DESALINATION OF SEA WATER USING REVERSE OSMOSIS METHOD

Abstract

In this research project, the performance of a B10 sea water desalination permeator was studied by conducting experiments using a small commercial RO unit. In this regard, pure water, salt (NaCl) solutions and sea water were used as feed solutions. The experimental results obtained were used to test the applicability of three permeator transport models, namely, the Ohya-Sourirajan approach, the complete-mixing model, and a lumped parameter approach introduced in this study. The lumped parameter approach attempts to model the performance of the B10 permeator by assuming that the shell-side concentration is a linear function of the permeator surface area. This one-dimensional approach results in the simplifying relation that the shell side concentration is given as the average of the feed and reject concentrations. With respect to the RO experiments, pure water tests showed that for a feed pressure of 55.15x10? N/m², feed flow rate of 200 cm³/s and a temperature of 28°C, the B10 product water recovery was 68.9%. Experiments using salt solutions showed that the product water flux is linearly dependent on the hydraulic driving force while the salt flux varies non-linearly with respect to feed pressure and feed concentration. The results showed that for a feed concentration of 35,000 ppm NaCl, an operating pressure of 55.15x10? N/m², a feed flow rate of 200 cm³/s and a temperature of 28 °C, the product recovery was found to be 24.0%. The salt rejection in this case was 97.6%. For the sea water experiments, the product concentrations were found to be generally lower as compared to the results for an equivalent concentration of NaCl feed solution. At a feed concentration of 33,000 ppm, an operating pressure of 55.15x10? N/m², a feed flow rate of 200 cm³/s and a temperature of 28 °C, for example, the product recovery was found to be 29.5% and the salt rejection was 99.2%. Indeed, the solute rejection was found to be higher than 99% for feed concentrations of less than 40,000 ppm sea water. Analysis of the experimental results showed that the use of the Ohya-Sourirajan approach produced some scatter in the nondimensional transport parameter, ?. This meant that using this approach to model the permeator performance would incur some inaccuracies in both product flux and salt rejection. Similarly, the complete-mixing model was found to be inadequate since the transport parameters could not be determined for some of the experimental points. In contrast, the lumped parameter approach gave good fit to all experimental points, with errors in product and salt flux of well within 10% for high feed concentrations and much less at lower feed concentrations. The simplicity of the lumped parameter approach also allowed its use in desalination economic studies. For any operating pressure, the capital and energy cost decreases with increasing product recovery, while the permeator cost increases, thereby exhibiting a minimum water cost at some optimum recovery. The use of an energy recovery system reduced the energy requirement cost of the RO plant by 30%. The water cost also decreases with plant capacity, being about US$ 1.40/m³ for a 37,854 m³ /day plant capacity. It should be emphasized that the product water cost can be reduced further by utilizing larger-sized permeators which have lower cost per unit membrane area. Indeed. recent economic studies have shown that it is possible to obtain a product water cost of less than US$ 1.00/m³ in a properly designed RO plant.