Water Journal : Water Journal May 2011
membranes & desalination refereed paper technical features 108 MAY 2011 water Attempts to further reduce the chloramine dose were unsuccessful. A reduction from 2mg/L to 1.5mg/L resulted in deteriorating RO performance. Flux Reduction The design flux for the Botany GTP (24.9L/m2.h for primary, 18.1L/m2.h for secondary) could be considered conservative when assessed against the FilmtecTM membrane system guidelines (2009) for well water (27--34L/m2.h). Although groundwater may normally be classified as well water for the purposes of RO system design, the brackish nature and presence of significant RBOC and upstream biological activity justifies classification as wastewater. The flux design guidelines for conventional filtered municipal effluent are 14--20L/m2.h. Due to ongoing process issues in the upstream clarifiers and media filters, iron residuals of 0.1--0.3mg/L in the RO feed water also contribute to RO fouling. While work continues to improve upstream iron removal, overall permeate flux was reduced to mitigate both iron and residual biological fouling. The GTP currently operates at fluxes 20% below the original design, and overall recovery reduced from 91% to 89%. In addition, interstage balancing has been retrofitted to each skid by installation of flowmeters and manual permeate throttling valves. Table 2 summarises the current operational parameters of the RO process. Due to groundwater extraction yields below prediction, the plant is only required to run at 50% capacity to achieve hydraulic containment. This is achieved by operating two of the four primary skids and one of the two secondary skids, and maintaining one primary skid and one secondary skid in standby. The fourth primary skid sits idle, without membranes. All the original membrane elements have been replaced over time. Although the originally installed membranes have been subjected to numerous chemical cleans, membrane performance has been maintained. Most of the replaced membranes suffered physical damage, particularly glue line failures. The Toray and FilmtecTM membranes in the primary skids have been replaced like-for- like, while the FilmtecTM BW30-400/34i membranes in the secondary skids have been replaced with BW30-440i membrane elements. With this arrangement, primary RO skid run times now exceed nine weeks, and secondary skid run times exceed four weeks. Periodic Flushing Increasing differential pressure between the feed and reject streams is caused by particulate fouling. Particles accumulate in the feed/concentrate channel spacer of the membrane elements, restricting the flow. Particulate fouling from feed contamination is unlikely in the secondary RO skids, having already passed through two stages of membranes in the primary skids. However, rising differential pressure, particularly on the second stage of the secondary skids (stage four overall), is the most common initiator for chemical cleaning at the GTP. This is attributed to precipitation of residual iron and concentration of residual organic carbon facilitating biological activity, even with a chloramine residual. The improved run time on the secondary skids has in part been attributed to an automated flushing regime. The membranes are flushed with permeate for four minutes every 24 hours. Recovery of differential pressure climb after a flush is not complete, nor immediate, as illustrated in Figure 9. The rise in differential pressure is arrested, and then falls slightly for about 12 hours after the flush before rising again. This response suggests further optimisation of flushing frequency may be possible. Successive flushes have a diminishing effect. When the differential pressure reaches an alarm point, a warm water flush is initiated. Permeate is heated to 38˚C and circulated via the membrane CIP system. Each array is circulated separately for 10 minutes, and then flushed for 10 minutes through both arrays to waste. The warm water flush provides a significantly greater recovery Figure 9: Secondary skid performance with periodic flushing. Figure 10: Salt passage throughout the flush cycle.
Water Journal April 2011
Water Journal July 2011