Review and Comparison of Methods for Soiling Modeling in Large Grid-Connected PV Plants

Abstract

Soiling in PV modules is one of the biggest issues affecting performance and economic losses in PV power plants; thus, it is essential to supervise and forecast soiling profiles and establish the best cleaning program. This paper analyzes different methods for soiling modeling in Large Grid-Connected PV Plants and discusses the different factors influencing soiling. Analytical models from environmental conditions are discussed in detail, comparing the proposed model by the authors (SOMOSclean) with another three relevant models from the literature (Kimber, HSU, and Toth), applying them to 16 PV power plants in Spain (total capacity of 727 MWp). Uncertainty between models and sensors is also measured, presenting the numerical results for a period of 2 years. While simpler models may offer straightforward implementation, they often fail to capture the full complexity of soiling dynamics, leading to increased RMSE error.