Prediction of dust particle size effect on efficiency of photovoltaic modules with ANFIS: An experimental study in Aegean region, Turkey


SOLAR ENERGY, vol.177, pp.690-702, 2019 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 177
  • Publication Date: 2019
  • Doi Number: 10.1016/j.solener.2018.12.012
  • Journal Name: SOLAR ENERGY
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Page Numbers: pp.690-702
  • Istanbul University Affiliated: Yes


In this study, the effect of coal dust in variable sizes and weight on photovoltaic (PV) modules' performance has been examined under laboratory conditions. Experimental studies have been performed under Standard Test Conditions (STC: Radiance: 1000 W/m(2); Cell temperature: 25 degrees C; Sun Spectrum: AM 1.5) for monocrystalline silicon (m-Si) and polycrystalline silicon (p-Si) PV modules. By using sieve analysis, the particle sizes of coal dust have been divided into six groups which are in mu m size and as follows: (- 38), (+ 38/- 53), (+ 53/- 75), (+ 75/-106), (+ 106/- 250), (+ 250/- 500). Artificial pollution has been created by uniformly distributing coal dust of certain size and weight onto PV modules. Three different weights of coal dust (5 g, 10 g and 15 g) have been employed for every single size of coal dust. In order to investigate the effect of any particle size and any weight of coal, the performance of PV modules has been investigated by measuring voltage, current and power. The data set consisting of electrical parameters has been used to develop a model by using Adaptive Neuro-Fuzzy Inference System (ANFIS). Comparison of experimental and ANFIS results have been given by calculating of Root Mean Square Error (RMSE) and coefficient of determination (R-2). The performance indices have been calculated as RMSE = 0.18719 and R-2 = 0.99803 for m-Si, RMSE = 0.87098 and R-2 = 0.99714 for p-Si PV modules. According to the results, for a given particle size and weight, the ANFIS model is quite successful in power estimation for PV modules.