Structural design and machine learning of non-membrane solar evaporators: A review

Abstract

Desalination using solar evaporators is a viable solution to alleviate freshwater resource shortages. Among these, non-membrane evaporators, which integrate complex nanostructures with macrostructures to achieve multifunctional evaporation benefits, have become a research hotspot. Specifically, non-membrane evaporators can be optimally designed from perspectives such as enhanced light absorption or solute recovery at the macrostructure level to improve their practical evaporation performance. Meanwhile, non-membrane solar evaporators can develop highly efficient internal water transport channels. These channels not only improve water delivery efficiency and minimize heat loss but also help suppress salt crystallization, thereby ensuring strong salt tolerance and high heat transfer efficiency. In comparison, non-membrane evaporators constitute a complex system formed by the coupling of multiple structures, requiring sophisticated material selection and precise optimization of structural parameters. Therefore, leveraging the predictive capabilities of machine learning algorithms can greatly facilitate the development and design of solar evaporators. This work reviews recent advances in and contributions of machine learning to material screening, structural design, and system integration in solar evaporators. The insights gained are expected to support deep integration and wide adoption of machine learning in the design and performance prediction of non-membrane solar evaporators.