Abstract Access to potable water with standard quality is an inevitable component of human's lives. Freeze desalination, by consuming lower energy compared to other techniques, relies on the exertion of a cold source which is then accompanied by simultaneous rejection of impurities from water. Regarding this, a numerical study on freeze desalination in a hollow cylinder is carried out to determine the effects of the design variables such as heat flux, hydraulic diameter, initial salt concentration, and freezing time on the ice mass, ice salinity, ice generation speed, and Nusselt number on the cold surface of the inner tube. Results show that increasing the value of heat flux from −250 W m 2 to −1000 W m 2 imposes 3.9 times faster ice generation speed while decreasing the desalination rate by 22%. In opposite, increasing the hydraulic diameter from 2 to 8 cm, reduces the ice generation speed by 41% but improves its quality by 23.5%. Also, ice production and salinity would both increase with the freezing time which imposes a limit for freezing duration. Multi-objective optimization and artificial neural network with the purpose of obtaining the highest ice mass with lowest salinity are employed. Finally, multistage freeze desalination is explored to reach WHO standards.

中文翻译：

---

间接冷冻海水淡化数值研究与优化

摘要 获得标准质量的饮用水是人类生活的必然组成部分。与其他技术相比，冷冻脱盐消耗的能量更低，依赖于冷源的使用，然后同时去除水中的杂质。为此，开展了空心圆柱体冷冻海水淡化数值研究，确定了热通量、水力直径、初始盐浓度、冻结时间等设计变量对冰体、冰盐度、产冰速度的影响。 ，以及内管冷表面的努塞尔数。结果表明，将热通量值从-250 W m 2 增加到-1000 W m 2 使冰生成速度提高了 3.9 倍，同时使海水淡化率降低了 22%。相反，将水力直径从 2 厘米增加到 8 厘米，制冰速度降低了 41%，但其质量提高了 23.5%。此外，冰产量和盐度都会随着冻结时间的增加而增加，这对冻结持续时间施加了限制。采用多目标优化和人工神经网络，目的是获得最高冰量和最低盐度。最后，探索多级冷冻海水淡化以达到世卫组织的标准。