CSP mirror soiling characterization and modeling

• Autor(a) principal: Conceição, Ricardo
• Data de Publicação: 2018
• Outros Autores: Silva, Hugo, Collares-Pereira, Manuel
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10174/24565; https://doi.org/10.1016/j.solmat.2018.05.035
• topic: Solar energy; CSP; Soiling; Modeling

Soiling stands as a major problem for solar energy conversion technologies, causing unwanted transmittance, reflectance and absorbance losses. In this paper, a TraCS (Tracking Cleanliness Sensor) is used to quantify soiling effect in a flat mirror and to calculate soiling rates between periods without rain. Environmental parameters such as vertical wind speed, air temperature, relative humidity and particulate matter in the atmosphere are used as predictors to model soiling. Relations and trends between input and output are analyzed using a simple linear regression model and also through an interaction model. Further investigation is performed with a neural network approach to assess its viability for this type of problem and also for comparison with the previous models.