Water. desalination + reuse

proJects Skids for Pillbara are water tested in the factory before shipment to the site. Table 1. Water quality following initial sewage treatment In addition, an ammonia level of 25-40 mg/L, the temperature fluctuation, and changing weather conditions makes the treatment process a serious challenge. The ammonia level makes chlorination an unfeasible option, and a membrane step is therefore needed to remove viruses and bacteria. The original plan of using a pressurized UF (Pentair X-Flow's XIGA) would have been able to treat this water and provide the required log removal. However, the feed turbidity would have required substantial pretreatment to reach allowable and sustainable feedwater quality to the UF membranes. The DAF was originally designed to take a major part of the suspended solids - the algae, but the changing conditions affect the steady performance of the DAF system. DMF process With the experience of the sand filter backwash application, the route of DMF was challenged to see if it could provide the same water quality with fewer process steps, taking out the DAF and the media filter to potentially reduce investment and operational costs. The alternative route of processing is given in Figure 2. The requirements for chemical pretreatment is far less and needed only for chemically enhanced backwashes to improve membrane performance (not shown in the figure). Taking out the DAF and media filter eliminates the associated operating expenses along with hard-tocontrol process, particularly of the DAF. DeMonstration plant To make an educated decision, a demonstration plant was constructed by Tenix in Australia and operated for a few months in late 2012. A full-scale module plant was built and sent to the site where it was run by remote control. The target was to find what flux rate, coagulant dosing, cleaning interval, chemicals, and recovery could be used for sizing the UF plant. The application was the first of its kind. All the questions boil down to how applicable this product is for the process of separating the feedwater into high-quality effluent. Because the UV treatment is after the UF, the UV transmissivity was of particular interest to achieve sufficient log removal from the process. The decision to use the Pentair X-Flow Compact 33 V membranes was based on the fact they can provide 4-log removal on viruses and accept the feedwater without pretreatment (though eventually with some coagulant). The size of the lumen allows for solids feed level up to 1,000 mg/L and has the validation and approval of the California Department of Health for this virus removal. The High Solids Package (HSP) applies a dead-end filtration process where feedwater is fed into the lumen (insideout filtration) to be filtered. Alternating feed sides, air scour and backflushing keep the process in stable mode, minimizing pressure (typically < 0.5 bar) as shown in Figure 3. The plant design used three modules in parallel configuration and was fed from the last facultative pond of the treatment plant at Karratha. Inside the containerized demo plant were a receiving tank and 500-micron screen, after which the flow was pushed into the membrane units. The filtrate was collected in a permeate tank from which the backwash water was taken. Continuous monitoring of feed and permeate was done on-line to facilitate remote control and performance monitoring. Given the high turbidity and algae levels, coagulant (ACH) was dosed into the feed line, just before the UF unit. This supports the filtration process. Process flows and pressures, as well as the dosing rate and quality measurements were continuously monitored and recorded. During the trial a Pressure Decay Test was performed regularly in order to prove the integrity of the membranes. Apart from the quality measurements done on a continuous basis (turbidity) this test complies with the requirements of the Water Guidelines of Western Australia. Beyond this testing an extra challenge test was executed by the WRC in Melbourne on the MS2-phage test. The result of this proved to have a log removal of 4.4-5.8, well above the maximum allowable, which is a credit to the membranes on such removal. The demonstration plant found practical results in flux rates that were used to design the Figure 2. Alternative process train with Direct Membrane Filtration. August-September 2013 | Desalination & Water Reuse | 19 |

• Cover