This paper was presented at IEX 96 in Cambridge,UK on July 15 th 1996 and is the copyright of the SCI

Transcription

1 This paper was presented at IEX 96 in Cambridge,UK on July 15 th 1996 and is the copyright of the SCI AN ECONOMIC COMPARISON OF REVERSE OSMOSIS AND ION EXCHANGE IN EUROPE P.A.NEWELL, S.P.WRIGLEY, P.SEHN & S.S.WHIPPLE Liquid Separations Technical Service and Development, The Dow Chemical Company, D Rheinmuenster, Germany 1 INTRODUCTION The comparison between ion exchange and reverse osmosis membrane technology as options for water treatment applications has been the subject of a number of studies and continues to generate a high level of interest in the literature 1. Most of these studies have been carried out in the USA with little reported in other regions. As technical developments in both product areas continue to be made, such as counter-flow regeneration packed bed systems, narrow particle sized ion exchange resins and high rejection,lower energy membranes, the need for monitoring economic performance remains. In addition, external factors such as water costs and disposal, power and chemical costs continue to change and are different around the world, further affecting the economics. The purpose of this paper is to make a comprehensive economic comparison reflecting recent technical developments and using European costs and practices. 2 DESIGN BASIS Four feed water salinities between 1.6 to 9.4 eq/m 3 (80 to 470 ppm as CaCO 3 ) were used, with a water analysis based on an averaged composition from a number of European sites. These waters cover the ranges found from the Nordic region (low TDS), through those in Central Europe (medium) to Southern Europe (high) The ratio of sodium to total cations was taken at 30% (the range was found to vary between 10 and 50%) and % alkalinity at 50% (range between 20 and 80%) for all salinities. Silica was set at 10 ppm throughout. The chemical and utility costs in Table 1 were used to obtain a base case cost of producing water. The chemical and power costs were then changed in a sensitivity study to determine the effect on the overall economics of the different systems. Two different types of water treatment systems were used : ion exchange only () and reverse osmosis followed by ion exchange polishing mixed beds (). The systems were sized to continuously produce mixed bed quality water (<1 µs/cm and 10 ppb SiO 2 )at flow rates of 50 and 200 m 3 /hour net. Operating costs include chemicals, power, labour and maintenance, together with water and waste water, which are an increasing consideration in water treatment economics 2,3. Surface water was used for both size plants with pretreatment consisting of flocculation,clarification and sand filtration.acid and antiscalant were dosed prior to RO and 5 micron filters were used. The assumptions for the cost analysis are given in Table 1. NOTE : This paper is taken from Ion Exchange Developments and Applications p58-66 Proceedings of IEX 96 edited by J.A.Greig and Published in 1996 by the Royal Society of Chemistry.The copyright belongs to the SCI.

2 Water Analysis, eq/m 3 Total 1.6 eq/m eq/m eq/m eq/m 3 Ca Mg Na HCO SO Cl SiO 2 (ppm) Temp (C) ph Costs : Energy 6.5 p/kwh Sulphuric acid 52 /te Steam 9.5 /te Scale inhibitor 3080 /te Caustic soda 195 /te 100% Lime 65 /te Feed water 4 p/m 3 Deprecation of Capital 15 years System Sizes 50 m 3 /h and 200 m 3 /h System Operating Rate 360 days/year Product Water Quality ex Mixed Bed <10 ppb Na,<10 ppb SiO 2 and <0.1 µs/cm Table 1 : Basis and Assumptions for Cost Analysis Identical water storage facilities were assumed for all systems,with 12 hour treated water storage for the 50 m 3 /h case and 5 hours for the 200 m 3 /h case.the cost of neutralising the waste water was included together with the disposal costs of the waste effluent. Labour costs were not assumed to vary across the options studied are are based on 1.5 manyears at 18,000/manyear.Annual maintenance costs were estimated at 3% of the capital cost of the plant. The capital number used includes the cost of an installed plant to an end-user but does not take into account the cost of land, buildings or taxes.the capital was depreciated linearly over 15 years. Three bed ion exchange system: The basis for the plant design is given in Table 2. Operational Sequence Preatreatment Specification Flocculation/clarifier Sand filtration Demineralised Water Train Cation resin unit Degasifier Layered bed anion unit Mixed bed Demineralised Water Storage Waste Neutralisation Neutralise to ph 7 Regeneration Cation regeneration Counterflow H 2 SO 4 or HCl Anion regeneration Time between regens Uniform size (UPS) gel strong acid cation UPS weak MP/strong base gel anion UPS gel strong acid cation/strong base anion Counterflow NaOH 24 hours eq/m 3 case 12 hours & 6.4 eq/m 3 cases 8 hours eq/m 3 case Regeneration time 2 hours Resin Life Cation resin 8 years Anion resin 4 years Table 2: Bases and Assumptions for Cost Analysis for System

3 The plant is a packed bed counter-flow regenerated design consisting of 2 x 100% streams with cation-degasser-layered bed anion-mixed bed polishers containing uniform particle sized resins.details of layout, vessel sizing, resin volumes and operating conditions are given in Fig.1. SAC D E G A S WBA SBA MB SAC/SBA Storage 50 m 3 /h Cation Anion Layered Bed Mixed Bed Ionic Load SAC WBA SBA Cycle SAC SBA Cycle Volume Volume Volume Time Volume Volume Time liters liters liters hours liters liters hours 1.6 eq/m eq/m eq/m eq/m m 3 /h Ionic Load SAC WBA SBA Cycle SAC SBA Cycle Volume Volume Volume Time Volume Volume Time liters liters liters hours liters liters hours 1.6 eq/m eq/m eq/m eq/m Figure 1 : Layout for System Service run-lengths of 24, 12, 12 and 8 hours were taken for the four different salinities and reflect typical practices in the respective European regions.for the low salinity feeds, silica loading is the limiting factor for the anion sizing and a maximum of 10g SiO 2 /l resin end-ofcycle was set. Warm caustic was also used for the 1.6 and 3.2 eq/m 3 feeds due to the higher silica loads. Mixed bed sizing was set according to specific flow conditions and a silica limit of 1.5g/l resin and is constant for a given flow rate. Regeneration with both 60g/l H 2 SO 4 and 50g/l HCl were considered, but it was found that the cost per unit of treated water were similar for both acid regenerants, as the increased chemical efficiency of HCl and lower resin inventories are off-set by the higher cost of the chemical. Only the economics using H 2 SO 4 are therefore reported. Reverse osmosis/mixed bed system: The basis for the RO/ plant design is given in Table 3. Operational Sequence System Pretreatment Flocculation/clarifier Sand filters Specification RO Train Pretreatment Acid & antiscalant addition 5-Micron cartridge filter RO Membranes Type Thin film composite, spiral wound Life 3 years Recovery 80% in two stages Feed pressure 14 Bar (200 psi) Degasifier Ion Exchange Polishing Mixed-bed UPS Strong acid cation/strong base anion - gel Waste Neutralisation Neutralise to ph 7 Table 3: Bases and Assumptions of Cost Analysis for System

4 This consists of 1 x 100% line with RO-degasser-mixed bed polisher for 50 m 3 /hour and 2 x 50% lines for 200 m 3 /hour. A system recovery of 80% was used in every case. Details of layout, numbers of elements and operating conditions are given in Fig.2. The mixed bed design is the same as for the system: only the service run-length is shorter due to the higher expected average silica leakage from the RO compared to the layered bed anion. Cart. Filter Array 1 Array 2 D E G A S RO Storage MB SAC/SBA Storage Concentrate Thin Film Composite RO System 50 m 3 /h Array 1 Array 2 Ionic Load PV Total.Elements PV Total Elements UPS Resin Mixed Beds Feed SAC SBA Cycle Pressure Volume Volume Time Bar liters liters hours 1.6 eq/m eq/m eq/m eq/m m 3 /h 1.6 eq/m eq/m eq/m eq/m Figure 2: Layout for RO-Mixed Bed Ion Plants Capital estimates for both systems at each salinity and flow rate are summarised in Table 4.

5 K Sterling Feed Water (eq/m3) Plant 50 m3/h m3/h Plant 50 m3/h m3/h Table 4: Purchased Equipment Capital Estimates 6 These estimates are derived from the individual equipment component costs used to make up the plant. The capital cost of the plant increases with water salinity, due mainly to larger vessel sizes. The RO capital cost is higher than, due in part to pretreatment requirements, but is almost independent of water salinity. 3 RESULTS AND DISCUSSION Base cases: The costs to produce water at each salinity and flow rate are given in Table 5 for the ion exchange base case and in Table 6 for RO/. Cost of Treated Water p/m 3 50 m3/h 200 m3/h Feed eq/m Operating Costs Chemicals Sulphuric Acid Caustic Soda Lime Energy Resin replacement Raw water/effluent Labour Maintenance Depreciation (15 years) Total Cost Table 5: Cost to Produce Water for System The results of the base case calculation given in Table 5 indicate that the cost to produce water using only ion exchange increases with feed TDS as expected,principally due to regenerant chemical costs. The effect of increasing plant size is to lower the cost to produce water as the regenerant costs increase proportionately but capital, raw water, labour and maintenance costs are relatively lower for the larger plant. The costs for vary between p/m 3 at 50 m 3 /hour and p/m 3 at 200 m 3 /hour. At 50 m 3 /hour, operating costs account for ~70% of the total cost with regenerants, raw water,labour and maintenance making the most significant contributions. At 200 m 3 /hour, operating costs increase to ~80%.

6 Cost of Treated Water p/m 3 50 m3/h 200 m3/h Feed eq/m Operating Costs Chemicals Sulphuric Acid Caustic Soda Lime Antiscalant Energy Resin replacement Membrane replacement Raw water/effluent Labour Maintenance Depreciation (15 years) Total Cost Table 6: Cost to Produce Water for System Table 6 shows that the cost of producing water using RO/ is also dependent on feed TDS, but much less so than for the system. This is due to the fact that the main cost contributors (power, water, labour, maintenance and capital) are relatively constant over the water salinity range considered. RO/ costs are around p/m 3 at 50 m 3 /hour and p/m 3 at 200 m 3 /hour. Operating costs are 72-80% of the total cost for the two plant sizes. The total costs to produce water are plotted against feed salinity in Fig.3 for 50 m 3 /hour and in Fig.4 for the 200 m 3 /hour base cases Figure 3 Cost of Treated Water 50 m 3 /h Plant - Base Case Figure 4: Cost of Treated Water 200 m 3 /h Plant - Base Case The break-even point for the two technologies is at 8 eq/m 3 TDS for the 50 m 3 /h case and 7 eq/m 3 for 200 m 3 /h.it should be emphasised that these break-even points are based on the assumptions specified in the base case given in Table 1.A sensitivity study was therefore ran to assess the effect of the changes in the cost of water,power and chemicals on the economics. Sensitivity study: The sensitivity of power costs is shown in Fig.5 for 50 m3 /hour and Fig.6 for the 200 m 3 /hour case. Power costs were varied over the range p/kWh (base case = 6.5p/kWh). For the RO/ system, this resulted in a change in the cost to produce water of +/-2.5p/m 3 compared to the base case, thereby affecting the break-even point with at 50 m 3 /h to 8.0+/-1.5 eq/m 3 and at 200 m 3 /h to 7.0+/-1.2 eq/m 3.

7 Power Cost (KWh) Top -9.8p 9.6 Mid - 6.5p Bot - 3.3p Figure 5 Cost of Treated Water 50 m 3 /h - Power Sensitivity The sensitivity of caustic regenerant price on economics is shown in Fig.7 for the 50 m 3 /hour/plant Caustic Pricing Top /te 9.6 Med /te Bot /te Figure 6 Cost of Treated Water 50 m 3 /h - Caustic Price Sensitivity The effect on RO/ economics is minimal. Over the range /ton 100% NaOH, the break-even point for a 50 m 3 /h plant ranges from eq/m 3. Finally the effect of raw water/effluent cost is shown in Fig.7.The costs range from the lowest - that of that of a surface water intake with little water loss - to an inexpensive borehole water Water Cost Top- 13 p/m3 9.6 Mid p/m3 Bot p/m3 Figure 7 Cost of Treated Water for 50 m 3 /h - Cost of Water Sensitivity Over the water cost range taken the breakeven point for a 50 m 3 /h plant ranges from eq/m 3.If,however, mains water is taken (e.g. 50 p/m 3 ) or the cost of effluent treatment is expensive,the costs of RO/ vs increase markedly and no break-even point is found under 10 eq/m 3

8 Direct comparison of the present results to existing literature data is difficult due to the many different assumptions and parameters used in each study. However, it is possible to make some qualitative comparisons. The base case salinity break-even points of the present study (7-8 eq/m 3 ) are substantially higher than the eq/m 3 ( ppm as CaCO 3 ) values obtained in previous USA studies 1,4 and goes against the trend of declining salinity break-even point observed in recent years 5. This shift to higher salinity can be explained by the economic advantages of the packed bed counter-flow regenerated system over co-flow systems and the higher power costs in Europe compared with the USA,both of which are unfavourable to RO. A similar conclusion was made in a European study where a break-even point of ~10 eq/m 3 was found 3. 4 CONCLUSIONS This economic evaluation considers the major factors contributing to the total cost of treated water by RO/ and alone in Europe. The effect of system size and the latest technology in both resins and membranes has been included. Major conclusions that can be drawn from this study apply to new water treatment plants and are summarised below:- 1. The break-even point above which it is more economical to use RO/ versus alone is 7-8 eq/m 3 ( ppm as CaCO3 TDS). This is significantly higher than comparable USA studies and reflects developments in packed bed counter-flow regenerated systems and higher power costs in Europe. 2. The break-even points derived in this study represents water salinities that are commonly encountered in Europe. For many new plant projects a clear choice of one technology often cannot therefore be made on purely economical grounds but other considerations and local factors must be taken into account, such as water availability, type and disposal regulations, existing installed capital and end-user technology preference or familiarity. 3. Although capital has a significant effect on the total cost of water for all options considered, operating costs represent the major portion at 70-80% of the total. 4. This study considers mainly surface water and low cost discharge of effluent from the water treatment plant into a river.unless the concentrate from an RO plant can be used elsewhere on site,increasing water costs and disposal costs will very likely penalise RO over ion-exchange. 5. Chemical costs for and electrical power costs for RO are the most important operating expenses and those that need to be carefully considered in the decision for a new plant. 6. This study has identified a number of factors that impact the economics and hence market position of versus RO technology now and in the future. Relative familiarity with technology combined with the economics shown in this paper could limit the growth of RO technology in industrial water applications.the development of lower energy membrane modules and a trend away from the use of regenerant chemicals could change this scenario. REFERENCES 1 see for example, A.F. Ashoff, UltraPure Water, July/August 1995 p VGB-Kraftwerkstechnik GmbH literature, May K. Grethe and C. Beltle, "Power station make-up water using RO and ion exchange for demineralisation" Steinmuellertagung S. Beardsley, S. Coker and S. Whipple, Watertech Expo '94, 9. Nov S. Whipple, E. Ebach and S. Beardsley, UltraPure Water, October Capital estimates provided by Satec Ltd.,Memcore Ltd.,ERG Plc. and Steinmüller GmbH