Calculating the impact of meteorological parameters on pyramid solar still yield using machine learning algorithms

• DOI: 10.1016/j.ijft.2023.100341
• Publication Date: 2023-03-28T16:30:34Z
• Authors: Mahmoud Elgendi; Mohamed Atef
• Keywords: Solar still; 0211 other engineering and technologies; Meteorological parameters; LR; 02 engineering and technology; ANN; QC251-338.5; Heat; Productivity

Deterimining the effect of individual meteorological and solar still parameters on productivity is challenging. This because are correlated experimentally isolating their impact difficult. However, locations that optimize can be determined by analyzing these yield. In this study, two algorithms, an artificial neural network (ANN) linear regression (LR), were used to predict yield still. Four important identified different in prediction models. The number input features was reduced using principal component analysis (PCA) a correlation function, which complexity ANN model. Based feature importance, Solar, Tf, Tout, Wind had most influence productivity, as demonstrated accuracy model complexity. LR trained identify models fit data with minimum error. perfomed better than model, mean absolute error (MAE), R-squared (R2), root square (RMSE), it nonlinear systems. performance influenced various parameters; however, determining specific each parameter challenging low-complexity ANN. By excluding have weak eliminating those strong dependencies, four accurately still's identified. approach 72% compared all 12 general parameters. researchers who conduct several days experiments determine yield; they compare findings for values addition, location based importance