Assessment of UF pretreatment prior RO membranes for seawater desalination

Among the desalination technologies available, reverse osmosis is currently considered as the most economical process. However, while RO membranes have been continuously improved over the years, they have had one major shortcoming; the susceptibility to fouling. Indeed, it appears that RO technology is limited in hydraulic performance by the raw water quality. In the case of direct seawater intake, extensive treatment is required upstream the RO process. Thus, the objective of this study was to evaluate the ultrafiltration membrane process as a RO pretreatment. The trials were conducted at Fuerteventura, Canary Island. The preliminary results showed that the Aquasource® UF membrane does not seem sensitive to high salinity of the seawater. Indeed, the mechanical resistance, as well as the specific flux of the UF membrane, remained stable. The dead-end filtration mode coupled with operation at low pressure allowed for very low power consumption, approximately 0.1 kWh per m3 of permeate. Concerning the RO membrane evaluations, the hollow fine fiber (HFF) membrane was more efficient in terms of salt rejection, but the thin film composite spiral wound (SW) membrane operated at lower pressure (i.e. lower energy consumption). To date, the study continues in order to develop an integrated membrane system (UF/RO) to treat surface seawater.

The Aquasource® UF membrane does not seem sensitive to high salinity of the sea water. Indeed, the mechanical resistance as well as the specific flux of the UF membrane remained stable. The dead-end filtration mode coupled with operation at low pressure allowed a very low power consumption, approximately 0.1 kWh per m3 of permeate. These results, however, have to be confirmed on surface intake. Comparison between RO modules, each one manufactured with a different technology, shows that hollow fine fiber (HFF) membranes operate at greater pressures than thin film composite (TFC) membranes to obtain similar permeated flows. This necessarily results in higher specific electrical consumption for the HFF modules. However, the resulting HFF permeate quality is superior than that produced with TFC membranes. Thus, the type of membrane selected for a specific situation depends on the final water quality required and the associated efficiency needed (i.e. water cost). At the time of preparation of this paper, further trials are on-going (flux at 140 L/h.m²) in order better define the limits of the UF membrane. Following this test, UF will be tested on surface intake and the UF water will feed the RO pilot. The RO pilot tests will be conducted side by side to evaluate the impact of the pretreatment on the RO hydraulic performances. Acknowledgments This research forms part of the EC Joule Programme Project JOE3-CT97-0053 undertaken by Suez-Lyonnaise des Eaux, CPERI of the University of Thessaloniki, AGBAR of Barcelona, Mekorot Water Company and the Water Research Institute (WRI) of the Technion—Israel Institute of Technology. Permission to publish this paper is gratefully acknowledged.