Optimization of single-slope solar still with different energy storage materials using the taguchi and machine learning methods

Abstract

This research enhances the efficiency of single-slope solar still by utilising various energy storage materials, including paraffin wax, blue metal stone, basalt stone, and kanche marbles, through the application of the Taguchi method combined with machine learning techniques. An L 16 orthogonal array was employed to evaluate the effects of critical factors, such as energy storage material, water depth (2 cm, 4 cm, 6 cm, and 8 cm), water inlet temperature (30°C, 35°C, 40°C, and 45°C), and solar intensity (400 W/m², 600 W/m², 800 W/m², and 1000 W/m²) on distillate yield. The Taguchi results indicated that paraffin wax, functioning as a phase change material (PCM), yielded the maximum distillate output of 3.317 L/m²/day and a water temperature of 60.25°C at a water depth of 2 cm, with a water inlet temperature of 45°C and solar intensity of 1000 W/m². The Taguchi analysis determined that water depth is the most significant factor, followed by energy storage material and solar intensity. A machine learning model was utilised to forecast the distillate yield and water temperature under various conditions, achieving a prediction accuracy of ±1.33% for productivity and ±0.98% for water inlet temperature compared to experimental data. A confirmation test was performed to validate the optimisation process, utilising the optimal configuration of paraffin wax, a 2 cm water depth, a water inlet temperature of 45°C, and a solar intensity of 1000 W/m². This resulted in a distillate output of 3.76 L/m²/day and a water temperature of 60.30°C. The confirmation test results indicated that the Random Forest model aligned closely with the experimental values, exhibiting errors of 0.80% for distillate productivity and 0.25% for water temperature. This demonstrates that the model is dependable and efficient in forecasting system performance under optimal conditions.

Topics & Concepts