Explicit prediction models for brackish water electrodialysis desalination plants: Energy consumption and membrane area. (1st June 2022)

• Record Type: Journal Article
• Authors: Siddiqui, M.U.; Generous, Muhammad M.; Qasem, Naef A.A.; Zubair, Syed M.
• Is Part Of: Energy conversion and management. Volume 261(2022)
• Publication Date: 2022-06-01
• Subjects: Electrodialysis -- Explicit model -- Regression -- Specific energy consumption -- Membrane area; Direct energy conversion -- Periodicals; Energy storage -- Periodicals; Energy transfer -- Periodicals; Énergie -- Conversion directe -- Périodiques; Direct energy conversion; Periodicals; 621.3105
• DOI: 10.1016/j.enconman.2022.115656
• Languages: English
• ISSNs: 0196-8904

Abstract

Highlights: An automatic regression model development strategy is presented. A global sensitivity analysis was carried out for electrodialysis desalination plant model. Regression models for electrodialysis membrane area and energy consumption were developed. Residual analysis was carried out to analyze the developed models. The results showed that the models are reliable with high accuracy. Abstract: With the demand for desalination systems to mitigate freshwater scarcity worldwide, there is a need for a reliable and straightforward model for the design and evaluation of desalination systems. Two critical metrics for membrane-based desalination systems are specific energy consumption and the required membrane size. The current study presents a strategy to develop simple explicit models for electrodialysis desalination systems. The procedure starts with determining important independent parameters using global sensitivity analysis on a dataset generated using a detailed analytical model of the electrodialysis system. Multi regression models are then developed using the model inputs standardized using their means and standard deviations. The set procedure was applied to an electrodialysis system configuration taken from the literature. After developing regression models for specific energy consumption and the total required membrane area, statistical and residual analyses were carried out to analyze the performance of the developed models. The developed models work for feedwater with salinity ranging from 1000 to 15000 ppm, while the permeate salinity is fixed at 200 ppm. The results showed that the developed regression models predict the specific energy consumption and the required membrane area with mean absolute percentage errors of 2.6% and 3.6%, respectively.