Experimental Investigations on Solar Still using Phase Change Materials (PCM) and its Performance Prediction using Artificial Neural Network

Abstract

A significant amount of water available on the Earth is in saline form, and a very scarce quantity available is freshwater. In order to meet the requirement of freshwater, sustainable desalination techniques need to be explored for its low-cost generation. Solar desalination is a sustainable method using renewable energy which produces freshwater from brackish water by utilising solar energy. As freshwater generation capacity of conventional solar still is low hence energy storage materials (such as PCM) are embedded to store energy during the day duration, which can be utilised during off-sunshine hours (night hours). The present study analyses the impact of energy storage materials (PCM) on the freshwater productivity of solar still with PCM and solar still with PCM and pin fins. For the study, an experimental setup was designed and fabricated; and the results were compared with the conventional solar still under the atmospheric conditions of Gandhinagar, Gujarat, India (23o10′ N 72o38′ E). The variation in the mass of PCM and the mass of brackish water was considered as study parameters. The result shows that the total accumulated productivity in solar still with PCM and solar still with PCM and pin fins improved by 24% and 30%, respectively, as compared to conventional solar still. Moreover, the freshwater productivity of the solar still with PCM and pin fins improved by 8% as compared to the solar still with PCM. The energy efficiency in solar still with PCM and solar still with PCM and pin fins enhanced by 36% and 47%, respectively, as compared to conventional solar still. The energy storage material stores the energy during the day duration, which will be released during the night duration, so the solar still with PCM and pin fins generates more amount of freshwater after sunshine hours. The cost per litre ($/L/m2 ) was 0.022, 0.019, and 0.017 and the payback period (days) was 113, 97, and 89, respectively, for the conventional solar still, solar still with PCM and solar still with PCM and pin-fins, respectively. The effect of the seasons on the freshwater productivity of solar stills using three different fatty acids (PCM- lauric acid, palmitic acid and stearic acid) as a thermal storage medium along with pin fins were compared for the different seasons (winter and summer). The accumulated productivity, energy efficiency, and exergy efficiency were observed as 4.43 L/m2, 47.9% and 2.23% for solar still with lauric acid and pin fins (SSLA) in January, which was the maximum vi among all cases in January. The accumulated productivity, energy efficiency and exergy efficiency were 5.74 L/m2, 56.2% and 2.96% for the case of solar still with stearic acid and pin fins (SSSA) in the month of May, which was the maximum among all the cases. The SSSA produces more freshwater productivity due to higher latent heat and thermal conductivity of the stearic acid. The cost per litre and payback period for the solar still with stearic acid and pin fins was 0.016 ($/L/m2) and 72 days, respectively, as compared to 0.022 ($/L/m2) and 99 days for the conventional solar still. As solar desalination experiments are time and resource-consuming methods, there is a need for a robust system to identify the forecast the freshwater productivity. In order to forecast the freshwater productivity of the solar still with PCM and pin fins (SSPCM) and conventional solar still (CSS), the LSTM (Long Short-Term Memory) and GRU (Gated Recurrent Unit) were considered. The accuracy of the LSTM and GRU was assessed using different statistical criteria. As compared to GRU, the coefficient of determination of the forecasted freshwater productivity using LSTM shows a better accuracy of 0.96 and 0.98 for the CSS and SSPCM, respectively. The LSTM can remember the data pattern for a longer duration, so LSTM can accurately forecast the freshwater productivity of the solar still. This study investigates the effects of eco-friendly biodegradable fatty acids (as compared to paraffin wax) on the performance and productivity of solar stills. Also, the forecasting and performance prediction of solar stills with PCM using recurrent neural networks (RNN) is explored. The Long Short-Term Memory model accurately forecasted the freshwater productivity of solar stills as compared to the Gated Recurrent Unit model.