Modelling and simulation of desalination process using artificial neural network: a review

Water is the natural, yet very essential, resource for survival of humans, animals and plants. However, only 3% pure water (present in lakes, rivers, as groundwater and frozen water) is available globally and 97% being saline is not suitable for drinking and agriculture purposes. Surprisingly, only 1% of this pure water is within reach of humans for existence. Hence, it is quite imperative to improve the water quality as well as its availability. Desalination, a process for converting the saline water into fresh water, may help in achieving this objective by providing water suitable for consumption by humans and animals, for agriculture and industrial applications. In this paper, we review various desalination techniques namely: reverse osmosis, vapor compression distillation, electrodialysis, multi-stage flash, etc., and their hybrids being increasingly used for treating seawater. Modelling and simulation of such processes is vital for improving water quality and quantity as well as understanding, analysis and reporting of the physical, chemical and biological results for appropriate process measurement and control. Artificial neural network (ANN) involves representing such processes with models inspired by the architecture of a biological neural network of human brain. An exhaustive review of ANN-based models, improvised recently to more effectively simulate process behavior for optimizing operating conditions, has been presented.