Desenvolvimento e identificação experimental de planta piloto de dessalinização por osmose reversa

• Ano de defesa: 2021
• Autor(a) principal: Queiroz, Dayse Maria Benevides de
• Orientador(a): Carvalho, Paulo Cesar Marques de
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Dessalinização por osmose reversa; Modelagem de sistemas reais; Método mínimos quadrados não recursivo; Acionamento fotovoltaico; Dessalinização da água; Osmose
• Link de acesso: http://www.repositorio.ufc.br/handle/riufc/70048

The reverse osmosis desalination (OR) presents a viable solution for regions where there is little availability of potable water, with increasing application in semi-arid regions, such as northeastern Brazil. In this context, this dissertation presents the development, installation, and identification of the mathematical model of a reverse osmosis desalination system. The plant structure in modular aluminum has dimensions of 1.45m x 1.00m x 0.48m, ensuring flexibility in the installation and ease of handling and configuration. The devices that compose the system are; two hydraulic diaphragm pumps, a pre-filtration assembly, pressure and flow sensors, human-machine interface (HMI) data acquisition system, a pair of reverse osmosis desalination membranes positioned in series, performing processing with and without feedback. To verify the performance of the desalination method, three trials with different levels of total dissolved solids (TDS) for the feed water were conducted for both the system with and without recirculation, but under equal operating conditions. The first trial, with 2,300 mg/L, brackish water level, the second trial, also with brackish water, with 3,900 mg/L, and the third trial was performed with saline water STD level, 5,130 mg/L. The configuration with feedback showed better performance in reducing STD and in the amount of permeate than with feedback, but it was only possible to obtain potable water in the process with recirculation with 2,300mg/L of feed water, the other trials performed removal of more than 50% of the initial amount, but still unsuitable for consumption. As the salinity increased, the nominal pressure of 50 psi was insufficient to achieve STD reduction. A single trial collected one set of data to identify, and another set of data to validate the estimated model. The identification was performed through the non-recursive least squares method, obtaining four ARX models, of 1a, 2a, 3a, and 4a order. After identifying the models, validation was performed through simulation and noise analysis. The 1a order model presented a performance of 68,92%, the 2a order model 72,8%, the 3a order model 74,6%, and the 4a order model 73,3%. Thus, it is noted that the 3a order model was the one that best described the developed RO desalination system.