Development of a Hybrid Model for Soiling Analysis in Photovoltaic Systems Under Different Climatic Conditions Considering Rainfall and Turbidity Data
Abstract
Soiling affects the performance of photovoltaic (PV) systems, especially in arid regions with high concentrations of dust and particulate matter. This accumulation increases maintenance costs for cleaning and reduces the Performance Ratio (PR) of PV plants, making it necessary to understand the deposition behavior over time in order to mitigate the effects of soiling on system performance. In this context, this study presents a hybrid model for soiling analysis integrating rainfall and Linke turbidity data, replacing particulate matter (PM10 and PM2.5) data, which is more commonly used in current models in the literature. The developed model dynamically calculates an index of accumulated dirt over time, considering different levels of rainfall intensity and using atmospheric turbidity as a control factor for deposition speed. Data validation was performed using data from field weather stations located at the photovoltaic plants studied in the western region of the state of São Paulo and in the western region of the state of Bahia, comparing the behavior of the simulated curves with the reference cell curves of the plants. The results achieved by the developed model showed better performance for correlation and reduction of error metrics in both locations compared to the Kimber model, demonstrating the efficiency of this approach for regions with low atmospheric data coverage.
Keywords
Soiling Model, Photovoltaic Systems, Particulate Matter, Linke Turbidity, Atmospheric data