Performance comparison of different mathematical models in the simulation of a solar desalination by humidification-dehumidification

• Published in: Desalination Vol. 437; pp. 184 - 194
• Main Authors: Lacerda de Oliveira Campos, Bruno, Oliveira Souza da Costa, Andréa, Cecília de Souza Figueiredo, Kátia, Ferreira da Costa Junior, Esly
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.07.2018
• Subjects: Mathematical model; Multi-objective optimization; Parameter estimation; Solar desalination; Parameter estimation; Multi-objective optimization; Mathematical model; Solar desalination
• ISSN: 0011-9164; 1873-4464
• DOI: 10.1016/j.desal.2018.03.014

Solar desalination by humidification and dehumidification (HDH) is an interesting desalination technique, which has been the subject of many studies, including mathematical modeling. The objective of this work was to propose a mathematical model that is more efficient to predict distillate production than the ones in literature, maintaining reasonable temperature predictions. The heat and mass transfer coefficients were estimated by the minimization of sum of squared errors of both temperature and distillate predictions, joined in a parametric global multi-objective function. Different modeling strategies were proposed and seven models were compared. For the studied system, not considering air saturation for the hot air stream improved distillate predictions significantly, but the introduction of sodium chloride property equations had no considerable impact in model performance. The distillate mean error of models C and D (4.3%) is 31% lower than the model in literature that presented the best predictions. A sensibility analysis showed that an increase in the column heights may cause beneficial and adverse effects in distillate production, so that there are optimum values for each specific system. In the studied process of the present work, the recommended heights for the condenser and the humidifier are 0.50 and 0.35 m respectively. •Seven mathematical models were developed and compared.•The estimation of the model parameters is made through a multi-objective optimization.•The distillate production relative mean error of the best model was 4.3%.•Condenser and humidifier recommended heights were investigated.