Nanomembranes - Nanotechnologies for Water Environment Applications - page 400

Industrial City was less than 4 (Khawaji et al., 2007), while 3.5 - 4.3 was reported at the
Addur SWRO plant (Alawadhi, 1997). Fluctuation in the SDI value from 3.0 to 6.3 was
noted in a pilot-scale sand filter column using seawater off the coastlines of Singapore as
feed (Chua et al., 2003). The addition of 10-5-1 μm cartridge filters after the sand filter
was able to minimize the fluctuations, producing the SDI values within a range of 2.8 -
3.8 (Chua et al., 2003). Similarly, a cartridge filter (5 μm) after the dual media filtration
was able to lower the SDI value to a range of 3.3 - 4.0 at the Addur SWRO plant
(Alawadhi, 1997); the majority of the SDI values, however, still were above the
recommended SDI for RO membrane (> 3.0). The adverse effects of the poor feed
quality to the RO membrane experienced at the Addur SWRO plant had led to an
increased frequency in the RO membrane cleaning caused by fouling of colloidal
materials and biofouling effects. Chlorination used to control biofouling was adversely
contributing to the fouling effects. After dechlorination, the nutrients produced from the
reaction between organics in the seawater and chlorine were made available to support
growth of the surviving bacteria (Alawadhi, 1997).
12.4.2 Membranes in Pretreatment for RO Desalination

In view of the drawbacks in using the conventional pre-treatment methods,
considerable research has been carried out to study the feasibility of using MF and UF
membranes to replace or follow the media filtration method. The pilot-scale study
carried out by Chua et al. (2003) applied hollow fiber UF (pore size 0.01 μm) and MF
(0.1 μm) membranes separately to pretreat seawater off the coastal of Singapore. The
SDI values for both pretreatment methods were consistently below 3.0 (which satisfied
the recommended value for RO feed). The SDI values were less fluctuated with MF and
UF membranes compared with sand filters. In Japan, two dead-end flow membrane test
units, namely Unit 1 (nominal pore size 0.1 m; hydrophilic modified polyethylene

material) and Unit 2 (nominal pore size 0.01 μm; polyacrylonitrile) were tested as
pretreatment for RO desalination (Taniguchi, 1997). The permeate SDI value of Unit 1
was within the range of 2.5 - 4.0. Due to the lower nominal pore size, Unit 2 was able to
achieve a better quality with an SDI value ranged from 2 to 3. A trial testing using UF
pretreatment with an air enhanced backwash at an RO facility at Jeddah Port on the Red
Sea in Saudi Arabia was also reported to achieve an SDI value of consistently below 3.0
(Pearce et al., 2004). This SDI value was 2 units lower than the conventional
pretreatment system used at the same facility. The results with different feed seawater
from different geographical locations have clearly demonstrated that UF membranes
(with a nominal pore size of 0.01 μm) are able to consistently meet the required feed
water quality of the RO membrane. The membrane pretreatment was also able to
enhance RO operation whereby cleaning was not required during the 6 months of trial
operation at the RO facility at Jeddah Port (Pearce et al., 2004). UF treatment was also
able to provide good removal of colloidal silica and the coliform group of bacteria (less
than 1 cfu/100 ml), while a higher flux (almost double) was achieved with MF as

μ