Solar Distillation Systems Enriched With Machine Learning Techniques: A Review

Abstract Solar stills have the advantage of using solar radiation as they are the simple thermal energy source for saline water and industrial water desalination. This paper focuses on a detailed review of how a solar distillation system performance evaluation can be made with ongoing higher-end Machine learning techniques explicitly helpful in optimising and evaluating the still performance. Complete research on the implementation of ML models is made in this study to draw the feasibility of implementing the appropriate supervised or unsupervised machine learning methods. A comparison of the two of deep learning models applied in the advancement of the solar distillation process is explained in this study. The need for performance assessment of solar distillation system with Machine Learning Techniques is analyzed, and further significant features and components of ML and DL Methods are clearly explained. Keeping the importance of the study in front, a comparative analysis is made from the observations found in the literature review. We conclude that the Classification ML Techniques with ANN are the most appropriate models to predict the solar distillate while the ANN-MLP, ANN-FF models are more accurate than the other models. Instead of a traditional statistical approach, a DNN Hybrid model with more hidden layers can be used in optimising the water depth.