Forecasting Sustainable Water Production in Convex Tubular Solar Stills using Gradient Boosting Analysis

Abstract

Water scarcity is an important global issue that necessitates the development of sufficient and sustainable desalination technologies. This study forecasts the productivity of two solar distillation technologies, namely the conventional tubular solar still (TSS) and the convex tubular solar still (CTSS). The research objectives included assessing the distillate yield of both solar stills and investigating the application of an advanced gradient boosting machine learning (ML) technique for forecasting distillate production. Compared to the TSS, the CTSS demonstrated a calculated increase in productivity which indicates its potential as an effective water desalination technology. The correlation analysis revealed that the TSS exhibited 10 significant correlations while the CTSS exhibited 4 correlations. The application of the gradient boosting model revealed exceptional predictive precision for both solar stills. R-squared (R2) for the TSS model was 0.86, the Root Mean Squared Error (RMSE) was 58.2%, and the Coefficient of Variation of Root Mean Squared Error (CVRMSE) was 29.3%. In contrast, the CTSS model displayed impressive performance metrics, including an R2 value of 0.99, an RMSE value of 1.2%, and a CVRMSE value of 4%. Valuable insights were provided for the enhancement of solar stills, in addition to highlighting advanced ML techniques for accurately predicting productivity.