Thermal behavior in solar distillation system using experimental and machine learning approach with scaled conjugated gradient algorithm

Abstract

This study investigates the thermal dynamics of solar stills by combining experimental methods with machine learning techniques. The experimental setup consisted of two single-slope basin solar stills, from which temperature data were collected for the still, water, vapour, and glass surfaces. A machine learning model, specifically an Artificial Neural Network (ANN) using the Scaled Conjugate Gradient algorithm, was used to predict temperature variations throughout the distillation process. The ANN model achieved high prediction accuracy, with Coefficient of Determination (R2) above 0.99 for all temperature components. The study highlights the effectiveness of integrating machine learning into solar distillation systems, providing valuable insights for the design of more efficient technologies. These findings contribute to the understanding of sustainable water production and highlights the significant role of machine learning in optimizing solar distillation. Experimentally, the maximum water temperature reached 58.5 °C, with the highest basin temperature at 61.4 °C. The vapour temperature followed a similar trend, peaking at around 60 °C, demonstrating the efficiency of the evaporation process. The ANN model showed an impressive reduction in prediction errors, with an average relative deviation (ARD%) of <0.07 % and a mean square error (MSE) of 0.001227.