Thermal, statistical, and machine learning-based analysis of a modified single-slope solar still for improved freshwater productivity

The demand for fresh water is increasing due to population growth, climate change, and industrial needs, making sustainable desalination increasingly important. This study considers experimental modifications to the single-slope solar still, including evacuated tubes, closed-loop R410-A-charged copper tubes, spiral fins, and wick materials, along with a passive external condenser. From March 2024 to February 2025, solar stills modified in four different configurations were tested and compared with a conventional solar still in the Jaipur climate. For thermal, statistical, and machine learning analyses, 1584 observations were used. The enhanced stills resulted in a 7°C–8°C rise in basin water temperature and a 40%–60% increase in freshwater productivity under peak summer conditions. The energy efficiency was 45%–73%, while for the conventional, it was 26%–40%. The attained exergy efficiency was 6.3%, compared with 1.8%–3.6% in the conventional system. The energy efficiency and distillate yield had uncertainties of ±3.0% and ±0.7%, respectively. Statistical validation showed paired t = −28.78, p < 0.001, and Friedman χ 2 = 154.16, p < 0.001. The SVM model was found to have a high R 2 of 0.9918, MAE of 0.085 L/m 2 , and RMSE of 0.116 L/m 2 . Water temperature and glass temperature are the most significant variables, contributing 47.93% and 24.32%, respectively.