Multi-objectives heuristic GP and machine learning optimization of a solar-powered humidification-dehumidification water desalination system

Abstract The escalating global demand for fresh water necessitates sustainable desalination solutions. Humidification-dehumidification (HDH) solar desalination systems offer a promising and low-impact approach to environmental sustainability. This paper details a thorough investigation into optimizing the performance of these systems. A detailed model is developed encompassing the solar collector, humidifier, dehumidifier, and condenser, incorporating governing equations for the transfer of heat and mass. This model is intended to be validated by a small-scale HDH desalination setup. Furthermore, the paper introduces a novel multi-objective heuristic gradient projection (MO-HGP) optimization technique. This method simultaneously considers objectives of maximizing freshwater production rate and system performance efficiency, leveraging heuristic principles and gradient projection to identify optimal system configurations. After applying the optimization and machine learning technique, a detailed analysis of performance improvements is compared to conventional approaches. Finally, to enhance efficiency and promote wider adoption, the research implements the integration of a solar tracking system (STS) into an existing HDH desalination unit. The theoretical and practical impacts of STS on increased solar energy collection and its direct correlation with higher freshwater production rates are analyzed. Through this integrated approach of theoretical modeling, and advanced optimization, including solar tracking, this paper demonstrates the potential for a significant average annual efficiency improvement of approximately 23%. This advancement substantially enhances the viability of solar-powered HDH desalination, particularly for remote areas with significant solar exposure and limited water availability, offering a pathway for more sustainable water resource management.