Comparative Studies of the Effects of Soiling of PV Modules and Systems  in Tropical, Subtropical, and Semi-Arid Climate Zones in Brazil  
Suellen C.S. Costa1, Antonia Sonia A.C. Diniz1, Lawrence L. Kazmerski1,2, Cristiana Brasil Maia1, Cláudio Dias Campos1, Daniel Sena Braga1, Pedro P. Brito1, Vinicius Camatta1, Sergio de Morais Hanriot1
1Pontificia Universidade Católica de Minas Gerais (PUC Minas),, Belo Horizonte, Brazil
/2University of Colorado Boulder, RASEI, Boulder, CO, United States

Brasil’s principally southern-hemisphere geographic location represents interesting, complex, and comparative opportunities for soiling studies, encompassing sub-tropical, tropical, and temperate climate zones. This paper summarizes our methodology, with results quantifying the soiling ratios (SRatio) and the soiling rates (SRate) for these different photovoltaic (PV) technologies installed in diverse climate-zones in Brasil (based on our climate-zone mapping). Specifically, these analytic procedures are validated for soiling monitoring stations composed of both polycrystalline-Si and thin-film CdTe modules. Meteorological data, electrical, and thermal parameters were collected and analyzed. A mathematical model determined the SRatio and SRate for each site (in compliance with the IEC-61324-1 monitoring standard). The results specify: (1) a mean SRatio for the period of analysis for Belo Horizonte (-19.92°N, Lo-43.99°W), Porto Alegre (-30.05°N, -51.17°W), Brotas de Macaúbas (12.0011°S, 42.6303°W), and Macapá (0.0356°N, 51.0705°W)–representing 4-climate conditions. From the SRatio characteristics, (2) the dry-season data were used to estimate the SRate through the Theil-Sem estimator. These data are correlated with the difference in the spectral responses of the two PV technologies. In addition, a method for analyzing energy production data of PV systems in order to extract the soiling rate was developed and is presented. The I-V characteristics confirm differences associated with non-uniform soiling, and Pm measurements were used to quantify SRatio. This method considers the change in energy production from the PV systems associated with the accumulation of soiling during the dry season, wherein SRate is obtained through the estimation of the slope of the characteristic using the Theil-Sen estimator. The “monitoring station” and the “systems” methodologies are compared and discussed. The meteorological variables were related to the estimated soiling ratio, and the resulted indicated low correlation coefficient (R2) for the dry periods and showing the complexity of the soiling process. The novelty and major contributions of this research work include: (1) the development of the methodology to estimate and compare soiling losses from the soiling monitoring stations and the photovoltaic systems data from sites in Brazil; (2) the first-time acquisition and comparison of the annual SRatio and SRrate for three locations in tropical, subtropical, and temperate zones in Brazil; and (3) the combined use of crystalline-Si and thin-film CdTe modules at the locations to evaluate and validate the spectral-effect differences and those of  non-uniform/uniform soiling on the performances (including temperature distributions).