanied by salts and silica. The size distribution showed the average size of the dust particles to be 47µm in diameter, the most frequent had a diameter of 6.5µm and the median 11.8µm.


Version intégrale (4,77 Mb)

Page publiée le 22 février 2018, mise à jour le 7 octobre 2019