The U.S. Environmental Protection Agency (EPA) is initiating a peer review of draft scientific modeling approaches to inform EPA’s evaluation of...
Part one of this article appeared in the February issue and described how nanofiltration, reverse osmosis and electrodialysis reversal are being run side-by-side at the Brackish Water Demonstration Facility in California.
One Year Performance Comparisons
Operating costs and characteristics of the nanofiltration (NF), reverse osmosis (RO) and electrodialysis reversal (EDR) systems were monitored over the first year of operation (February 1999 to January 2000). The records kept during the first year of operation included all the labor costs, chemical costs, power consumption and downtimes. Therefore, EDR, RO and NF can be fairly compared using this data. The SCADA system permits the equipment operators to record the use of chemicals, power and time of operation for each membrane system separately. Maintenance costs were relatively low due to the one-year warranty coverage. However, mechanical improvements recommended by the operations staff were significant.
The average total dissolved solids (TDS) of the raw water feed to the plant is 1,000 mg/L. The system is designed to produce water of 370 mg/L TDS and 150 mg/L of hardness. These levels match the water quality of the state water imported from the Calleguas Municipal Water District (CMWD). Table 4 gives the average water analysis for the raw water and final blended product. Table 5 shows the product and reject quality from the three membrane systems.
Representative plant flows are shown in Table 6.
The total manpower cost was $168,000, which includes two operators for nine hours/day at a productive hourly rate of $32/hr. Labor was fairly evenly split between the technologies, with 16 percent for EDR, 15 percent for RO, 16 percent for NF and 53 percent for treatment plant operations including general and preventative maintenance, laboratory analyses and reports. Labor related to the EDR system included bi-weekly wash-downs and stack probing. CIPs were performed every 1,000 hours of runtime. Daily Silt Density Indexes (SDIs) on the NF and RO systems, as well as CIPs every two to three weeks due to particulate and biofouling, were the major labor requirements for these systems.
The daily cost for EDR chemicals was $30.87, which included hydrochloric acid, antiscalant (Argo AS120) and sodium bisulfite. The daily chemical cost for NF was $34.89. This included antiscalant (Argo AS120, Permacare 191 and Argo AF200) and sodium bisulfite. Cleaning chemical costs were $1,302 for EDR, $12,635 for NF and $841 for RO. The NF chemical costs are highest because proprietary cleaners were used for CIPs related to particulate and biofouling. The cost for NF antiscalant pretreatment also is about double per gallon in comparison to RO.
Power consumption was 1,275 kWh/day for EDR, 1,460 kWh/day for NF and 1,690 kWh/day for RO. Power cost is $0.069/kWh. The VFD pumps on the EDR system helped save on the power consumption. Considerable energy savings can be realized if the NF and RO pressurizing pumps were converted to VFD operation. The current RO and NF systems require an orifice plate and motorized valve to control inlet pressure and flow. If VFDs were used on the NF and RO, the power reduction is estimated to be 40 percent.
The total downtime for the EDR was 236 hours for the year for CIPs, stack probing, stack wash-downs and valve maintenance. The downtime for RO was 600 hours and for NF was 530 hours. This included moving membrane elements, CIPs and valve repair. Therefore, EDR had the largest production for the year at 342 million gallons followed by NF with 239 and RO with 224.
The RO and NF systems had problems with biofouling and particulates. Rotating the elements was effective and improved flows and pressure losses. The biofoulants clogged the first foot of the first element of the first stage. The systems are shutdown when a large differential is noted in the element. The first membrane element then is rotated to the last position in the second stage, and the last element in the second stage is rotated to the first position in the first stage. This cleans the fouled membrane and allows for continued operation. There was a manufacturing flaw in the NF membranes and the lead elements became delaminated during the first year of operation. The membranes were replaced through the warranty from the manufacturer.
The EDR currently has minor problems with biofouling and particulates but it has not affected the production, and the cleaning procedures reversed any trouble that was experienced. There was a design flaw in the water flow spacers. While they performed adequately, the spacers were replaced during 2000 with a new design.
Table 7 summarizes the costs per thousand gallons of water produced ($/kgal). The inevitable question is which technology worked out the best. Each system performed the job it was required to do—produce 1 mgd of 370 mg/L TDS and 150 mg/L of hardness water—but each had its own unique problems. The EDR has outperformed the spiral wound RO and NF systems in terms of cost of water produced, due to less downtime for CIPs and maintenance, less power consumption and better performance with particulate and biofouling agents that are present in the raw water.
There are many issues for continued study. Proposed pump control changes to the NF and RO systems may make them more cost effective. The fixed speed pumps on the RO and NF could be replaced with VFD pumps. This would present savings on power. The plant may experiment with different antiscalants, inter-stage pressure controls and additional pretreatment such as cartridge filters or particle dispersant chemicals. It will have to be determined if the power savings will offset the gain in the new operational costs. The EDR system performance also may improve with the proposed additional raw water pretreatment.
The raw water that currently is supplied to the plant is chlorinated. This soon will be supplied as chloraminated water and the plant operation will become capable of adding ammonia to convert the chlorine to chloramine. Filmtec has supplied documents stating that the NF and RO membranes can operate with the presence of chloramines in the water. The chloramines should provide help with the biofouling problem.
The NF may out-produce both the RO and EDR in the next year since the membrane defect problem has been corrected. Future EDR maintenance costs may increase, including reversal valve seal replacement and stack tear-downs for cleaning and spacer or membrane replacements.
The new Surface Water Treatment Rule regulations on microbial barriers and TTHM/HAA5 pass-through may require another step to the EDR process if the UWCD groundwater feed to the facility does not meet the more stringent regulations.