Plant Case Histories IDA 2005 (Singapore) Paper: Optimizing Membrane Plants via Correlating Fouling with Critical Flux

These case histories are based on actual design and operating history data collected from 2 membrane water purification plants:

CASE HISTORY A:
Performance Comparison of RO and NF Systems
Request Paper on Real-time Membrane Fouling Monitoring - Port Hueneme NF-RO Plant Case History

The nanofiltration train (NF-A) and reverse osmosis train (RO-A) at Port Hueneme Brackish Reclamation Demonstration Facility in Southern California have both suffered from biofouling since startup in November 1998 until the plant implemented a raw water choramination scheme in June 2000 (around 13,500 hours of operation). Moreover, the nanofiltration membrane elements also suffered from a manufacturing defect that was discovered later. Plant operators monitor the membrane systems' "health" by inputting daily average analog values for flows, pressures, and conductivities into Dow Filmtec^®'s ROSA^® program configuration file. They use the expected fouling factor of 1.0 and the set points for flows as published by Filmtec^®. These set points are then verified against operating product flow and differential pressure trending graphs.

Recently, the plant provided operating history data from both trains for performance evaluation using MASAR^®'s unique measurement of the Fouling Monitor (FM) (blue curve) against ASTM-normalized flux decline trend, using ASTM D 4516-00 standard method (red curve). Not only did MASAR^® show the progress of biofouling in both trains (rising FM values) up until chloramination was instituted, but it interestingly showed a significant improvement due to more effective disinfection ONLY in the RO train! The NF train's FM performance showed a much lower degree of improvement conceivably due to biomass "sloughing off" which could be caused by elements leak. The plant's daily monitoring procedure, just like the ASTM normalized flux decline trending, failed to pick up this important difference in Train NF-A's actual behavior as MASAR^® did, and actually showed a misleading trend!

Had the plant been using MASAR^®, it would have given the operators a true and early indication of the performance and fouling status of both trains. Corrective measures would have then been immediately implemented after investigating the sources of deterioration. That early determination would have had an obvious positive impact on the plant's operational efficiency and cost-effectiveness.

We wish to thank Mr. Joe Richardson, former Water Utility Operator II at Port Hueneme Water Agency for providing Demonstration plant's operating data and input on its performance history.

CASE HISTORY B:
Biofouling Plant Performance
If above link does not open a window on this page, it may be minimized at the bottom bar.

The City of Clifton, Colorado, designed and built a 2.4 million gallons per day nanofiltration plant to treat Colorado river water and supply the city and nearby areas with drinking water. The plant was partially started in September 1997 with NF Skid 1. It was shutdown in the Spring of 1998 due to low river TDS. The plant was re-started in August 1998. Shortly after, Skid 1 was cleaned after about 1,400 hours of operation to control the rising membrane pressure drops and permeate flux decrease due to suspected biofouling in the system. Two other cleaning cycles followed in November and December 1998. Operating data for the entire period since startup in 1997 till middle of December 1998 were provided to MASAR Technologies, Inc. for performance and fouling evaluation.

MASAR^®'s unique analysis of the Fouling Monitor (FM) history showed that biofouling actually started to develop in Skid 1 as early as beginning of December 1997 (after only 200 hours of operation) where the FM started to rise from an average of 0.5 % at 50 hours to 3.6% at 290 hours, while the ASTM-normalized flux continued to show a improvement (as shown in the top graph). When Skid 1 was cleaned for the first time, the FM was 4.0%, indicating some type of moderate to significant fouling already in progress. During the same period, the measured membrane DP showed a very slight, hardly noticeable and certainly not alarming, rising trend while Osmonics' calculated Fouling Factor showed a very stable performance with no apparent change from baseline value of 1.2 (as shown in the bottom graph). Both indicators took a sudden turn for the worse in late August 1998 prompting the need for cleaning in September.

Had the plant been using MASAR^® to monitor its performance, the operators would have been able to realize the development of a serious fouling problem very early within the first month before significant deterioration in pressure drops and permeate flux started to occur. Corrective measure would have then been taken and plant performance restored and maintained at startup levels.

We wish to thank Mr. Dave Payne, Water Treatment Plant Supervisor at Clifton Water District and Mr. Peter Eriksson, Senior Applications Development Engineer at Osmonics, Crossflow Filtration, for providing plant's operating data and input on its performance history. (Ref. - Small Systems: Nanofiltration as Salt Reducer, Behind the Clifton, Colorado, Project, Peter K. Eriksson, Water Conditioning & Purification, pp. 60-64, April 2001, Publicom, Inc., Tucson, Arizona.)