Comparative Cost of UF vs. Conventional Pretreatment for SWRO Systems Abstract P. Glueckstern and M. Priel, Mekorot Water Co., Tel-Aviv Surface seawater intended for desalination by RO need extensive pretreatment to control membranes fouling. Currently, conventional pretreatment is applied, consisting mainly on inline flocculation followed by media filtration. This kind of pretreatment usually yields satisfactory results, provided a well-designed seawater intake system usually a submarine type is used. However, considering long-term in order membranes performance stability, membranes manufacturers usually limit the operating flux to avoid frequent chemical membranes cleaning. Recently, Mekorot in cooperation with other partners, namely: Hydranautics, USA, Ben Gurion University, Royal Scientific Society Jordan and Sultan University, Oman, has concluded a three year research supported by the MEDRC, to test and evaluate the comparative field performance of capillary UF technology versus conventional pretreatment at two seawater sites: Eilat (Red Sea) and Ashdod (Mediterranean). Details of this research including operational results were reported at the EDS conference in Toulouse (July 2002). It is well known that the new backwashable microfiltration and ultrafiltration membranes technology widely used for treatment of polluted surface water has superior quality performance. The main reason for not adopting this technology for pretreatment in SWRO systems until lately, stems mainly from economic reasons and to some extent to lack of operational experience with seawater. Following the recent experience in pilot units and the continuously cost reduction, on account of the technological improvements and worldwide competition by several membranes manufacturers, this option is now seriously considered for application in seawater systems. In the presentation, the comparative economics of conventional versus membrane pretreatment for large SWRO systems will be analyzed, assuming several design approaches and projections of future cost reduction in membrane pretreatment systems. 1. Introduction Due to advancement in MF and UF technologies, their implementation for water filtration was dramatically expended in the last several years, and several hundreds of systems are in operation for municipal drinking water systems worldwide [1]. In the recent years, these new backwashable MF and UF technologies are widely considered as pretreatment in RO plants operating with polluted surface water such as river water, agriculture and municipal wastewater [2,3].
Continuous cost reduction in MF and UF membranes cost and especially their reduced cost in large systems [3], makes this option attractive also for pretreatment in large SWRO systems operating on surface feedwater, originating from an open intake source. Recent studies [4,5], indicate that the still higher investment in these technologies (in comparison with the conventional in-line flocculation followed by media filtration), is partially, and in some cases fully, compensated by cost reduction in their successive RO systems, by enabling higher permeate flux rate and higher permeate recovery. Mekorot had started to investigate the actual performance of backwashable UF capillary technology back in 1997 by conducting field testing at two sites where a polluted brackish source and a seawater source were desalinated. The first study was conducted in the framework of EU-Joule multi-institutional, multi-national research program [2] and the second one - in a research program supported by the MEDRC [6]. The MEDRC study included inter alia field evaluation of hybrid membrane systems, consisting of UF pretreatment and an RO seawater unit. The study was conducted by the Red Sea (Eilat site) and by the Mediterranean (Ashdod site). Details of the study, including operational results were reported at the last EDS conference in Toulouse, July 2002 [6]. The aim of the present paper is to analyze the comparative cost of UF vs. conventional pretreatment for implementation in large SWRO systems. 2. Design and Cost Basis Based on the performance of the Ashdod seawater pilot plant tests and economic data from various sources, and from the 10,000 m 3 /d Eilat SWRO plant in particular, an economic analysis of the comparative desalination cost of water from conventionally versus UF pretreated SWRO plant was evaluated. The main design and economic assumptions used for the above analysis are summarized in Table 1. The comparative investment cost and unit water cost of a 90,000 m 3 /d (30 Mm 3 /y) SWRO plant are summarized in Table 2 and Table 3 respectively. Flow diagrams and energy balance for the two compared 90,000 m 3 /d SWRO plants are shown in Table 1Figure 1 and Table 1Figure 2.
Table 1 Main design data and economic parameters used in the comparative economic analysis Seawater solid contents -TDS, mg/l 40,000 Seawater temperature, 0 C - range 16 30 - average 22 Pretreatment type CONVENTIONAL media filtration ULTRAFILTRATION (UF) Filtration rate, UF flux 8 m 3 /m 2 -hr 97 l/m 2 -hr Water losses 4.4% 8.1% RO product recovery 45% 50% RO membrane flux 13.5 l/m 2 -hr 15.5 l/m 2 -hr SWRO plant capacity 90,000 m 3 /day Number of units 8 7 Train capacity 11,250 m 3 /day 12,857 m 3 /day Product Chloride content < 240 mg/l Cl < 200 mg/l Cl Chemicals costs 4.82 cents/m 3 2.70 cents/m 3 Energy Consumption 3.57 kwh/m 3 3.56 kwh/m 3 Economic Parameters: - Fix charges - Electric power price - Plant insurance - RO & UF membranes replacement for progressive and future projection 10% 4.5 cents/kwhr 0.3% 14.3%/year 10.0%/year
Figure 1 SWRO desalination plant 90,000 m 3 /day, 30 M m 3 /year ULTRAFILTRATION pretreatment - Principle flow diagram Figure 2 SWRO desalination plant 90,000 m 3 /day, 30 M m 3 /year CONVENTIONAL pretreatment - Principle flow diagram Table 2 A 90,000 m 3 /day SWRO plant - comparative investment costs Pretreatment type Conventional UF INVESTMENT, K $ - Infrastructure 15,500 15,000 - Pretreatment System - Media Filt./ UF (excl. membr.) 9,000 12,200 - UF membranes 4,400 - Total Pretreatment System 9,000 16,600 - RO System - RO System excl. membranes 35,700 32,000 - RO Membranes 4,200 3,700 - Total RO System 39,900 35,700 - Total excl., membranes 60,200 59,200 - UF & RO Membranes 4,200 8,100 - Total incl., membranes 64,400 67,300 System Capacity, m 3 /day 90,000 90,000
Annual Production, M m 3 /year 30.0 30.0 Specific Investment, $/m 3 -day 716 748 $/m 3 -year 2.15 2.25 Table 3 A 90,000 m 3 /day SWRO plant - comparative annual and unit water costs K $/year cent/m 3 Pretreatment type CONV. UF CONV. UF Capital - 10% of investment incl. membranes 6,440 6,730 21.49 22.46-0.3% insurance 193 202 0.64 0.67 Total capital cost 6,633 6,932 22.13 23.13 Fixed O&M cost - Labour for 22 oper. @ 40 K$/Year 880 880 2.94 2.94 - Spare parts and oper. matherials @ 1% of inv. excl.mem. 602 592 2.01 1.98 - Membrane replacement @ 14.3% / year 601 1,158 2.00 3.86 - Overhead @ 5% 104 132 0.35 0.44 Total fixed O&M cost 2,187 2,762 7.30 9.22 Variable O&M Cost - Energy @ 0.045 $/kwh 4,814 4,808 16.06 16.04 - Chemicals 1,445 809 4.82 2.70 - Overhead @ 5% 313 281 1.04 0.94 Total variable O&M cost 6,571 5,898 21.93 19.68 Total 15,391 15,591 51.35 52.02 3. Analysis of results 3.1. Base Case The base-case assumes a higher specific RO permeate flux and recovery ratio for the UF design, but is quite conservative regarding all other parameters, such as membranes replacement cost and plant availability. The total investment in the 90,000 m 3 /d SWRO plant is estimated as some 64.4 M$ for the case where conventional media filtration is used and 67.3 M$ for the plant using UF pretreatment (see Table 2). The unit water cost calculated at the economic parameters defined in Table 1, amounts to approx. 52 cent/m 3 with the UF pretreatment (Table 3). Comparative cost breakdown of investment and unit water cost are shown in Table 4 and Table 5, respectively. From the investment cost breakdown, shown in Table 4, it is quite evident that the much higher investment in the UF pretreatment, 184 $/m 3 -d vs. 100 $/m 3 -d, is partially balanced by the reduced investment in the RO plant (397$/m 3 -d vs. 443 $/m 3 -d). This lower investment is obtained by operation at a higher RO membrane flux (15.5 l/m 2 -hr vs. 13.5 l/m 2 -hr).
Operation at a higher RO membrane flux is made possible due to the higher filtrate quality, but this is still pending on an approval of the membranes manufacturer. The unit water cost breakdown (see Table 3) shows that the cost components contributing to the higher cost in systems where UF is employed versus systems with conventional pretreatment, are the capital and the fixed O&M cost, including UF and RO membranes replacement cost. As all other cost components (labor & maintenance) being quite similar, the additional UF membranes replacement cost makes the difference. Table 4 Comparative investment costs in UF filtration Vs conventional MEDIA filtration for the 90,000 m 3 /day SWRO plants Filtration method MEDIA UF UF Vs. MEDIA M $ $/m 3 -day M $ $/m 3 -day M $ % Infrastructure 15.5 172 15.0 167-0.5-3.2% Pretreatment 9.0 100 16.6 184 7.6 +84.4% RO System 39.9 443 35.7 397-4.2-10.5% Total Investment 64.4 716 67.3 748 4.1 +4.5% Table 5 Comparative unit water cost breakdown of UF filtration Vs conventional MEDIA filtration for the 90,000 m 3 /day SWRO plants Filtration method MEDIA UF UF Vs. MEDIA K $/year cent/m 3 K $/year cent/m 3 cent/m 3 % Capital cost 6,630 22.13 6,932 23.13 1.00 +4.5% Fixed O&M cost 2,187 7.30 2,762 9.22 1.92 +26.3% Variable operating cost including 5% operational overhead - Energy 4,814 16.06 4,808 16.04-0.02-0.1% - Chemicals cost 1,445 4.82 809 2.70-2.12-44.0% - Total variable op. cost 6,571 21.93 5,898 19.68-2.25-10.3% Total 15,391 51.35 15,591 52.02 0.67 +1.3% The considerably higher fix O&M cost of about 1.9 cent/m 3 is partially compensated by the somewhat lower variable O&M cost, consisting of energy and chemicals costs. The UF system consumes similar specific energy (see Figure1 and Figure 2) much lower cost - about 44% - of chemicals (see Table 3). 3.2. Subjective approaches regarding selecting the membrane pretreatment for SWRO systems Depending on the customers, designers and manufacturers conditions and requirements, several approaches can be used when selecting the appropriate pretreatment for the plant: 3.2.1. Very conservative approach In this approach no credit regarding better feed water quality is given. In this case the RO system is designed with a conservative flux of about 8 gfd (13.5 l/m 2 -hr). Availability, RO
replacement and UF membrane replacement cost figures are not reduced in anticipation for future cost reduction. 3.2.2. Progressive approach In a case where both the customer and the system supplier have confidence in the improved performance of the integrated membrane system for desalination of seawater, relying mainly on comprehensive field testing, many of the potential credits are taken into account. In this approach it is reasonable to assume a considerably reduced RO and UF replacement cost (10% instead 14.3% per year) and to assume future UF membranes cost to be reduced by 50% 1. 3.2.3. Progressive approach at sites with extreme changes in seawater quality At locations where conventional pretreatment cannot yield the quality needed for plant operation during the whole year, due to occasional severe seawater quality deterioration, and longer downtime because of need of more frequent chemical cleaning of the RO membranes, a lower plant availability of 310-320 operating days, instead of 333 has to be taken in account. 3.2.4. Other factors Other factors have to be considered: System s availability, especially for implementation in the locations where seawater is of poor quality, or where land at seashore sites is scarce or where land cost is very high. The salinity of the product. These factors can be economically assessed only for specific site conditions. The land required for a SWRO plant operating at a high RO membrane flux using membrane pretreatment is only 50% or less of the area needed for a system using a conventional pretreatment, while the product salinity is considerably lower (see Figure 3 ). 1 Average replacement cost
Figure 3 SWRO permeate salinity Vs. seawater temperature and RO membrane flux Thus, if for example a product salinity of 150 mg/l Chloride is required, the SWRO with the conventional pretreatment would require a second stage desalination of about 40% (36,000 m 3 /day) while for the SWRO using UF pretreatment, a second stage of only 27% (24,000 m 3 /day) is needed. 3.3. Comparative results of cost according to the various approaches The resulting unit water costs according to the various assumptions/approaches are summarized in Table 5. Table 6 Comparative results of the various approaches Pretreatment method Case/Approach Availability ULTRAFILTRATION CONVENTIONAL % Investment K $ Unit Water Cost cents/m 3 BASE 91.2 67,300 52.02 VERY CONSERVATIVE 71,500 53.85 PROGRESSIVE 59,700 50.03 FUTURE PROJECTION 59,700 47.23 BASE 91.2 64,400 51.35 REDUCED AVAILABILITY 87.7 52.55 REDUCED AVAILABILITY 84.9 53.54
4. Selection of preferred pretreatment UF Vs. CONVENTIONAL The selection of the preferred pretreatment is pending, to a great extent, on the specific site conditions, on subjective approach philosophy and on the perspective of the UF technology. For reasonable good site conditions, including seawater quality and conservative or even very conservative approach to the expected benefits of UF vs. conventional pretreatment, the lower investment and lower unit water cost will justify the conventional option as the right decision. On the other hand, for sites limited in size and inferior seawater quality affecting plant availability, the UF will be seriously considered, especially if a progressive approach to the benefits of feed quality is adopted. 5. Near and long term projections The current cost per m 2 of UF membranes in this study is still much higher than the cost of RO membranes (30-40 $/m 2 vs. 12-15 $/m 2 ), but is being continuously reduced. By assuming a 50% cost reduction, the total UF pretreatment may eventually become equal to the conventional media filtration followed by cartridge micron filters polishing. This development will significantly reduce the unit water cost - by about 9% comparing to the estimated base case cost (Figure 4). A summary of the resulting unit water costs at the evaluated different approaches and scenarios is illustrated in Figure 4. Figure 4 Breakdown of unit water cost of CONVENTIONAL and UF pretreatment (90,000 m3/day SWRO) at various operating and economic assumptions
6. Conclusions At present, the conventional pretreatment is in most cases the preferred technology for pretreatment, because of its lower investment cost but also because of lack of experience of implementing UF to pretreat seawater. By adopting a more progressive approach after extensive pilot testing, and especially after further cost reduction, the UF/MF pretreatment method will most probably get more and more acceptance for application in large SWRO systems using surface seawater, originating from an open intake source. Beside the above analysis, it has to be emphasized that in addition to the considerations of all relevant site conditions, international tenders for both options are essential to explore real market prices, especially in the atmosphere of the continuously rapid technological developments and cost reduction of MF and UF technology. 7. References 1. W. T.Bates, Capillary UF as RO pretreatment, Presented at the International Water Conference, Pittsburg, PA. USA, October 1999 (IWC-99-22). 2. P. Glueckstern, P. Priel, A. Thoma and Y. Gelman Desalination of brackish fishponds effluents, EDS conference. Paris, France, 2000. 3. E. William, P. Mehul, A. Kevin, Selecting microfiltration equipment for a water purification mega project. Presented at the IWW conference, held in Muelheim on the Ruhr, Germany, Sept. 2002. 4. M. Wilf and K. Klinko, Effect of new pretreatment methods and improved membrane performance on design of RO seawater systems, IDA World Congress on Desalination and Water Reuse, Madrid, Spain, Oct. 1997, Vol. 1,357. 5. P. Glueckstern and M. Priel, Advanced concepts of large seawater desalination systems in Israel, Desalination, 119 (1998) 33-45 6. P. Glueckstern, M. Priel, M. Wilf, Field evaluation of capillary UF technology as a pretreatment for large seawater RO systems, Desalination, 147 (2002) 55-62.