Water Journal : Water Journal September 2012-1
membrane technology refereed paper technical features 64 SEPTEMBER 2012 water chlorine contact time from 20 to 25 mg.min/L (Table 1). Membrane challenge testing was undertaken during June/ July 2010 to increase the membrane LRV accreditation for viruses. Challenge testing at a pilot or full-scale has been adopted as an approach to validate membrane treatment processes for pathogen removal (Jacangelo et al., 2006; Tazi-Pain et al., 2006). This paper describes the method and results of full-scale virus challenge testing at the GRWTP. Testing The challenge test design was undertaken by the Australian Water Quality Centre (AWQC) and United Water (Veolia Water) in consultation with the health regulator. All microbiological cultivation, sampling and analysis were undertaken by the AWQC and are reported in Keegan (2010), while the plant was operated by the City Green Alliance with United Water as operators, and performance is reported in City Green Alliance (2010a). The GRWTP had been in operation since December 2009 and, therefore, the membranes were approximately seven months old before the first challenge test in June 2010. Membrane challenge testing The challenge testing was based on the virus removal correlation through an intact membrane. The testing was conducted using introduced MS-2 bacteriophage, which is widely used as a challenge particulate for pathogenic enteric viruses (Jacangelo et al., 2006; Jacangelo et al., 2008; Arkhangelsky and Gitis, 2008; Langlet et al., 2008). MS-2 is similar in size to pathogenic viruses and is recommended in the US EPA (2011) generic protocol for challenge testing of UF membranes. MS-2 can be readily cultured and enumerated by laboratories. The stock MS-2 was prepared, enumerated and stored at --80°C until required. MS-2 was added and continuously mixed in an 800L batching tank filled with effluent from the feedwater basin. The effluent was tested to ensure there was no chlorine residual. The MS-2 was dosed continuously using a peristaltic pump from the batching tank into the suction side of a single dedicated pump in the membrane feedwater pumping station, drawing from the feedwater basin (Figure 2). During testing, the membrane pump station was operated continuously and pre-UF chlorination was isolated. The test design involved sampling the feed and filtrate of a membrane unit at four discrete time intervals (1, 12, 20 and 28 minutes) throughout a 30-minute filtration cycle. Four membrane units (#2, #4, #5 and #6) were tested individually on a single day and this was repeated over three days (Day 1 = 17/6/10, Day 2 = 8/7/10 and Day 3 = 28/7/10). The challenge test results were analysed by determining the LRV using the three methods outlined by the US EPA Membrane Filtration Guidance Manual (2005): • Method 1: LRV calculated from log10 of average of all influent samples and average of all effluent samples for a filter run; • Method 2: LRV calculated from average of the LRV calculated for time (t) paired samples for any filter run; • Method 3: LRV calculated as minimum of LRV calculated for time-paired samples for any filter run. Quality assurance included the use of the AWQC accredited to ISO9001, while United Water was accredited to ISO9001 and ISO14001. Quality controls included: background sampling for MS-2 prior to dosing; a trip control (MS-2 stock); an MS-2 survival assessment; and proof of mixing whereby MS-2 concentrations were compared from the membrane pump station to the inlet of the membrane unit. All samples were collected in triplicate. All samples consisted of approximately 250mL and were stored on ice for return to the laboratory. All sampling locations were equipped with flame-resistant sampling taps that could be heat sterilised using a portable butane torch. Enumeration was undertaken within 24 hours of sampling. MS-2 was analysed according to the method of Adams (1959) with E. coli 700891 as the host bacterium. Serial dilutions of the sample were prepared in tryptone soy broth. A 100 μL volume of sample and 20 μL of log10 phase E. coli was added to a molten 0.5% tryptone soy agar (TSA) overlay (cooled to 50°C). The mixture was poured onto solidified TSA supplemented with 150 μg/mL ampicillin contained in petri- dishes. After the top agar layer set, the petri-dishes were inverted and incubated for 16--24 hours at 35--37°C. The plaques were enumerated with a plaque identified as a circular zone of clearing, 1--5mm in diameter, in the lawn of host bacteria. Each sample was analysed for phage numbers (PFU) using three dilutions and one plate per dilution. The raw data were analysed to select those dilutions giving counts of between 20--200 plaques per agar plate and outliers were identified and removed from the analysis (Keegan, 2010). MS-2 numbers in the water samples were calculated by multiplying the mean number counted from the triplicates by the dilution factor, divided by the total volume plated. Membrane system and operation The GRWTP UF membrane system consists of a Siemens-Memcor CP Unit and L20 V module system comprising eight units, operating in a x7 duty and x1 standby arrangement, capable of 60-65L/s per unit. Each unit contains 120 modules, and each module contains approximately 10,000 hollow fibre PVdF membranes, with a nominal pore size of 0.04 μm and outside to inside filtration. The plant is controlled via an iFIX SCADA system with the operation of the UF system controlled by a dedicated PLC incorporating Siemens proprietary control logic. Based on an operator-specified flow requirement, the UF system outputs a pressure requirement on the feed side of the UF units and the upstream feed pump station operates to maintain the requested pressure. The feed flow control valves on each UF unit modulate to maintain the required flow rate (flux) through the on-line units as specified by the Siemens control system. Membrane integrity monitoring involves Pressure Decay Testing (PDT) and filtrate turbidity monitoring via dual on-line turbidimeters. During challenge testing, two membrane units were run at a time to match the capacity of a single pump (120 L/s) at the feedwater pump station. The time at which a unit was exposed to the MS-2 before the filtration cycle was minimised and sampling occurred Wetrack online instrumentation UF feed manifold UF discharge manifold To UV MS2 Dosing pump Batching Feed basin UF feedwater pump Figure 2. Schematic of full-scale testing apparatus.
Water Journal November 2012-1
Water Journal August 2012