Browsing by Author "Maree, Patsy"
Now showing 1 - 1 of 1
Results Per Page
Sort Options
- ItemSurface water ultrafiltration pre-treatment through in-line coagulation and shallow bed media filtration(Stellenbosch : Stellenbosch University, 2017-03) Maree, Patsy; Burger, A. J.; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.ENGLISH SUMMARY: High concentrations of organic and microbial matter in surface water cannot always be removed sufficiently through conventional water treatment methods. Ultrafiltration (UF) membranes are preferred for this purpose, as the low feed pressures required are cost effective, and a smaller footprint area is required compared to a conventional plant. UF membranes effectively clarify and disinfect surface water. One notable hindrance to the use of UF for surface water treatment is the tendency of the membranes to foul. Particularly with the first annual rains, very high organic and solids loadings enter rivers. UF membranes can foul rapidly under these conditions. Therefore, this project investigated the use of in-line coagulation and integrated shallow-bed media filtration as lowcost pre-treatment prior to UF. The media filter and UF were filtered and backwashed at the same intervals, using UF filtrate to simultaneously backwash both units. A shallow media bed with a depth of 250 mm promoted fractional particle retention and floc modification, by promoting contact between destabilised coagulated particles. The media filter, which acted as pre-filter (PF), was characterised for low turbidity river water of 12 NTU, represented by a prepared (synthetic) suspension of 20 mg/L humic acids (HA) and 30 mg/L bentonite. Design and operating parameters were chosen such that the PF could produce filtrate with turbidity below 1 NTU as feed to the UF, for 40-60 minutes. Backwashing the combined system at 4 PF displacement volumes (for 2 minutes) effectively restored the retention capacities of and initial pressure drops over both filters. The system was operated at an average downflow rate of 10 m/h through the PF and a flux of 46 LMH through the membrane. Keeping the parameters selected during characterisation fixed, the performance of the combined system was tested with higher turbidity synthetic suspensions. The improvement brought about to membrane performance by the PF was of interest. For the low (12 NTU) and high (31 NTU) turbidity suspensions, reversible fouling occurred on the membrane, with a large percentage of the loading removed by the PF. No discernible irreversible membrane fouling took place. For low turbidity water, pre-treatment completely reduced TMP over the membrane, while at high turbidity water (25 mg/L HA, 100 mg/L bentonite), TMP was reduced by 70%. For the latter, breakthrough occurred after 15 minutes for 10 m/h PF filtration rather than after 50 minutes. With a very high solids feed loading (400 mg/L bentonite) at 8 m/h PF downflow, the filtration duration had to be shortened to 50 min for effective PF backwashing. The PF could still reduce TMP by 60%, and allowed a more gradual turbidity filtrate profile to be filtered by the UF. High organics loading of 70 mg/L HA did not present the same problem, and were effectively coagulated in such a manner that 99+% of TMP could be reduced through pre-filtration. On-site testing confirmed lab findings that most of the turbidity load could be removed by the PF, indicating that overall system recovery can be increased by prolonging filtration duration. However, this work unfortunately took place at a time when the feed water was rather clean (6 NTU). On-site test work needs to be conducted with high turbidity feed to evaluate the performance of the unit under such adverse conditions.