Dear Reader! The publication of this document does not pursue for itself any commercial benefit. But such documents promote the most rapid professional and spiritual growth of readers and are advertising of paper editions of such documents. You can download (copy) this file for educational purposes only. PLEASE NOTE that ANY COMMERCIAL USE of this file IS STRICTLY FORBIDDEN You should delete it immediately after your reading is completed. Under international legislation you are responsible for prorer use of this file. All copywrite reserved by owner.Membrane Technology in Water and Wastewater TreatmentMembrane Technology in Water and Wastewater Treatment Edited by P. Hillis Northwest Water Ltd, UK Rc 6D SCI‘The Proceedings of the Conference being held by the Water Chemistry Forum of the Royal Society of Chemistry and the European Desalination Society on Membrane Technology in ‘Water and Wastewater Treatment a the Univesity of Lancaster, UK, on 27-29 March 2000. Special Publication No. 249 ISBN 0-85404-800.6, ‘A catalogue record for this book is available from the British Library © The Royal Society of Chemistry 2000 All rights reserved. ‘Apart from any fir dealing for the purpose of research or private study, or criticism or review as permitted under the terms of the UK Copyright, Designs and Patents Act, 1988, ‘this publication may not be reproduced, stored or transmitted, in any form or by any means, without the prior permission in writing of The Royal Society of Chemisty, or in the case of, reprographic reproduction only in accordance with the terms ofthe licences issued by the Copyright Licensing Agency in the UK, or in accordance with the terms of he licences issued by the appropriate Reproduction Rights Organization outside the UK. Enquiries ‘concerning reproduction outside the terms stated here should be sent to The Royal Society of (Chemistry atthe address printed on this page. Published by The Royal Society of Chemistry, ‘Thomas Graham House, Science Park, Milton Road, Cambridge CB4 OWF, UK For further information see our web ste at www.s.0rgPreface “These proceedings contain the papers presented at the Intemational Conference on ‘Membrane Technology in Water and Wastewater Treatment held at Lancaster University from the 27th-29th March 2000, The contents are set out in the order of their presentation at the conference. The meeting was organised and sponsored by the Royal Society of Chemistry Water Chemistry Forum, the European Desalination Society and the SCI Separation Science and Technology Group; with the support of the Institution of Chemical Engineers, the Chartered Institution of Water and Environmental Management and the UK Drinking Water Inspectorate ‘The conference was designed to bring together membrane experts from around the ‘world to the United Kingdom to share their experiences of the continuing development of ‘membrane technology in the water industry. The aim of this event was to promote the use of membrane technology using MF, UF, NF and RO, for a range of water treatment applications. The conference was organised to benefit the following: ‘water companies industrial operators plant designers and consultants R&D organisations ‘membrane equipment manufacturers chemical suppliers legislative bodies ‘The mecting focused on all aspects of membrane technology for water and wastewater applications. It covered the treatment of ground and surface water, backwashwater, seawater and industrial and domestic wastewaters. ‘This book will serve as a useful reference to the status of membrane technology at the end of the second millennium and give an insight into the future of membranes as we ‘enter the third millennium. Peter Hills Editor and Conference ChairmanContents Case Studies Keynote Lecture: Membrane Case Studies, Past, Present and Future JS. Taylor and S.J. Duranceau ‘Sea Water Reverse Osmosis ~ The Largest Plant in British Waters N. Marsh, J. Howard, F. Finlayson and S. Rybar Drinking Water Sources in Kuwait M. Safar and ¥. Al-Wazzan ‘Nanofiltration for Colour Removal ~7 Years Operational Experience in Scotland E. Irvine, AB.F. Grose, D. Welch and A, Donn ‘Ultrafiltration for 90 MLD Cryptosporidium. and Giardia-free Drinking Water. ‘A Case Study of the Yorkshire Water Keldgate Plant FNM, Knops and B. Franklin Application of a New Generation Microfiltrtion Process for Large Scale Water and Wastewater Treatment WT. Johnson and A. Patterson Water Quality and Treatment ‘The UK System of Approval of Products Used in Contact with Drinking Water T. Ogunbiyi Immersed Membranes for Drinking Water Production P. Coté, C. Gingerich and U. Mende Phosphate and Iron Removal from Seepage and Surface Water by Microfiltration JAMH. Hofman, N.C. Wortel, ET Baars and J.P. van der Hoek Reuse of Filter Backwash Water as a Source for Drinking Water Production: Piloting and Implementation of a Full-scale Ultrafiltration Plant ‘A. Briigger, K. Vossenkaul, T. Melin, R. Rautenbach, B. Golling, U, Jacobs and P. Ohlenforst 28 2 41 49 37 1 78 85vil ‘Membrane Technology in Water and Wastewater Treament Improved Performance of Drinking Water Microfiltration with Hybrid Particle Pre-treatment T. Carroll and N. Booker River Trent On Tap — Comparison of Conventional and Membrane Treatment Processes BE, Drage, LE. Upton, P. Holden and J.Q. Marchant ‘The Use of Electrodialysis at Amsterdam Water Supply J.P. van der Hock, LA.M.H. Hofman, P.A.C. Bonné and D.O. Riinbende Applying Electrodialysis (EDR) Technology to Underground Water Treatment E.Sgarbi Nanofiltration for Drinking Water Treatment from a Eutrophied Lake in Taiwan HoH, Yeh, SH, Lin, SJ. Kao and G.T. Wang Fouling and Cleaning Keynote Lecture: Membranes and Microorganisms ~ Love at First Sight and the Consequences H.-C. Flemming Optimising Membrane Performance — Practical Experiences LY. Dudley, F. del Vigo Pisano and M. Fazel Fouling Characteristics of Membrane Filtration in Membrane Bioreactors ‘MAH. Thomas, S.J. Judd and J. Murrer Cleaning of Membranes in Water and Wastewater Applications R Krack Water Reuse Water Reuse for the Next Millennium —Membrane Treatment atthe Millennium Dome LH, Khow, A.J. Smith, A. Rachwal, A. Donn and C.V. Meadowcroft ‘Wastewater Reclamation Case Studies, the Benefits of Outsourced Membrane Systems D. Threlfall Comparison between Different Out-to-in Filtration MF/UF Membranes for the Re-use of Biologically Treated Wastewater Effluent E. Van Houtte, J. Verbauwhede, F. Vanlerberghe and J. Cabooter 93 100 108 18 128 139 150 158 166 175 186 190Contents Industrial Applications ‘Sulphate Removal Membrane Technology: Application to the Janice Field GH. Mellor, R.C.W. Weston, G.F. Bavister and A, White ‘SASOL’s Experience in the Desalination and Re-use of Acid Mine Drainage and ‘Ash Water 1G. Niewwenhuis, G.H. Du Plessis, M.P. Augustyn, B. Steytler, AJ. Viljoen and LW. Van Der Merwe Recovery of Wool Scouring Effluent Utilising Membrane Bioreator (MBR) ‘Technology as Part of the Activated Sludge System followed by Two-stage Reverse Osmosis (RO) Membrane Concentration AR. Bennett Performance on a Real Industral Effluent using a ZenoGem® MBR D. Mallon, F. Steen and K. Brindle ‘Membrane Technology in Wood, Pulp and Paper Industries J. Wagner Case Studies of Wastewater Re-use for Petrochemical, Power and Paper Industry B, Durham Posters Practical Experience with a Membrane Bioreactor for Wastewater Treatment- semi-cross-flow Ultrafiltration S. Geibler, K. Vossenkaul, Th. Metin, P. Ohle, E. Brands and M. Dokmann ‘Treating Highly Coloured Waters: Design Innovations and Implications AB. Grose, D. Welch and E. Irvine ‘Treatment of Leachate by the MBR Process (Membrane Bioreactor) ‘AU. Robinson Operation of a Zero Discharge Wood Pulp Effluent Treatment Plant G. Bateman Integration of Maintenance and Operation into the Design of Reverse Osmosis ‘Membrane Networks HJ. See, VS. Vassiliadis and D1. Wilson Microfiltrtion and Reverse Osmosis of Knostrop Final Effluent ‘M. Barton 201 au 219 26 233 241 251 253 255 287 258 260x Membrane Technology in Water and Wastewater Treatment Modelling Temperature and Concentration Polarisation in Ultrafiltration of Non- Newtonian Fluid under Non-isothermal Conditions SP. Agashichev ‘Novel Methods of Hollow Fibre Membrane Integrity Monitoring S. Williams, A.J. Merry and C.V. Meadowcroft ‘Comparison of Chemical Pretreatment Methods for Nanofiltation of Cold, Soft and Humic Waters J. Yli-Kuivila, R, Litkanen and R, Laukkanen In-situ Ultrasonic Measurement of Fouling and Cleaning Processes in Spiral- wound Membrane Modules G.-Y. Chai, A.R. Greenberg and W.B. Krantz A Novel Way to Treat Textile Wastewater with Nanofiltration and Adsorption Th, Melin and L. Eilers 261 268,Case StudiesMEMBRANE CASE STUDIES, PAST PRESENT AND FUTURE J.8, Taylor, PLD., PE. S.J. Duranceau, PhD., P. ‘Alex Alexander Professor of Engineering Dir. of Water Quality and Treatment Civil and Environmental Engineering Department Boyle Engineering Corporation University of Central Florida 320 East South Street Orlando, FL 32816 Orlando, FL 32801 USA, USA, 1 INTRODUCTION Membrane case studies have played an essential part in the development of membrane technology for drinking water treatment. Continuing advances in regulatory constraints and aesthetie criteria for consumer water quality have driven the water community to seek new technologies, which meet these criteria. Foremost among regulatory constrains are disinfection requirements, disinfection by-product and corrosion regulations. Consumers hhave become aware of regulatory violation through mandated public notification, and they have always been aware of the appearance, taste and odour of drinking water. Before the requirement of advanced technologies to meet higher water quality regulations, design, construction and successful operation of conventional water plants was well established and did not require pilot studies. However, such is not the case today. Pilot or case studies in all developed countries are establishing productivity, water quality and ‘estimated cost for advanced treatment process construction and operation. 2. OVERVIEW OF MEMBRANE PROCESSES Understanding membrane application requires understanding of the characteristics of drinking water membrane processes. Reverse osmosis (RO), nanofiltration (NF), electro- lysis reversal (EDR), ultrfitration (UF) and microfiltration (MF) are the membrane processes, which have application to drinking water'', Combinations of membrane processes with other processes have become Known as integrated membrane systems. Although a conventional NF process consists ofa pre-treatment and posttreatment process before and after the NF, which could be described as integrated, this is described as conventional. The coupling of a MF and a NF or coagulation, sedimentation and filtration with a NF are accepted examples of IMS's. The basic characteristics of these processes ‘are shown in Table 1. Although many factors affect the solute separation by these process, a general understanding of drinking water application can be achieved by associating minimum size of solute rejection with membrane process and regulated contaminate.4 ‘Membrane Technology in Water and Wastewater Treasment 3 REGULATIONS ‘The US water quality requirements determine membrane selection. Many of the regulatory constraints for drinking water can be related to control of inorganic, organic or pathogenic solutes in the finished product. The specific application of membrane processes to drinking water applications is shown in simplified format in Table 2. The ‘word Yes indicates the membrane process can remove significant amounts of contaminate specified by the rule, and No indicates the membrane process can not remove the regulated contaminate Table 1 Characteristics of Membrane Processes Process Mechanism Exclusion Regulated Solutes, Pathogens Organics Tnorganies EDR c 0.0001 um None None All RO SD 0.0001 1m C,B,V BPs, SOCs Al NF s.D 0.001 ym C,B,V BPs, SOCs__ All UF s 0.001ym —C,B,V None None MF 8 01pm — CB None None Mechanism: C=charge, S=sieving, Diffusion Pathogens: (Ceeysts, Bbacteria, V=viruses Organics: DBPsdisinfection by-product precursors, SOCs-Synthetic Organic ‘Compounds Examples of community water quality objectives are total hardness, taste & odour, and. colour. Some community water quality objectives may be classified as secondary standards or those standards, which do not affect consumer health. Community water ‘quality objectives are shown in simplified form in Table 3. Once the treatment objectives are known, potential membrane systems can be determined for meeting these goals. Some ‘general statements can be made regarding current treatment concerns: 1, Diffusion controlled membranes (RO & NF) are required for control of inorganic contaminates such as total dissolved solids (TDS), total hardness (TH), chlorides, cfc, and DBP precursors 2. Charge controlled membranes (EDR) ean remove TDS, TH, chlorides ete 3. Size exclusion controlled membranes can control particles, turbidity and cystsCase Sales s ‘Table 2 Summary of Membrane Process Applications for Drinking Water Regulations us “Membrane Process Regulation/Rule EDR RO NE UF MF SWIRESWIR | no yes yes ye yes oR » yes ye ves yes Ter to ve ye no no ioc es yes 1 no 0 SoC 70 Ys = 38 = Radionuclides yes(-Rn) |} yesGRn) | yes(-Rn) | no no DBPR no yes 1 no no GwoR m0 yes 1 yes 1 ‘Arsenio yes s yes 0 no Sulphates vs ys 1 no m0 SWTR -Surface Water Treatment Rule Enhanced Surface Water Treatment Rule “Coliform Rule Lead and Copper Rule “Inorganic Rule (Phases 1, 4, ¥) > “Synthetic Organi Chemicals (Base Neutrals and tractables) DBPR -Disinfection By-Products Rule GWDR-Groundwater Disinfection Rule ‘Table 3 Summary of Membrane Process Applications for Drinking Water Regulations Parameter “Membrane Process TED RO NF UF WF ps Yes yes yes 0 no TH Yes yes yes no no TRO No yes yes no no Toc No yes ves No no Colour No yes yes 0 no Fe& Mn No. yes yes No. no 4 REQUIREMENTS FOR CASE STUDIES 4.1 Documentation Case studies are investigations of singular or combined processes with specific goals for production, water quality and cost. Accurate project documentation is required to develop and report pilot production and water quality. Documentation of laboratory work should bbe associated with a change of custody and quality control. Determination of precision6 ‘Membrane Technology in Water and Wastewater Treatment and accuracy of analytical samples in the laboratory or field is essential for meaningful interpretation of reslls. A traceable paper trail of sample analysis should be described in project documentation and referenced in a projec report. Work and sampling logs should be used for documentation of fieldwork. Work or operational logs have daily entries that deseribe taking operational data, mechanical repair or any other activity that is assoc with operation of pilot systems. Sample logs are separate from work logs in that specific ata is recorded. Typical examples of field data include pH, pressure, flow, and temperature. 42 Productivity Productivity is essential to any water treatment facility. Productivity is affected by design of the membrane process and fouling. Designers can select membranes for specific treatment characteristics. Once selected, a designer can select operating conditions for that membrane process. A primary consideration affecting productivity is fouling. The four primary mechanisms of fouling are scaling, plugging, adsorption and biological growth. The primary means of controlling fouling by mechanism and unit operation are show in Table 4. General Comments can be made regarding pre-treatment requirements. 1. Scaling control is typically required for all RO/NF membrane systems in either surface or groundwaters and is achieved by acid and/or antiscalent addition. 2. Plugging control is typically required for all RO/NF membrane systems in either surface or groundwaters and is achieved by feed water turbidities and SDI’s less than 0.2 NTU and 2 respectively. 3. Bio-fouling control is typically required for aerobic surface or groundwaters and is achieved by NH;C! or addition of other bactericidal agents 4. Organic fouling can occur in surface water systems with TOC > 3-6 mg/L. and is typically reduced by coagulation, sedimentation and filtration. However, the significance of organic fouling is not known. ‘Table 4 Fouling Control by Pre-treatment System ‘Mechanism _| Process ‘IAS MEIUF CSF NHCl_AOC Removal Scaling | + : a . Plugging |- + a 5 ‘Adsorption. | - + tS is Bio-fouling_|- : + + 43 Evaluation of Membrane Systems Performance of membrane units can be measured inthe field or laboratory by functioning ‘membrane systems in the form of (a) small cells, (b) bench units or () pilot plants. Small cells have historically been used to test film characteristics and have not yet been shown to be representative of actual production. A testing protocol for small cells has been published’. There results indicated that the Phase II MCLs for DBPs could be met in four of the five waters tested, the rate of productivity decline was exceptionally high and that DOC was diffusion controlled. While the DBP results were comparable with previous studies, the productivity results were not. ‘Their results have shown that small cells areStudies 7 suited for membrane screening quality studies but are not adequate for productivity assessment. ‘Membrane manufacturers use 4°x40” elements, .5°x40” elements or 2.5°x20” elements in pilot studies for plant scale up. The preliminary results of the information collection rule (ICR) were presented at the 1999 WQTC in Tampa, FL? In excess of forty ‘membrane studies have been done through the United States to comply with the ICR requirements. The studies typically consist of flat sheet laboratory work; single element field work, multi-staged pilot plants and full scale plant documentation. ‘The studies have shown that DBP precursor removal requirements were exceeded for haloacetic acids for either stage | or 2 MCLs, but thatthe ninety-fifth percentile of THMs exceeded the stage 2-MCL (40 ug/L). ‘The higher THMs were attributed to high bromide concentrations (30 ug/L) and variable bromide rejection (70 % to 0%) by membranes. “Membrane case studies require that mass transport of water and solutes through the ‘membranes be described quantitatively. The equations used to describe flow through single element are shown in (1) through (7) with reference to the membrane element shown in Figure 1. PRETREATED PERMEATE) FEEDWATER aor > aan (CONCENTRATE(C) ac Pore Figure 1. Basie Diagram of Mass Transport in a Membrane ‘Mass transport in pressure driven membrane processes can be described as convection or diffusion controlled, Models for deseribing mass transfer in membrane systems have been presented by several investgators.""“®, These equations can be used to predict permeate water quality of any membrane system. Although membrane systems have been shown produce water quality that exceeds most regulatory requirements, models as shown in equation (8) or (9) are essential for predicting the cost and performance of membrane systems from pilot plant data. Equation (8) is used to describe diffusion controlled mass ‘transfer, which includes inorganics, ic, alkalinity, hardness, TDS, sodium, chlorides, ete. Equation (9) is used to describe sieving controlled mass transfer, which includes TOC, DBP precursors, most SOCs and organics in general However, TOC has been shown as diffusion controlled in a surface water application at Tampa FL for CA and CTF membranes". The TOC from the CTF membrane was so low that no limit for lux and recovery was projected for precursor control, however flux and recovery were limited for the CA membranes as TOC rejection was significantly less. Existing research has clearly shown consistent production and concentrate disposal are the limiting constraints for membrane systems. Concentrate disposal is essentially a regulatory problem but consistent production is a research and development problem. ‘There is simply inadequate information on the fouling of membranes by moderate organic surface waters8 ‘Membrane Technology in Water and Wastewater Treaament Reference Equations: o Fig = Ke(ar-ant)=2F @ 2, R= 3} : ® Q,=9.+9, @) O/C, = OL. + 2,6, o & AC “Concentration gradient OWL) (CCL) Gy = Feed stream solute concentration anny Cc = Concentrate stream solute ‘concentration (M/13) Gp = Permeate stream solute concentration aw) C3 = Solute concentration atthe membrane surface (M13) k= Diffusion coefficient from the surface to the bulk (13/12 AP = Pressure gradient (1), (PpePo)2- Pp) ‘An = Osmotic pressure (L) (y+ e)/2-Ap) Ky = Solvent mass transfer coefficient Cup ©) 2, . 0 a mM kice et g Tya-W) , 6 © 24 2r-R ” = oc, o y= Water fl (191029 Fs = Solute flux (M/L2) Solute mass transfer coefficient (L/t) Feed stream flow (34) ‘oncentrate stream flow (L3) Recycle ratio 4 = Sieving pass coefficient Fouling indices can be used to indirectly estimate pre-treatment requirements for ‘membrane systems®, These investigations have shown that fouling indices do not statistically correlate to the rate of productivity decline in diffusion controlled membrane systems. Regardless fouling indices can be used to get a crude esti The silt density index (SDD), modified fouling index (MFI) and ‘treatment requirements ‘of membrane pre- the mini-plugging factor index (MPFI) are shown in equations (10), (11) and (12).Case Sudles 9 mle) sor=—; (10) mri~(ovy ay Mrri~ 2 (2) Where: time to collect inital 500 ml; y= time to collet final 500 ml; T= time between t, and i,(15 min): Q=flow; V=volume ‘These indices that can be used to predict the required pre-treatment for RO or NF ‘membrane processes. These indices are determined by monitoring volume collected verses time of filtration through a 0.45 jm filter under a constant pressure of 30 psi Approximate values of RO and NF indices are shown in Table 5. Table 5 Site Raw Water Quality SDS-HAA6 (ug/L) 2032 234 3,561 100. >800 Alkalinity: (mg/L as CaCO;) 67 148 110 120° 100 Total Hardness (mg/L asCaCOs) 90 215. 130120. 150 188 (mgt) 47° 41 NA OS ps (mpl) 151 332430 300 400 ‘SPC (cfw100 ml) 1,816 ND ND ND ND MIB (g/L) <20 ND ND ND ND Geosmin (ng/L) 84 ND ND ND ND Iron (mg/L) 035 0.0317 0.82 0.30.3 Bromide (mg/L) 0.063 007 ND <0 <0. Manganese (mg/L) 0.029 <0.002 NA ND ND Chloride (mg/l) NG 335 100 80 40 Sulphate (mg/L) 8s 605 2 80 20 ica (mg/L) 104 63 ND 15 15 trate (mg/L) 016 153 NA S < Pesticides ___(_e/L) 2m = ->270=«>270=>270 caLP xX 2270 270-3270 2270 cap X 22m = ->270- 270 >270 CALP XX >270 On-going - : cap X 2270 >270- 270 >270 LFCI x S No fouling LFCt MMF = LFCL xX. Lrct IMF oLFCL XX LFCI Lrcl__X 7 = Ftional experiment scheduled Monochloramine dose was Tow and pre- existing levels of biogrowth are thought 10 have reduced effectiveness of this experiment ‘Table 15 TWD Plant and Membrane Systems Water Quality, TOCTTHM TIAA TurbidiyGecamin MIB TON—TDS mg wgyl ug NTU ag —ngL gil PlatCSF 32 «9050.14. NANA 10271 CSF-CALP 05 57 19 0.08 = NIA NA 690 ZMFCALP 08 60 57 009 32 <10 29 82 CZME-CALP 06 35 23° O1 13° 49 «13 205 MMF-CALP 05 34 23 009 16 92 24 203 CSFLFCl <05 33 4 009 <10 <10 15 26 ZMFLFCL <0S 6 7 009 <10 <10.24 35 MMF-LFCL <05 4 27 008 <20 —<10 22,12 63 East St. Louis, Minos ‘Six membrane systems have been investigated at East St. Louis Illinois. Productivity results for the ESL systems consist of CSF water fed into three separate NFs. Two TFC NFs and one CA NFs have been investigated to date. Membrane systems consisting of CA UF membranes are being investigated at ESL but CA UF and NF membranes have bbeen destroyed due to oxidation. No evidence of chemical oxidation has been observed. Both the NF and the UF systems employ addition of chlorine during operation, however biological oxidation is possible if material is not removed from the CA membrane surface. Excessive silt was observed on the CA NF, autopsy of the CA UF membrane is not complete.Case Sudles ” ‘The TFC membranes are shown to foul at a greater rate than the CA membranes. ‘Scaling significantly reduced productivity. Reducing the feed pH to 5, which may have caused aluminium precipitation in the NFs, controlled calcium carbonate scaling. Only CSF pre-treatment results have been reported today. As shown in Table 16, these results indicate that NF cleaning frequency decreases as NF flux and recovery increase, and that ‘CA NF foul less than TFC NFs does. These results also indicate that CA membranes are susceptible to degradation and not as durable as TFC membranes. Table 16 ESL Cleaning frequency by System and NF Flux and Recovery TNF Cleaning Frequency (days) Syste ux (ge) recovery(%) 10/55 10/75 15/55 15/75" csr-TFCl 30 o 8 No fouling CSF-TFC2 40 wo 7 ‘No fouling CSFCA TBD 2 1 rep rece 9 30 ‘The ESL water quality is shown in Table 17. The ESL project is currently in progress however the data indicates that the CA membranes do not reject TOC as well as the TFC membranes. CA NFs may produce waters that are more likely to exceed stage Il DBP MCLs than TFC NFs may. Pathogen rejection data was collected from the ESI. investigation and used to ereate the data shown in Table 18. As each ESL pilot unit was challenged separately by cysts, ‘bacteria and phage, measured rejection of the organisms was possible. This data was coupled assuming 2.5, 1.5 and 0.5 log rejection of eysts, bacteria and phage by CSF and compared to pathogen rejection by membrane system and CSF. This data indicates pathogen rejection by membranes is superior to conventional CSF. ‘Table 17 Finished ESL Plant and Water Quality ——____ESE-Plant_CSF-TFCI_CSF-TFC2_CSF-CA_ STOO gm Seger 0.3 en 2 ae ae THMSDS ug/L 30° 2 31 48 HAASDS ug/L 20 20 35 50 Turbidity NTU 0.13 ___0.05__0.06__0.05 "Monochloramiine addition controlled DBP formation Table 18 Calculated Pathogen Log Rejection for ESL Parameter CSF CSF CSF UF UF CUR NF-l NF-2 NF-3 NF-l NF-2 NF-3 Gas 70469" TBD TBD TBD Bacteria 5.5" 59 18 49 17 Virus 48° 47 Tl 46 70 NPDOC 1.0 Tt 10) 0.7 Inorganics 1 03 1 03 03 ¥ Assumes 25, 7: csF 73 log removal for Cysis, Bacteria and Viruses respectively by8 ‘Membrane Technology in Water and Wastewater Treament 64 City of Melbourne, Florida ‘The Melboume membrane systems consisted of two different TFCs NFs, which were preceded by parallel MF, CSF and GAC pre-treatment. The conventional NFs systems consisting of A/AS and static MF and NF could not operate for more than 2 hours due to plugging. Comparison of cleaning frequencies indicates thatthe alum CSF pre-treatment ‘produced the least fouling water or @ 20-day runtime, BAC pre-treatment was utilized to remove substrate and control biological fouling. ‘The short BAC pre-treatment run was due to high turbidity and plugging of the NF membranes. However this data indicates that BAC did not effectively control biofouling and additional filtration following BAC would be necessary to control NF fouling. ‘Assuming MF pre-treatment reduced plugging, an assumption can be made that the increased runtime following CSF was due to a reduction in organic or biofouling. CSF does partially reduce bacteria loading onto the NF, but alum CSF water is highly biologically active. Although an exact determination of fouling mechanism is not possible, it can be stated that higher TOCs from groundwaters given to conventional NFs application run for six months to one year before cleaning. Such sources are anacrobic and have a reduced and different bacterial population. Additionally in the limited data reported in these studies, there has been no instance where addition of a bactericidal agent id not increase runtime. Table 19 Melbourne Chemical Cleaning Frequencies GACBF -0.01283 2 Assimes intial Kw of 0.2 efpst ‘Table 20 Melbourne Plant and Water Quality Summary Parameter Plant _Permeate ————GACE MEE ACSFE GACF ACSE-F— Doc mg/L 10s 02 05 OS 02 THMFP ug 360 18 = 3145 20 HAA(S)FP ug/l. 262 5 ise ale) 10 Colour cpu 9 <1 4 1 <1 03 01 01 On 02 2428 173 11291461128 460 162 230 «16349 345458 2% 1SCave Sales 9 65 City of Fort Myers, Florida Fort Myers is a full-scale plant, which was evaluated as part of an AWWARE project” ‘The membrane system consists of pumping water from the Caloosahatchee River to rapid infiltration basins (RIBs) for infiltration into a shallow aquifer. Water is pumped from the aquifer to a conventional NF plant, which employs acid and anti-scalent addition to control scaling. The production results forthe Fort Myers system are reported from 1993- 1997. ‘The utilisation of the available membranes for Fort Myers is shown in Figure 3. The plant was initially constructed for 12-MGD capacity and begins operation in 1993 at slightly over 60 % of the available capacity. Capacity in this sense is defined as'a percentage of utilized membranes that are available membranes for operation. The data in Figure 3 shows the utilisation of membranes increased from slightly over sixty percent 0 slightly over eighty percent from 1993 to 1997. sBesssasss. WERAETI ° r a Figure 3. Fort Myers Membrane Utilization Over Time ‘The data in Figure 4 shows the average production from 1993 to 1997. This data shows that the average production decreased from slightly over 7 MGD to slightly more than six MGD from 1993 to 1997. Consequently utilisation increased and production increased from 1993 to 1997. Increasing utilisation of membranes (membrane surface area) and decreasing production is an indication of membrane fouling in an operating facility.20 ‘Membrane Technology In Water and Wastewater Treatment CALENDAR YEAR Figure 4 Fort Myers Production Over Time ‘The Fort Myers systems recovery is shown by stage overtime of operation in Figure 5, These results indicate that frst stage recovery decreased from approximately seventy percent to fifly percent over five years of operation. Sccond stage recovery increased from twenty percent to 30 percent and thid stage recovery remained constant over the same period. The results indicate colloidal deposition and removal in the first stage, which caused extensive fouling of the first stage. The increased resistance in the first stage increased second stage productivity. This observation is similar to first and second productivity relationships when a flow restriction is places on a stage one permeate line, whichis commonly done to prevent excess stage one productivity. Figure 5 Fort Myers Recovery by Stage Over Time ‘The Fort Myers normalised flux and MTC is shown in Figures 6 and 7. Both of these figures clearly show a consistent fouling over time of operation. The mechanism of fouling is plugging. The water quality data in Table 6 for the Fort Myers NF shows iron and sulphides are present in a slightly aerobic source. ‘The Fort Myers wells are less than 30 feet deep and subject to intermittent serobic conditions. Such an environment isCase Sudies a consistent with the chemical precipitation of iron or sulphur, and biological growth. Over five years, there were 161 estimated cleanings of individual banks inthe Fort Myers NF. ‘The mpd fouling and chemical condition ofthe Fort Myers feed water indicates plugging as the primary mechanism of fouling. “The Fort Myers raw, feed and permeate water quality is shown in Table 21. Although the Fort Myers productivity was unsatisfactory, the water quality was good. Only THM were observed to be higher than the Stage II MCL. The NPDOC while below 1 mg/L was high for typical NF permeate. Both NPDOC and THMs may decrease when the productivity problem is solved. eonsoaoshESSR Figure 6 Fort Myers Normalised Fes Over Time vee3 7263 IBA ARMA SOS | ORORS ANOS 11286 SIT Figure 7 Fort Myers MTC Over Time2 ‘Membrane Technology in Water and Wastewater Treatment Table 21. Fort Myers Water Quality ‘Parameter Raw Feed Permeate NPDOC mg/L 2 1s 08 Colour CPU 13 8 5 Ca mg/L 375 8 Mg mgL 4 14 3 HS mg/L. 0 2 «0 Do mg/L 8 0.14 014 Fe mg 03°03 «02 Alkalinity mg/LCaCO; 145 208 42 SO, mg/L. 23 21 159 Bromide mg/L 0.334 0.458 NR Na mg/L aie seal a mg/L 2 60 33 ‘Turbidity NTU 068 0.16 0.03 THM up 1095 735. 69 HAA up 80137718 7 CASE STUDY COSTS Costs for case studies would include planning costs, labour costs, permits, equipment rental and/or purchase, cletrcity, chemicals and laboratory analyses. Not unlike other ‘engineering planning, design and construction costs, there are contingencies that arise and ancillary expenditures that will be incurred during the course of project implementation. The costs expended for the case studies described herein ranged between $250,000 USD ‘and $400,000 USD, and included several funding sources (utility support, government research funds, university matching funds, in-kind donated funds). These costs should be viewed as value engineering and investment costs, as the information derived from performing detailed pilot investigations will impact bottom-line operation and maintenance costs. Also, pilot investigations can also be used to implement training of ‘personnel and other resources prior to construction and operation ofthe full-scale facility. For example, a pilot investigation may find that cleaning frequencies are twice of that predicted, resulting in design of additional pre-treatment facilities (prior to construction) and increased operation budgets atthe onsite of start-up. Ifa detailed pilot investigation ‘nad not been conducted, then sigificantly higher costs (upwards of 20 to 30 percent ofthe ‘original design estimate) would have to be expended to change-order and or perform a new design and construction step afer acceptance of the original construction occurred. In addition, the operation costs would have been projected to be much lower than actual; causing a significant alteration of the total amortised cost for which bonding is typically based upon. The key point here is to say that pilot investigations are mandatory for confirming costs and performance of membrane processes fora site-specific application.Case Studies 2 8 FINDINGS AND CONCLUSIONS 1. RO, NF, EDR, UF and MF are the five membrane processes, which have major drinking water application. Case studies are required to determine design criteria for applications of membranes in drinking water treatment. Membranes are necessary 10 ‘meet the regulatory constraints of increased water quality eriteria; consequently, ‘membranes offer a broad range of regulatory compliance to utilities. 2. Membranes can be effectively used singularly or in combinations of pre-treatment unit ‘operations and conventional NF/RO systems to enhance NF/RO productivity Correlation of membrane process and contamination removal can be made by pore and contaminate size. 3. Plugging, scaling, organic adsorption and biological growth are major fouling mechanisms for membranes, which must be assessed in applications. Site specific pilot investigation is required to determine significance of fouling mechanisms. Basic fouling mechanisms can be assessed by incremental evaluation of combinations of fouling mechanisms. RO'NE fouling can be assessed by observation of MTC decline ‘over time of operation; plugging and scaling must be controlled in every RO/NF application. 4. ROINF fouling in groundwater systems is typically significantly less than in surface ‘water systems although TOC in CSF surface waters is less than conventional RO/NF ‘TOC in groundwater systems. This is consistent with biological fouling of aerobic sources and decreased significance of organic fouling. 5. Organic fouling is a significant fouling mechanism for RO(NF applications but is typically not as significant as biological fouling. 6. Biological growth is a very significant fouling mechanism in every RO/NF application involving aerobic raw water sources (surface waters and shallow wells), and must be directly assessed for RO/NF fouling. Biological and chemical oxidation must be controlled for CA membranes in any treatment application. 7. NF pathogen rejection at ESL by TFC NF exceeded conventional CSF pathogen rejection. Combinations of membranes (Miss) exceed pathogen rejection by conventional treatment. 9 SUMMARY ‘The implementation of membrane case studies is important to support decision making activities when designing, constructing and initiating a membrane process for potable water production, The value ofthe study is often dependent on the manner in which the study was performed. Site specific variables impact the results ofthe pilot investigation. Pilot investigations assist in confirming design criteria, waler quality goals, concentrate disposal parameters, membrane replacement estimates, operation costs, cleaning costs, maintenance estimates and construction costs. Pilot investigations are also used to familiarise and train operations personne! in the understanding of the technology relative to fll-scale operations prior to start-up. Pilot investigations also ean provide information related to concentrate water quality that can assist in determining disposal or reuse options. Fouling is pethaps the most important variable that must be delineated ina pilot4 Membrane Technology in Water and Wastewater Treatment investigation, as many have found that without site specific pilot investigations, the construction of those facilities will often result in membrane plants that do not produce design water quantities, typically because fouling issues were not identified prior to construction and operation ofthe facility. References 1. Allgeier, S.C., Summers, RS. 1995. “Evaluating NF For DBP Control With The RBSMI", Jounal AWWA. 87:3:87, (March 1995). 2. Allgeier, SC. 1999. “Analysis of Nanofilration Under the ICR.” Prac. AWWA Water (Quality and Technology Conference, Tampa, FL, Nov. 1999, 3. AWWARF, 1998. “Investigation of Inegrated Membrane Systems”, Quarterly Reports, 1996-1998, work in progress, Denver, CO 4, Duranceau, S.J. 1990. “Modeling of Mass Transfer and Synthetic Organic Compound Removal in a Membrane Softening Process.” Doctoral Dissertation, University of Central Florida, Orlando, FL. 5. Lovins W. A... Taylor J. S.Kozik R., Abbasedegan, M., LeChaevallier M. and Aty, K. 1999. “Mult-contaminant Removal by Integrated Membrane Systems”, Proceedings of AWWA Water Quality and Technology Conference, Nov 1999. (6. Mortis, K. M., 1990. “Predicting Fouling in Membrane Separation Processes,” Master ‘Thesis, University of Central Florida, Orlando, Florida 7. Duranceau, S.J. 1990. “Modeling of Mass Transfer and Synthetic Organic Compound Removal in a Membrane Softening Process.” Doctoral Dissertation, University of Central Florida, Orlando, FL. 8. Malford, L. A., Taylor J. S., Nickerson, D. M. and Chen Shaio-Shing, “NF performance at fall and pilot scale.” Journal AWWA, Vol. 91, No.6, June 1999. 9.Sung, Larry. 1993 "Modeling Mass Transfer in Nanofiltration,* Doctoral Dissertation, University of Central Florida, Orlando, FL. 10, Reiss, C. R, Taylor J. S., Owen C., and Robert C. 1999, “Diffusion-Controlled Organic Solute Mass Transport in Nanofitration Systems.” Proceedings of AWWA ‘Water Quality Technology Conference, Tampa, FL, Nov. 1999. 11, Taylor, J. $3; Duranceau, $J; Barrett, W.M. Goigel, J.F. 1989. "Assessment of Potable Water Membrane Application and Research Needs, AWWA Research Foundation Report, Denver, CO. 12, USEPA. 1992, “Reduction Of Disinfection By-Product Precursors By Nanofitration”, LUSEPA/600/SR-92/023, April 1992. 13. USEPA. 1998 “Removal of Multi-Contaminates by Integrated Membrane Systems”, ‘Quarterly Reports for Project in Progress, 1996-1998, WERL, DRWD, Cincinnati, OH‘SWRO - THE LARGEST PLANT IN BRITISH WATERS Mr N Marsh and Mr J Howard Ms F Finlayson and Dr S Rybar Jersey New Waterworks Weir Westgarth Ltd Malcaster House 149 Newlands Road Westmount Road Catheart StHelier Glasgow Jersey G44 4EX (Channel Islands 1 ABSTRACT ‘The paper describes the operational experience with the newly constructed SWRO plant in Jersey. The plant replaced a multi-stage flash distillation plant, commissioned in 1970 by Weir Westgarth, which was operational until 1996, At the time of supply the MSF plant ‘was considered to be the most energy efficient desalination process with a Gain Ratio of 12to 1 to minimise running costs Projected Operating Costs for the new SWRO plant are compared with the actual running costs as well as with actual running costs for the original MSF Plant. The environmental impact of both installations is also addressed, For this particular site the benefits of membrane technology are clearly demonstrated Keywords: Seawater; Reverse osmosis; Operating costs; Environmental impact 2. INTRODUCTION Jersey is the most southerly ofthe Channel Islands, measuring approximately 14.5 km by '8 km and is situated approximately 24 km from France. The geology of the island is varied and complex but incudes ancient shales with intrusions of granite, andesits, and fhyolite, overlaid by sand deposits. “The annual average rainfalls 847.4 mm (133 mm year average). ‘The Jersey New Waterworks Company Limited (NWW) was founded in 1882 and isthe oldest registered company operating in the Island. The resident population of the island is approximately 85,000 which rises to about 120,000 in the summer. The SNWW currently supplies approximately 85% of the island6 ‘Membrane Technology in Water and Wastewater Treaiment 3 HISTORY 3.1 Operation, Water Supply and Demand [Natural aquifers account for only 3% of the raw water supply. Unfortunately, some of the highest yielding catchments have the smallest reservoirs, A network of raw water ‘pumping stations has been developed to fill the reservoirs from remote stream sources and also provide a flexible means to supply Handois and Augres Water Trestment Works. ‘The maximum storage available is 2,677,340 m’. This represents only 4 months demand. The maximum daily demand up to June 1998, was 27,534 m? (6,06 mgd), in 1994, This maximum daily demand indicates 2 shortfall of 5,223m/day and based on this, the capacity ofthe SWRO plant was fixed at 6,000 m’/day. 3.2 How the Plant Performed ‘The Multi-Stage Flash Distillation Plant (MSF) was constructed between 1968 and 1969 by Weir Westgarth and was commissioned in 1970. The rated output of the MSF plant ‘was 6,715 cubic metres (1,500,000 imperial gallons) per day. The plant has produced a total of 6,550,000 m’ of product water. The plant was used extensively during the summer droughts of 1976, 1989 and 1990, where 950,000 m*, 877,000 m* and 1018,000 m* of ‘water were produced respectively. 3.3. Process Description ‘The existing distiller was a high performance unit (design GOR 12 to 1), to minimise energy consumption. It consisted of 47 stages, 44 stages were designed to recover heat and 3 stages to reject heat. Each stage operated at progressively increasing negative pressure, until Stage 47 was almost at full vacuum. At Stage 47 the water boiled at 37 °C On entering each stage a portion ofthe brine flashed and via demister pad and condensed of tube bundles. The resulting steam passed upwards to condense on tube bundles to produce the distillate. The feed water was dosed with sulphuric acid, to reduce calcium carbonate deposits, anti scalant to assist in keeping solids in suspension, and a de-foaming agent. ‘The distillate being of high purity was chemically agaressive and so lime was added to modify the hardness and tus prevent damage to the pumping main to Val de La Mare Reservoir. The distillate was then blended with natural water sources prior to being ‘pumped to the water treatment works. ‘The plant was decommissioned in 1998, 28 years after commissioning, as part of the ‘contract for the new Reverse Osmosis plant. 3.4 Economics of the MSF Plant The capital cost of the plant, including 5,500 m of distillate main was £1.2M in 1970. ‘The Weir Westgarth Contract for the MSF distiller, boilers, oil tanks, and all mechanical plant and instrumentation was £775,000, ‘The fuel cost to run the plant excluding all other costs at 1998 prices would be £4,083 er day.Case Sudies a 4 CHOICE OF NEW TECHNOLOGY ‘Advances in membrane technology over recent years have had a significant affect on the choice ofthe process for the proposed plant. The plants single purpose plant with alow load factor, with Reverse Osmosis (RO) providing the most atractve process option in terms of flexibility of operation, space requirements and minimal visual impact in a tourist area, “The RO process offers the following advantages, compared to thermal processes = lower energy requirement = shorter start-up and shut down times - lower maintenance requirement = operational flexibility = no atmospheric pollution 5 NEWPLANT DESCRIPTION ‘The new sea water desalination system consists of the following main stages (Figure 1): = sea water soreens and intake pumps = quarry pool pumps = pre-treatment with auxiliaries - high pressure system pumps and RO trains with energy recovery = product pumps and product water stabilisation ‘The plant is designed to have the daily output of 6000 m’/day of permeate with a quality of less than 400 mg/l TDS, based on a feed water quality of 37,734 mg/l TDS. Full output can be maintained over the sea water temperature range of 7.5°C to 19,5°C. The sea water salinity isin reality slightly higher and TDS based on individual ion analysis are around 38,000 mg/l. Feed water conductivity fluctuates around 54.0 mS/cm. 5.1 Sea Water Intake Screens and Intake Pumps The sea water intake station is located on the shoreline. Water from the sea passes through ‘the submersed tunnel to the intake pit. Due to the high amount of seaweed, present during rough seas, a macerator is installed inthe intake pit. Three vertical intake pumps - two in duty one in standby - pump seawater to the quarry poo! tank via an old cement coated cast iron pipeline. An, automatic, self cleaning filter is installed before the seawater enters the quarry pool tank with a filtration rate of 800 microns. The filter, which has operated satisfactorily since April, is fully automatic and self cleans on a pre-set differential pressure or timer. The quarry pool has an estimated capacity of 32,000 m? and in the past debris in the seawater settled in the bottom of the quarry pool. The role of strainer is to eliminate further sedimentation and decay of particles larger than 800 microns,welBeIq Mold “O'U'M'S Aesiep “1 adeCase Studies » 52 Quarry pool pumps Under normal continuous flow conditions the quarry pool level will fluctuate by approximately one meter. The quarry pool pumps are designed to supply the plant with sufficient flow and pressure up to 7.6 meters below the normal operating level. This allows drainage of the quarry pool after long term plant shut down. Five quarry pool pumps are provided, four duty and one stand by. 53° Pre-treatment The role of the pre-treatment is to filter seawater to @ quality acceptable to feed into RO membranes and includes: DMF’; backwash pumps and backwash tank; ar scour blowers; cartridge filters; pre-treatment chemical dosing, The dosing of flocculant upstream of the media filters agglomerates the suspended solids and colloidal matter Under automatic control the flocculant is regulated to guarantee quick and optimum flocculation. The flocculant used is an organic type, based on the coagulation tests carried out on site in November last year and Weir Westgarth’s operational experience from existing installation’s (Gibraltar). In addition a set of compatibility tests was carried out to find if there is a potential danger of increased membrane fouling due to the reaction between organic coagulant and membrane, Test and operational results indicated no increased fouling potential due to the use of organic flocculant. System sterilisation is based on intermittent shock chlorination using sodium hypochlorite. It was expected that shock disinfecting - chlorinating would be required once every two weeks for about an hour. During the first three months of operation this hhas had to be carried out weekly and for 3-4 hours Flocoulated colloidal matter and suspended solids are removed in dual media filters. FFour such filter chambers are provided in two pressure vessels. The filters normally operate at filtration velocity of 8.8 mvh. With one chamber in backwash, the velocity in the remaining chambers reaches 11.8 m/h, ‘The projected operation time between two backwash cycles is 24 hours, corresponding to a dual media filter layer capacity for suspended solids of approximately 2.9 kg/m? or 2.8 kg/m’ of filtration media at 13 mgll of suspended solids infeed water. ‘Operational data gathered to date have indicated an average quarry pool SDI of 28 and an average DMF outlet quality of 4.5 - 4.8 SDI ‘The filtered water enters @ common header connected to four (33% duty) 10-mieron cartridge filters These provide the final pre-treatment step. The cartridge filter outlet is connected tothe high pressure pumps suction header. Scale inhibitor is added upstream of the cartridge filters by proportionally controlled dosing pump sets (3 x 50% pumps). 54 RO Membranes 4 x 1,500 m'/day membrane streams are provided. Each stream consists ofa high pressure pump; turbo-booster and set of membranes to produce 1,500 m’/day of permeate. The high pressure pump selected is of horizontal centrifugal design.30 ‘Membrane Technology in Water and Wastewater Trearment ‘The turbo-booster utilises the high pressure brine rejected from the reverse osmosis membranes, Transferring pressure energy in the brine reject stream to the feed stream, reducing the electrical power required during normal operation. Plant conversion of 45% was selected to ensure the power requirement was within the 2MVA power supply. With 45% conversion the projected specific energy requirement of the plant is 5.75 kWH/m’ for process load alone (pre-treatment and RO part). It is important to point that all pumps used in this installation are designed and manufactured by WEIR PUMPS Lid. ‘Normalising plant performance against the first day of operation using ASTM D4516- 85 (Reapproved 1989) and a temperature correction factor from the membrane ‘manufacturer assess tran performance. Membrane performance has been as good as predicted by the manufacturer with normalised salt passage of 0.27 - 0.3% at design flow rates, SS Control System ‘The system provided offers the flexibility and performance normally attributed to distributed control system (DCS), and is designed to provide minimum operator intervention and maintain the necessary interlocks for safe operation To facilitate the control of the new SWRO plant a PLC based system has been selected with a PC based topology incorporating a SCADA software package for ease of operatic, "The system architecture has three major components 1, Programmable Logic Controller (PLC) ‘A host PLC with direct connections to field signals and communications interfaces to PC based visualisation system (VS) or main machine interfaces (MMI) and remote terminal units (RTU). 2, Remote Terminal Units (RTU) AAs the geographic layout of the plant is diversified field cabling is minimised by introducing the use of remote terminal units (RTU) which are located in fietd junction ‘boxes to remotely collect the field instrumentation signals and relay them back to the PLC via a bus system, 3. Visualisation System (VS) or Man Machine Interface (MMI) Visualisation system (VS) or man machine interface (MMI) run on PC's with communication to the PLC. ‘The PC runs a software package and represents the process ‘configuration as MIMIC diagrams with individual displays of plant items or equipment ‘and faceplates for control points. A common point database resides on the system and is shared between the VS/MMI and the PLC, however the VS/MMI is contained and configured separately on the PC. The VS/MMI also offers trending, facilities, alarm handling, operator prompts and messages and data storage. ‘A direct link between Weir Westgarth Ltd and the plant control room enables process supervision or operational data transfer at any time 56 For comparison we have used a price of Sp per kWh of electricity. This gave a designed daily energy running cost of approximately £2,025 per day (compared to £4,083 for the MSF plant) (Table!)Case Studies 3 ‘The new RO plant avoids the buming of 45,400 litres of heavy grade oil per day and ‘consequent emission of 2.2 tonnes of SO;. This provides a significant environmental ‘improvement. Chemical costs were in the region of 8p/day for the MSF plant. The RO plant has similar chemical costs but utilises less hazardous compounds, ‘The new plant generates no gaseous or odorous emissions. The main effluent streams are brine and cleaning fluids. The characteristics of liquid effluents as well as the rejected flows can be summarised as follows: = brine isthe major liquid effluent produced and its concentration in chemical ‘compounds is approximately double the concentration of the feed seawater. The chemistry ofthe brine varies according to the chemistry of the raw seawater. The ‘wave motion rendering it harmless quickly disperses it = organic coagulant is used in the process as an altemative to feric chloride to avoid pollution of the sea with staining iron salts. The organic coagulant chosen is biologically degradable. = sodium hypochlorite used for shock disinfecting is neutralised before discharge tothe sea = preservation chemicals and cleaning chemicals are neutralised in the neutralisation pit before they are discharged to the sea 6 SUMMARY ‘The MSF Distillation Plant constructed between 1968 and 1969 by Weir Westgarth Lid was decommissioned and replaced by a new desalination plant with capacity of 6000 m’/day, also designed and built by Weir Westgarth Ltd. The contract was signed 5* of January 1998, with the plant handed over to the client on 29* June 1999. The design, construction, commissioning and reliability tests were executed within 18 months. The plant employs the latest water treatment technologies and minimises operating costs. ‘Table 1 Running costs ROCESS | OUTPUT] ENERGY COST PROCESRACLEANING ] CART FICTER] TOTAT geDay) (eDAY) cewicaL. —_|RERLACEMENT_| cost MSF [6517 | 4083 (Fuel Oi) £0.08" = [0.706%] 0.626in" SWRO | 6000 [2025 (Eleatigy | S008? BOOT [EOI a? design 0.338%m" ‘SWRO | 6000 | 2022(Electrci) | E008 FO_OGTRT [EDS a sctual 0337im References 1 EGDarton AG Tumer EDA Canagua 1996 2 WR Quems (Private Communications)DRINKING WATER SOURCES IN KUWAIT M, Safar and ¥. Al-Wazzan Water Desalination Department, Water Resources Division, Kuvait Institute for Scientific Research PO Box 24885, Safat 13109, Kuwait ABSTRACT Kuwait's water resources are limited to desalinated water, brackish water and ‘wastewater. In Kuwait, seawater desalination is the main source of freshwater required by all demand sectors. Other marginal supplies come from groundwater and bottled water. Freshwater is commercially produced by seawater desalination by two main processes, ie., multistage distillation flash (MSF) and reverse osmosis (RO). ‘This paper contains detailed information regarding drinking water in Kuwait in terms of system configuration, production rate and water quality standards, It also focuses on the implementation of membrane separation systems (ie., reverse osmosis) for the production of freshwater from both seawater and brackish water, Keywords: Multistage flash, membrane, bottled water, groundwater, 1 INTRODUCTION ‘The history of civilisation is closely linked to the availability of water, which has always been the primary source of economic growth, being directly related to food production and thus of the communities’ concern, Invariably, the history of nations reveals that the survival of any civilisation is directly related to water management. The shortage of potable water, especially in hot climatic zones, causes social discord and the disruption of economic development’ ‘The World Bank has pointed out in it water management strategy that the water situation in the Middle East is precarious. Per capita renewable water resources have fallen ftom approximately 3500 m? to less than 700 m° * Thus; water must now be viewed as a limited resource not as a sectored input. The shortage in natural water supplyCate Smdies a for domestic purposes is more acute for the Arabian Gulf area, where the demand for water increases annually ata rate of 3 or more percent? ‘Natural resources of freshwater in Kuwait are very limited because Kuwait is situated in an arid coastal region characterised by high temperature, decreased humidity, little rainfall and high evaporation rate. Therefore, Kuwait has always had to search for other sources to secure freshwater to meet with its growing demands. Kuvait in the past relied mainly on rainwater found near the surface in shallow wells, but due to the growth of population, that scant source became no longer sufficient to cater to the growing demand. So, in 1939, Kuwait tumed to the Shaat Al-Arab for freshwater supply brought by boats, and a primitive storage and distribution network was established. The water ‘transported ata rate of 8500 IGPD increased to 80000 IGPD by the end of 1946%. This situation prevailed until the influx of oil wealth and the first oil shipment in 1946. Kuwait then had the funds necessary to invest in modem water production facilities that could cater to freshwater demand, ‘This paper describes in detail all sources of freshwater in Kuwait in terms of system configuration, production rate and water quality standards. It also focuses on the implementation of membrane separation systems (ie., reverse osmosis) for the production of freshwater from both seawater and brackish water. 2. DRINKING WATER SOURCES IN KUWAIT Freshwater in Kuwait can be classified into four main categories: underground water, bottled water, desalinated water by reverse osmosis (RO), and 100 ML/day). Unit sizes tend to be relatively small (typically 100 ML/day. ‘This paper briefly reviews the history of CMF development. It deseribes the fist small scale commercial project using CMF-S at Marulan in Australia and the successful bid for the AQUA 2000 project, the first large scale CMF-S plant and the largest nicrofiltation plant inthe world ~ 126MLIday, due to be commissioned lat 2000, 3 EVOLUTION OF FOURTH GENERATION CMF-S TECHNOLOGY USF/Memcor introduced its CMF microfitration technology in the early cightes. ‘The first generation of systems were relatively expensive and were therefore limited to high value low volume process streams such as wine, fruit juice, and pharmaceutical spplicaions. They were typically manufactured info small skid mounted units using standard plumbing fitings Further cost reductions and simplifications were made in the second generation Mi and M2 series by using injection moulded manifolding and doubling the area of each module (by increasing the length). Microfiltation was now ‘becoming cost effective for small-scale water and wastewater treatment and in 1991 the first municipal drinking water plant was installed in Tooborac, Victoria, Australia (170 kLday) The third generation of microfitration systems, the MIO and MIOC series, were introduced in 1992 specifically t0 allow high volume processing of water and wastewater using repairable modules. A microfilration block of this type contains 2 to 50 modules depending on the required filtration capacity. One 90-module block filters 25 to 4.5 MLiday depending on the feed water quality. Large plants duplicate the 90- module block ‘The development of the M10/M10C technology coincided with the introduction of the ‘surface water treatment rule in the US and recognition that membranes offered an elegant solution tothe problem of removing chlorine tolerant organisms such as Crytosporidium and Giardia, | _ 200 This resulted in a rapid growth in | 3 CMF installations. Growing | acceptance of the technology and : a Cr 1000 ‘e i further developments have seen system costs tumble (Figure 1). There are now over 100 Memcor CMF potable water plants with a f capacity of between T and 125 |" Mbidey "installed or under oe construction, with a total MF are capacity in excess of 500 MLdey. Figure 1 Cost reduction overtime USF Memcor's most recent development is the CMF-S (submerged) series. This fouth generation systm isthe result of several years of R&D and represents a natural progression of the CMF technology ino very large-scale filration applications. The impliiaion ofthe process and stem allows it to be applied cost effectively a apaiies greater than 100 ML/day sing unit multiples of proximately 20 ML/day.Case Sues 0 For the first time this allows membrane filtration to compete on @ cost par with conventional processes on even the largest scale filtration applications. 4. THE CMF-S PROCESS In the CME-S process the membranes are submerged in an open tank and filtrate is withdrawn from the modules under suction. The membranes operate in direct flow using the same nominal 0.2um polypropylene membrane as the conventional CMF process has ‘employed for many years. A periodic backwash is required to dislodge the solids and chemical cleaning is employed at intervals to restore membrane performance. A CMF-S Backwash Earlier CMF generations used a unique ges backwash to keep the membrane surface clean. The CME-S backwash is based on the same principles with the exception that the air is applied extemally tothe fibre surface rather than through the membrane wall. The key to its success is a design that allows the air to be delivered evenly into the depths of highly packed membrane module. A short filtrate backwash supplements the scrubbing action of the air to help further dislodge solids. Solids dislodged from the membrane are drained from the tank at the ‘end of the backwash cycle. This process ensures that there is no accumulation of solids within the membrane tank that can affect the membrane performance or the water uality. The result is a backwash process with efficiency similar to conventional CMF. 42 CMF-S Clean-In-Place (CIP) (Chemical cleaning is achieved pumping cleaning solution from the CIP tank to the cell containing the membrane modules. The membranes are contacted with the chemicals ‘and periodically aerated. At the completion of the cycle the cleaning solution is recovered tothe CIP tank. The membranes are then rinsed and retumed to service. 43. CME-S Integrity Monitoring and Control Both CMF and CMF-S use three key steps to achieve control of system integrity ~ pressure decay testing, identifying leaks and isolating leaking sub-modules. Pressure decay testing allows the integrity of the barrier to be quantified without needing to rely ‘on measurements of treated water quality. The test is also more sensitive to changes in integrity than any current water quality testing methods can achieve. Air is applied to the filtrate side of the system at a pressure below the bubble point (ypically 100 kPa). The air displaces filtrate from the lumen through the membrane wall. Once all the liquid has been expelled from the lumen the system is isolated and the pressure decay monitored. The rate of pressure decay is a direct measure of the ‘membrane integrity and can be related to log removal of particles larger than the pore size (eg. Cryptosporidium) (Hong et a.,1999). If pressure decay testing indicates a drop in integrity the source of the leak is easily Jocated by visual inspection over the top of the tank. Groups of modules can be isolated for later repair if required.© [Membrane Technology in Water and Wastewater Treatment 5. APPLICATION OF CMF-S AT MARULAN, AUSTRALIA 5.1 Background to the Project ‘The Marulan project is the first commercial installation of the CMF -S technology. The town of Marulan is located 20 km east of Goulburn and has a population of 650 people. ‘The town’s water supply system was installed initially in 1967 at a capacity of 1.1 ‘ML/day. It comprised reticulation and chlorine disinfection but no water treatment. In order to meet the multiple objectives of increasing capacity, environmental flows in Summer and security of supply, the local authority: Mulwarree Shire Council decided to: ‘© Construct a 35 ML reservoir to ensure water availability during dry periods. © Negotiate with a private landholder to supply 300 ML/year to meet requirements from the Healthy Rivers Commission to meet environmental flows. ‘* Install a remotely operated microfiltration unit (1-5 ML/day) which would meet ccurrent and future water quality guidelines, remove waterbome parasites, cope with ‘occasional algal blooms and which could be expanded to 2 ML/day. 52 Process Description and Performance ‘The quantum change with CME-S is in the arrangement of the system so as to eliminate the need for pressure housings. The membrane modules are suspended in an open tank from a series of removable manifolds. On the Marulan plant there are a total of 56 ‘modules installed on 7 manifold racks, each containing 8 membrane modules, as shown in Figure 2. Figure 2. Membrane tank at Marulan showing submerged membrane and membrane rack assembly 53. Performance Result ‘The plant was commissioned in August 1998 and atthe time of writing has been operat- ing for just over 12 months. Demand on the plant has been relatively low compared to the design flow of 1.5 ML/day and the unit is usually only required to operate for 4 ~ 8Case Studies a hhours a day. The low demand and reduced flux means that the fouling rate is very low. ‘The TMP stayed relatively stable for the first 6 months of operation, then an increase in demand saw the fouling rate increase. The first chemical clean was after 8 months of operation. Since then, the unit has been cleaned every 2 ~ 3 months, without the plant reaching high TMP, purely to ensure no regrowth in pipework. Mandan 86619 Prtormance Figure 3. Filtration Performance of CMF-S plant at Marulan WIP Operator involvement has been minimal. The plant is remotely monitored using a SCADA package and data is collected by an in-built proprietary datalogging system supplied with the CMF-S unit. The operator makes 2 daly visit to check the plant ‘operation and chlorination system. Pressure decay tests carried out on the unit indicate that the log removal of Cryptosporidium and Giardia willbe atleast 5 log. Regular pressure decay tests and an alarm if values exceed a preset limit, are used to ensure removal is always greater than at least 4 log, 6 OTHER APPLICATIONS CMF-S has now been tilled on over 14 different sites with widely varying water and ‘wastewater qualities. In all cases CMF-S has operated suecessflly, producing treated ‘water of quality equal to conventional CMF in similar applications. "As well as remov- ing turbidity and organisms, coagulant has been directly dosed onto the CMF-S to remove colour and organic contaminants 7. SCALE-UP OF CMF-S TO LARGE SCALE FILTRATION ~ AQUA 2000 71. Project Background ‘The first lange scale CMF-S plant willbe part of the AQUA. 2000 project for Coliban Water, a 25 year BOOT scheme to provide drinking water to the Bendigo area in ‘Victoria, Australia, USF Australia successfully tendered for this project with the CMB-Se ‘Membrane Technology in Water and Wastewater Treament process in 1998. Currently design and construction is underway and commissioning is due late 2000. AQUA 2000 project includes four water treatment plants spread over an area of 1600 square km, with a total capacity of 165MLiday. The largest plant, at Sandhurst Reservoir, will be the largest microfiltration plant in the world with an initial capacity of T26ML day. 72 Process Description ‘The requirements of the BOOT contract have ensured the AQUA 2000 water treatment plants can cope with both current and future water quality requirements, most significantly inthe area of removal of particles, algae and disinfection by-products. ‘The pre-treatment for the microfiltration uses Trident adsorption clarifiers, with coagulant dosing to remove metals and colour. Powdered activated carbon can be dosed in the event of algal blooms. The building block of the CMF-S process is the 576 module cell, described in section 7.3. The largest plant in the AQUA 2000 project is at Sandhurst Reservoir and requires 6x $76 module cells to reach the capacity of 126 ML/day. USF Memeor Research tested the CMF-S and Trident adsorption clarifier processes on site at Sandhurst reservoir for eight months. ‘The trial validated the combination of the combined Trident-CMF-S process, removal of particles, colour and organics — particularly the removal of disinfection by-products. 73. Large Scale CMF-S The large-scale version of CMF-S is illustrated in Figures 4 to 6. The hollow fibre filtration membrane is fabricated into repairable sub modules having a nominal filtration area of 13 m?. Four sub modules are arranged into ‘a sub manifold assembly and up to eight sub manifolds connected via a stainless stcel manifold 10 a removable module rack The filtrate connection is via the top of each sub-manifold and backwash airflow via the bottom. Module racks are assembled into an open filtration cell. One back washable filtration cell can contain up to 18 module racks, i.c. 576 sub modules. Each cell has a nominal capacity of 20 ML/day when filtering typical surface water. Figure 7 shows the 6 x 576 ‘module cells for the 126ML/day Sandhurst Reservoir plant. ‘The module racks are suspended in the cell and are lifted directly from the cell for servicing. For plant illustrated in Figure 6, two rows of three cells are configured back- to-back with a central feed channel in-between. Backwash channels are located at the Fig 120 My et ‘neoCase Suies 6 opposite sides. Each cell incorporates local filtrate pump and valves to permit individual flow contol, backwash and CIP. ‘The CMF-S process has a small footprint. Each cell is approximately 6m x 3m. The 6 cell layout for the 126ML/day plant at Sandhurst Reservoir shown in Figure 7 occupies & space of less than 20m x 20m, excluding system ancillaries. Filtration cell size and construction is adaptable based on a modular approach to meet specific application requirements. Individual cell size typically ranges from 160 to 576 total membrane sub-modules. The back-to-back configuration shown can be expanded to.a 5 x2 train that would have a nominal capacity of 200 ML/day. Multiples of this train are then used as building blocks for larger capacity systems. Total CMF-S plant capacity is not limited. Trials ae already underway to validate the 32-module rack design shown in Figure 5. ‘These confirm exact scale-up from a single module to 32 modules (Figure 7). Seale-up from I rack to 18 racks (a full cell) is expected to be equally efficient. This is because: The open tank design with submerged modules allows membranes to be in intimate contact with the feedwater at all times, regardless of cell size. Raw water distribution throughout the unit is assured by level. Backwash air is distributed through manifolds directly into groups of four sub-modules so close control of air flow to each are completely removed from the system by draining the cell tthe end ofthe backwash cycle This eliminates | 65: ose iene Soll er Oe any possi of | Mart a ae Bap are ar raldistibution within the cll : that may lead to loss of sealeup efficiency E i a z 3 & “TMP (KPa) gosBSSS8388 Figure 7 ~ Scale-up from a single to module 10 a rack of 32 modules 8 costs As CMF-S is a submerged process, pressure vessels are not required to house the modules. This greatly simplifies the system design and reduces the capital cost. In addition the design of CMF-S lends itself to scale-up more readily as large membrane area can be installed in each cell. ‘The disadvantage of a submerged process is that itis TMP limited. It is not possible to run the system to TMP's greater than the pump design will allow, without adding liquid height above the membranes. The practical outcome is thatthe operating TMP range of ‘CMF-S is less than in a pressurised process. ‘This means that a lower flux (or greater membrane area) is needed if the CIP interval is tobe the same for both processes. A design flux for CMF-S between 75% and 95% of (CMF is typical, depending on the feed water. However, any increase in membrane area is more than offset by a lower power cost (due to lower TMP) for the CMF-S process.a Membrane Technology in Water and Wastewater Treament ‘The power consumption for the CMF-S process is up to 30% of that for conventional CMF. ‘Chemical consumption is approximately the same, and maintenance costs are expected to be lower due to fewer valves and a less demanding operating cycle. The net result is that the operating cost of the CMF-S process is usually less than for conventional CMF (typically about 5% lower). 9 CONCLUSIONS The CMF-S (submerged) series isthe fourth generation CMF system and represents ‘natural progression of the CMF technology into very large-scale filtration applications. The simplification ofthe process and system allows it to be applied cost effectively at capacities greater than 100 ML/day using unit multiples of approximately 20 ML/day. ‘© The CMF-S process is to be used in the largest microfiltration plant in the world (126ML/day) as pat of the AQUA 2000 project due to be commissioned late 2000. ‘© Many ofthe features ofthe conventional CMF process have been retained, including ‘the use of air for scrubbing the membranes, integrity testing using pressure decay, and the same 0.2um polypropylene membrane, thus building on the experience of ‘over 700 CMF installations worldwide. ‘© The Marulan plant represents the first commercial installation of the new fourth generation CMF-S process. Although rated at 1.5 ML/day the plant demand is @ flow. The plant has been running for an average of 6-8 hours per day at reduced flux since August 1998. ‘© CME-S has now been trialled on over 14 different sites with widely varying water and wastewater qualities. Trials have confirmed that both the backwash efficiency and treated water quality for CMF-S are equal to conventional CMF in similar applications. ‘© Operating costs are overall lower than conventional CMF due principally to significantly reduced power requirement. This is partially offset by a lower flux (higher membrane area per unit flow) due tothe limited TMP range. * The simplicity of the process, particularly for large scale systems, significantly reduces the capital cost of the system. Designs have been developed and component ‘manufacture is underway for systems that use cells, with a nominal capacity of 20 ‘MLiday, to build systems of greater than 100 ML/day cost effectively. References Hong, SK; Taylor, 1S.; Miller, F; Rose, J; Gibson, C Owen, C. Johnson, W-T. “Removal of Microorganisms by MF Process: Comrlation between Integrity Test Results and Microbial Removal Efficiency”. Proceeding of AWWA Membrane Technology Conference, Longbeach, CA, April 1999.Water Quality and Treatment‘THE UK SYSTEM OF APPROVAL OF PRODUCTS USED IN CONTACT WITH DRINKING WATER Dr Toks Ogunbi ‘Technical Secretary ‘Committee on Chemicals and Materials Drinking Water Inspectorate Ashdown House 123 Vietoria Street LONDON SWIE 6DE PREAMBLE ‘The Secretary of State has powers to make regulations controlling substances, products ‘and processes used in the treatment and provision of public water supplies in England and Wales. 1 INTRODUCTION The development of the regulatory framework for the authorisation/approval of products used in contact with public water supplies may be examined through three time periods ~ pre 1989, current/post 1989 and the future. 2: THE PERIOD BEFORE 1989 Prior to 1989, products used by water undertakers were introduced into the public water supply under a ‘voluntary’ system. The Committee on Chemicals and Materials did not have any statutory powers but was aocepted as watchdog whose opinions were respected though not legally binding; it had bark but lacked bite The Commitice was established in 1966 by the then Minister of Housi Government. Prior to the Water Act 1989 the Committee operated on advisory basis, initially to approve chemicals for use in public water supplies. The scope of the Committe’s activities was later extended to swimming pools, desalination plants and materials of construction for use in public water supplies. The Committe's terms of reference were further extended to consider the environmental impact of chemicals and ‘materials submitted for approval. The Committee approved product if use in contact with water was considered to be unobjectionable on health grounds and where appropriate, the conditions of approval took account of the environmental impact of the product. The ‘Committee periodically published statements listing approved products, the last being the 15th Statement, issued in 1989, Water undertakers were not legally obliged to use approved products but did so by and large.6 ‘Membrane Technology In Water and Wastewater Treament 3 THE PERIOD SINCE 1989 ‘The advent of formal regulation of products used in contact with public water supply may be seen as the Water Act 1989. Section $3 gave the Secretary of State powers to make regulations controlling substances, products and processes used in the treatment and provision of public water supplies. Regulation 25 of the Water Supply (Water Quality) Regulations 1989 came into force on 1 September 1989. Regulation 25 concems the introduction of substances and products for use inthe treatment and provision of public water supplies. Under Regulation 25(1, a water undertaker shall not, otherwise than for the purposes of testing and research, apply to or introduce into water which is to be supplied for drinking, washing, cooking or food production purposes, any substance or proviuct unless it (a) hhas been approved by the Secretary of State for the Environment and the Secretary of State for Wales — [regulation 25(1(a)}; or (b) is considered by the water undertaker to be unlikely to affect adversely the quality of the water [regulation 25(1}(0); or (©) has been used by a water undertaker during the 12 months prior to 6 July 1989 [regulation 25(1 (0); oF (@) is listed in the 15th Statement or any supplement issued by the Committee on Chemicals and Materials of Construction for Use in Public Water Supply and Swimming Pools [regulation 25(1)4)] ‘Under Regulation 25(4) the Secretary of State has powers to prohibit the use of any substance or product which water undertakers were previously authorised to use under regulations 25(1)(b), (©) or (d). The Secretary of State may also revoke or modify conditions of any approval given under regulation 25(1)(@). Unless itis in the interests of public health for prohibition, revocation or modification to take effect immediately, six ‘months’ notice must be given in writing to water undertakers and approval holders ion 25(8) requires the Secretary of State atleast once in each year to issue a list of all substances and products for which approval has been granted, refused, modified, revoked or prohibited. ‘Regulation 26 (1) provides for the Secretary of State to require a water undertaker to make an application for approval of any process orto prohibit the use of any process. ‘Regulation 26(4) makes provision for the publication of a similar list for processes. Under Regulation 27 the Secretary of State may require the applicant to pay a charge Which reflects administrative expenses of determining an application for approval ‘Regulation 28 relates to offences resulting from any breach of regulations 25 and 26, or from the provision of false information in support of an application for approval 3.1 Situations where Regulation 25 is Not Applicable Regulation 25 applies to substances and products which come into contact with water in the treatment and distribution of publi including use in aw water storage reservoirs, borehole installations and pipelines. Use in the following circumstances falls outside the scope of regulation 25 (2) afterthe time of supply i. within consumers’ installations, (©) in eatment and distribution of private water supplies; (©) _intreatment of swimming poo! water, and (© application of pesticides and weed control agents to raw water storage areas,Water Quay and Treament e Use of substances and products in circumstance (a) fells within the scope of the Water Regulations Advisory Scheme, Although there is no legal requirement to use approved substances and products in circumstance (b) the Inspectorate has advised that this be done wherever practicable ‘The Committee continues to operate a non-statutory approval system for substances and products used in swimming pools. The voluntary approval scheme for chemicals used in the treatment of drinking water by membrane and distillation processes has been discontinued. All products approved under ths eategory are to be reviewed for approval under Regulation 25( 1, Circumstance (4) falls within the scope of the Control of Pesticides Regulations 1986, 3.2 The Committee on Chemicals and Materials. ‘The Committee on Chemicals and Materials was recently re-consttuted in the light of the Nolan Committee's recommendations, as a non-departmental public body. It now consists of a Chairman and five members with a range of expertise including engineering, materials science, toxicology and water treatment, an independent member represents ‘water consumers” interests Products are recommended for approval if they are considered to be unobjectionable on health grounds, in the light ofthe best available evidence atthe time. Approval is given solely forthe purposes of Regulation 25 ofthe Water Supply (Water Quality) Regulations and should not be taken o imply any recommendation as to the technical merits of product in contact with water for public supply. Products such as reverse osmosis, nano-, ultra- and microfitrtion membrane elements and systems are usually tested for their leaching characteristics ‘The Committee carries out a risk assessment and where the nature and level of the leachate determinands are such that they are considered to be of low or negligible risk to ‘the consumer such a product may be recommended for approval by the Secretaries of State for the Environment and for Wales, ‘The Drinking Water Inspectorate does not grant the approvals. Accordingly the use of the DWI logo for advertising or any other promotional material isnot allowed 33 Cost of Approvals, ‘The Secretary of State currently does not charge for the approval of products. Applicants should however expect to pay the designated test laboratories for any tests the laboratories. ‘may carry out in respect of their products. 3.4 Scotland and Northern Ireland ‘The Water Supply (Water Quality)(Scotland) Regulations 1990 and the Water Quality (Norther Ireland) Regulations 1994 make provisions for the control of substances and products. These regulations do not contain provisions for separate approval systems, although the Secretary of State for Scotland issues duplicates of approval letters and other legal instruments issued in England and Wales. Approvals issued by the Secretary of State are recognised in Northem Ireland, but separate approval letters are not issued there.” Membrane Technology in Water and Wastewater Treaoment 4 FUTURE DEVELOPMENTS ‘The development of legislation in the UK has inevitably been influenced by developments in the EU. This is in part due to the requirements of the Single European Act and the need ‘to remove technical barriers to trade; but also as a public health measure. 4.1 The Role of CEN Standards European Standards for water treatment chemicals are being developed by Working Group 9 of CEN Technical Committee 164, These standards are published by the British Standards Institution in the BS:EN series. Subject to complying with national conditions ‘of approval, chemicals for drinking water treatment which conform with a BS:EN may be used without the approval of the Secretary of State. It is envisaged that in future, water uiltes will be able to make greater use of European Standards when procuring chemicals and construction products. There will be a need to make to make an assessment of whether the European Standard and manufacturers intended use of the product are appropriate. Water utilities will beable to use any product which they are satisfied meets the European Standard without recourse to the Committee con Chemicals and Materials. This should ultimately eliminate the need forthe remaining traditional use exemptions. 4.2 The End of Traditional Use Exemptions DWI Information Letter 9/97 confirmed that the intended prohibition of unapproved uses of substances and products in water supply pipes, service reservoirs and water towers will be effective from 1 April 2000. So far relatively few manufacturers and suppliers have made applications for the approval of these products. The prohibition removes the traditional use exemption which exists for cementitious pipes and linings, and bituminous linings, ete. From 1 April, 2000 all such products will have to be approved under Regulation 25(1)a. Those UK manufacturers and suppliers who seek and obtain approval for all such products will be advantageously placed for the advent of the European Approval System (EAS). 43 Introduction of the EAS For the past ten years there have been efforts simed at harmonising the approval systems across Europe, for construction products (pipes, storage tanks, reservoirs, eto) used in ‘contact with public water supplies. DG III of the EC has recently announced an initiative ‘which will ensure closer alignment between the approval systems of all member states resulting in the development of the EAS. The EAS will operate through national regulatory bodies resulting in a “European common list’ of mutually recognised products which will be eligible for use by all water utilities throughout Europe without any further testing or assessment. This ‘common list’ is likely to be supervised by a system of regular uit inspections for material sources and manufacturing procedures. ‘After the inception of the ‘European Common List’, UK applicants will stil able to get their product onto that list by application to one of the national regulatory bodies which ‘operates the EAS, In the UK itis likely that the Committee on Chemicals and Materials will fulfil the national regulatory role with its activities extended to cover products used within consumers’ premises,IMMERSED MEMBRANES FOR DRINKING WATER PRODUCTION Pierre Coté Christian Gingerich and Ulrich Mende ZENON Environmental Inc., Zenon GmbH 845 Harrington Court, Nikaulos-Otto- Str4, Burlington D-40721, LIN 3P3, Hilden, Oniario, Germany Canada 1 INTRODUCTION In the passed ten years, membrane filtration has been proven as a technology offering high quality, safety and reliability for the production of drinking water. Over the same period, prices have decreased by a factor of more than five, making membrane filtration cost ‘competitive with conventional clarification and disinfection technologies. ‘There have been intense and rapid technological changes over the same period, but membranes and systems have not been standardised yet. Some key choices have been made: ‘membranes are in the range of loose ultrafiltration to microfiltration, with pore sizes ranging between 0.01 to 0.2 tim, materials are polymeric, and the configuration is the hollow fibre ranging in diameter from 0.5 to 2.0 mm. However, for about half of the products available, filtration is from the inside-out, while for the other half, it is from the outside-in. In addition, most membranes are housed in pressure vessels and the feed is pressurised; the alternative configuration is a shell-less module with the driving pressure applied by suction ‘on the permeate sie. Zenon Environmental Inc. has pioneered the development of shell-less immersed membranes. These authors are convinced that this configuration is best technological platform for a wide variety of applications in water treatment. This paper examines the key product and operating features of ZeeWeed®, Zenon’s immersed membrane, presents an analysis of the benefits to users and describes various full-scale applications. 2. IMMERSED MEMBRANE AND MODULE. ‘The ZeeWeed® hollow fibre has a composite structure illustrated in Figure 1. A. strong ‘support provides strength and flexibility, while the membrane permeation properties can be ‘optimised separately. The inside and outside diameters of the hollow fibre are 0.9 and 1.9 ‘mm, respectively.n ‘Membrane Technology in Water and Wastewater Treatment ‘Seppo oe Figure1 The ZeeWeed? hollow fibre sructre Various membranes are available for different applications. The OCP chemistry used drinking water production, is an ultrafiltration membrane with a pore size of 0.04 um. itis hydrophilic, chlorine-resistant and can be cleaned in a pH range of 2to 10.5. Teper : ene camber Te a i Sak (eee shoe) teste Posie = Figure2 The ZW-650 module ‘A module consists of hollow fibres mounted on a vertical frame with permeate extraction from bottom and top headers (Figure 2). Each header contains a layer of potting resin that the fibres cross so that their lumens are connected to chambers where the permeate is collected. The hollow fibres are slightly longer than the distance between the top and bottom headers. Two rigid rectangular pipes connect the two headers and keep them at a fixedWater Quality and Treaoment n distance. One of them is used to carry the permeate from the bottom header to the top header. The other one is used to cary air to the aerators integrated into the bottom header. ‘The module dimensions are length of 2 000 mm, width of 700 mm and thickness of 200 mm, Two to twelve modules are assembled into a cassette, which constitute the building block for systems. A standard cassette of 8 modules contains 480 m? of membrane surface area and has a capacity of 30 to 40 m’/h under a trans-membrane pressure up to 50 kPa. 3 IMMERSED MEMBRANE SYSTEM ‘An immersed membrane filtration plant is composed of parallel tanks into which cassettes ‘are immersed. One tank normally corresponds to one production unit (one train) and is quipped with dedicated permeate pump and blower. This approach allows the construction of large trains of up to 20 000 m'/d. When compared to pressurised membrane systems, ‘where the train size is limited to approximately 4 000 m’/d, this greatly limits the number of ancillary equipment, pumps, blowers and valves. ‘This design is very flexible and can easily be upgraded for flow by simply adding cassettes {nto the tanks. Where it does not freeze, the membrane tanks do not need to be covered by a building. ‘A complete system also includes a number of shared ancillary equipment: back-pulse pumping unit, permeate storage tank, clean-in-place unit and membrane integrity testing Feeding is normally by gravity and requires minimum pre-screening of 10 mm, Pormeato > Roject Figure3 Immersed membrane tank design with plug flow Small systems are operated as completely stirred tank reactors (CSTR). However, the ‘common design for larger systems is to implement plug flow in a long tank (PFR), feeding at ‘one end and continuously bleeding the purge atthe other end, as illustrated in Figure 3 ‘This unique plug flow design offers significant benefits. First, it allows continuous ‘withdrawal of concentrate without interuption of production. Secondly, high recovery (the fractions of feed recovered as permeate), up to 99% is easily obtained in a single step. ‘Thirdly, the feed water gets concentrated as it flow along the tank and only the downstream cassettes see the most concentrated water. This is illustrated in Figure 4 where the CSTR and PER configurations are compared for feed water containing 10 mg/L of suspended solids.n ‘Membrane Technology In Water and Wastewater Treatment ‘et cecentton nth tan mah Figure4 Concentration gradient in the membrane tank for the CSTR ‘and PFR modes of operation in a 12 cassette tank 4 IMMERSED MEMBRANE OPERATION ‘The history of membrane filtration technology isa succession of improvements where higher fluxes have been sought to reduce cost. However, these efforts have been mostly directed at reducing capital cost of pressurised membrane systems where the membrane modules have low packing density and require a lot of ancillary equipment. When the cost of implementing membrane technology is examined with a life eyele analysis approach, it becomes clear that energy and membrane replacement costs are significant factors. ‘The design philosophy for a ZecWeed® system is to reduce the operation requirements to ‘the level of simplicity of conventional water treatment equipment such as sand filters. Ths is achieved by operating the membrane in quasi-direct filtration (dead-end), at low trans- membrane pressure and moderate fluxes. Gentle aeration agitates the hollow fibres and induces an airlift circulation pattem in the tank. Periodic back-pulses are used to de- concentrate the surface of the membrane. ‘This mode of operation corresponds to low fouling conditions. First, there is no pressure Joss on the feed side of the membrane and the trans-membrane pressure is practically uniform along the length of each hollow fibre since permeate is withdrawn from both ends. ‘Sccondly, operation under suction imposes a practical limit to the trans-membrane pressureWater Quay and Treatment 15 and operating flux, which remain in the stable pressure-controlled region of the filtration eure, ‘These, gentle operating conditions correspond to low energy consumption (0.05 to 0.15 Whim’, depending on the application) and a long membrane life. The mode of operation of a ZeeWeed” plant corresponds to low chemical cleaning requirements of 2 to 6 times per year. Cleaning is always done in-situ, using the back- pulsing unit to deliver the chemical solution to the membrane. The primary cleaner is chlorine, which is a common chemical in drinking water plants; the cleaning mixing water does not need to be heated. Cleaning can be done into an empty tank or a full tank. Empty ‘tank cleaning offers the benefits of using full strength cleaning solutions and generating lower volume of waste. Full tank cleaning offers the possibility of prolonging the soaking time. Membrane integrity is verified with an automated pressure decay test. This test is conducted by pressurising the lumen side of the membranes with air and measuring the decay in air pressure over time. The location of compromised fibres can be determined by ‘observing the stream of air bubbles on the top of the membrane tank. Since the fibres are not ina shell, compromised fibres can easily be repaired in the field. 5 IMMERSED MEMBRANE APPLICATIONS ‘The ZeeWeed® membrane was first commercialised in 1994. The rapid development of the technology is summarised in Table 1, which shows the number of plants operating, and ‘under construction or design. Initial plants were small (< 100 m/d) focused on wastewater ‘treatment using membrane bioreactors (ie. the replacement of clarification by membrane filtration in the activated studge process). ZeeWeed® started to be used for drinking water production in 1996. Currently, Zenon is designing a 100 000-m3/d ultrafiltration plant for Olivenbain, CA. Tablet __ZeeWeed®fl-scale pants (Q3, 19) ‘Number of Plants Plane size (md) | 700 | 100-1 000 [1000-10000 [> 10 UOT Tadustrial = aoa eee Process Water | 1 4 ° |__ Wastewater | 9 2 1 0 Municipal Drinking Water | 1 9 u 6 Wastewater [229 21 1 (Tot a 35 7 Because the ZeeWeed® membrane is resistant to most oxidants and tolerant to high concentration of suspended solids, it can be coupled with a conventional process such as ‘oxidation, coagulation, adsorption, or biological treatment, for the removal of both dissolved ‘and particulate contaminants. Even for a simple direct filtration plant, this flexibility ean be used to modify the level of treatment in response to variable water quality by complementing16 Membrane Technology In Water and Wastewater Treanment ‘membrane filtration within the same unit process. For example, powdered activated carbon ‘ean be added to treat seasonal occurrences of taste and odour or pesticides. The following applications have reached full-scale development: ‘© Direct filtration of surface water; ‘Coagulation coupled with filtration for cold and low alkalinity coloured surface water; (Oxidation coupled with filtration for iron and manganese removal; Concentration of conventional water treatment plant residuals; Pre-treatment of surface water to reverse osmosis; Filtration of tertiary effluents for water reuse; ‘Activated sludge coupled with filtration for wastewater treatment. Many papers that describe the performance of ZeeWeed® in various applications are listed inthe reference section,'?>45#? 6 CONCLUSIONS Today, some ten years after commercialisation in the water field, membrane filtration is a proven technology. In essentially all new water treatment projects, membranes are now considered and are often selected over conventional technologies. However, the universal adoption of membrane technology by the water treatment community may be impeded by the lack of standardisation of configurations and operation modes. These authors submit that immersing membranes in a tank represents the best technological platform for standardisation, forthe following reas ‘© Filtration with hollow fibres from the outside-in is more tolerant to variable feed water quality. ‘Immersed membranes are conducive to the construction of large modules, cassettes and treatment trains, which significantly reduces the requirements for ancillary equipment, and thus translate in lower costs. ‘© Gentle filtration conditions, at low trans-membrane pressure and limited flux lead to low fouling rates and ease of operation. ‘© Immersed membranes are flexible tools that can be used for direct filtration or ‘coupled with a conventional process such as oxidation, coagulation, or adsorption for the removal of both dissolved and particulate contaminants. ‘© Immersed membranes can be used to build new plants or to upgrade existing plants by immersing the membranes directly in clarifiers or sand filters.Water Quality and Treatment n References 1. S. Kroll, B. Vestby, The Production of Ultra High Purity BFW from a Variable Surface Water Supply using Membrane Based Technology, presented at Power Gen International, New Orleans, December 1, 1999. 2. G. Best, D. Mourato, M. Firman, S. Basu, Application of Immersed Ultrafiltration ‘Membranes for Colour and TOC Removal, presented at the AWWA Annual Conference, Chicago, June 22, 1999 3. H, Husain, P. Cté, ZenoGem: The Zenon Experience with Membrane Bioreactors for Municipal’ Wastewater Treatment, presented at the 2" Intemational Meeting on ‘Membrane Bioreactors for Wastewater Treatment ~ MBR 2, Cranfield University, June 2, 1999, 4. P. COté, D. Mourato, C. Guingerich, J. Russel, E. Houghton, Desalination, 1998, 117, 181. 5. T. Lebeau, C. Leliévre, H. Buisson, D. Cléret, L. Van de Venter, P. C6té, Desalination, 1998, 117, 219 6. P. Cété, 8. Monti, L. Bell, D. Bonelli, The ZenoGem Process for Pharmaceutical Wastewater Treatment, presented at the 1AWQ 2" Intemational Conference ~ Advanced ‘Wastewater Treatment, Recycling & Reuse, Milan, September 14-16, 1998. 7. Mourato, D., Benson, M., Carscadden, G., The Role of Particle Count Analyzers as an On-line Integrity Tool for Immersed Membrane Plants, presented at the AWWA Annual Conference & Exhibition, Dallas, June, 21-25 1998, 8. P. Cété, H. Buisson, C. Pound, G. Arakaki, Desalination, 1997, 113, 189.PHOSPHATE AND IRON REMOVAL FROM SEEPAGE AND SURFACE WATER BY MICROFILTRATION JAM, Hofiman', N.C. Wortel?, E.T. Baars'.P. van der Hoek! * Amsterdam Water Supply 2 Grontmij Consulting Engineers Provincialeweg 21 P.O. Box 14 1108 AA AMSTERDAM 3730 AA DE BILT. Phone: +31 20 6510200 Phone:+ 31 30 6943210 Fax: +31 20 6976880 Fax: #31 30 6956366 1 INTRODUCTION Drinking water forthe City of Amsterdam and its surrounding municipalities is produced at two production plants. The total capacity of both plants together is 101 million my. ‘The production plant East, also called the River-L ake Water Works, uses mainly seepage ‘water from the Bethune polder (BP\(25 Mm'/y). In dry seasons an additional in-take from the Amsterdam-Rhine Canal (ARC) is used (approx. 6 Mm’/y). The seepage water from the Bethune polder originates from area with peat-sols, resulting in a relatively high organic content, «high colour, and presence of ammonia, iron and phosphate. The water is pre-treated at treatment plant ‘Loenderveen’ with a double coagulation and sedimentation, storage in lake, rapid filtration. The purpose of the coagulation sys- tem is mainly phosphate and turbidity removal. This is necessary to prevent eutrophica- tion of the storage lake. A low phosphate level (20 + 5 ug/l P) is however desired to ‘maintain the selfpurifying effet in the lake. The storage lake itself has three major ob- jectives: removal of ammonium by nitrification, improvement of hygienic quality and damping of water quality peaks (mainly chloride from the ARC). ‘After pre-treatment the water is transported over a distance of 20 km to the Weesperkar- spel treatment plant were the main treatment to drinking water quality takes place. ‘To fulfil future drinking water demands and improve drinking water quality, research is conducted to optimise the existing treatment process. Also treatment capacity extension is investigated for the River Lake Water Works. To increase capacity, more ARC water will be treated. When the existing treatment plants at Loenderveen and Weesperkarspel ‘were designed in the 70°s it was expected that the treatment capacity would grow 10 a ‘maximum capacity of 61 Mm’ly. Parts of the treatment infrastructure were already built at this final capacity. However, growth of drinking water demand is not as rapid as was expected in the 70's. Therefore capacity extension in small flexible steps (around 2 Mm'y) may be beneficial. Because ofits modular design, membrane filtration offers this fexibility. For the pre-treatment system, microfiltrtion and ultrafiltration technologies offer good alternatives. Ultrafiltration was investigated earlier’. Results indicated good water quality as well asa relatively stable operation. However uitrafiltration showed a relatively high chemical demand (hydrochloric acid, hydrogen peroxide, sodium hypochlorite)Water Quay and Treatment ~» ‘This paper presents the results of indicative study on the application of a Memcor Cross Flow Microfiltration system on Bethune polder water and ARC-water. Characteristic of the Memeor system is its outsde-in filtration type and its air back wash. The study was conducted in close co-operation with Grontmij Water & Waste Manage- ‘ment. The objectives ofthis study were: ‘* removal of phosphate ‘© removal of pathogenie micro-organism ‘© removal of suspended solids fouling behaviour of the membrane modules. 2. EXPERIMENTAL 2.1 Pilot Plant ‘The pilot plant consisted of a small Memcor CMF system, containing three MIOC mem ‘brane modules. Its total membrane surface area was 45 m7”. Air backwash was applied every 25 minutes. This interval was chosen based on experience with the membranes on ‘ comparable source. No optimisation ofthe backwash interval was done. For flux enhancement and phosphate removal, aferic chloride dosing system was pre- sent inthe feed system. Ferric chloride was added directly before the feed pump. The raw ‘water from the ARC or the Bethune polder was prefiltered with « 200 jm microstrainer. 2.2, Experimental Programme ‘The experimental period was divided in two parts: In the frst eight rans Bethune polder ‘water was treated: in the second five runs ARC water was used. During each run the pilot plant was operated ata constant flux with an air back wash every 25 minutes. Each rin continued until the maximum allowable Trans Memibrane Pressure (TMP) was reached. ‘After that, the membranes were cleaned thoroughly, using detergent cleaning agents. Also, the feric dosage was varied, to investigate the necessity for phosphate removal and to investigate its influence on the membrane fouling rate. Moreover, the feed and perme- te water were sampled to stady the water quality ‘The membrane fouling rate was determined by dividing the TMP increase by the run length in days. Table 1 and Table 2 give an overview of the experiments conducted dur- ing both experimental periods. Table 1. Experiments conducted on Bethune polder water pyar a3_[5 [or 3 [06 [os7[0.a6 3119.71.53Py Membrane Technology in Water and Wastewater Treatment ‘Table 2. Experiments conducted on Amsterdam Rhine Canal water li fiz [i3 [i4_ [las Sa aS 60 70 }70 [70 70 [96 35] 37 To fo [kPafday 173 73 9 3. RESULTS AND DISCUSSION 3A Process Stability Figure 1 shows the raw TMP data on the Bethune polder water treatment; results on ARC-water are similar. Each run ean be clearly identified by the “saw teeth” in the raph. It is observed that for each operating condition (ux, feric dosage), different fouling rates exist. In general itis found that fouling rate increases with flux and de- creases by addition of feric chloride to the raw water. Furthermore, it ean be seen that most lines are more or less straight filtration curves, indicating the occurrence of ‘nor- smal" cake filtration 200 ‘TMP (kPa) PPP PPE EEO Ww 8 8 Date Figure 1. TMP for the runs on Bethune polder water. Using these TMP data to calculate the membrane fouling rate and ploting them as function of the applied flux gives good opportunites for optimising the membrane op- eration, The results of this exercise are shown in Figure 2. It was found that the mem- brane fouling rate increases exponentially with the applied fux. Figure 2 shows that for ‘ARC-water, the ferric addition reduces the fouling rate by a factor of approximately 2 For the Bethune polder water the effect of ferric addition i less clear, probably because ‘of the natural iron content ofthe water. ‘The optimum fux can be found by selecting a chemical cleaning interval, eg. 7 days Using the maximum allowable TMP (180 kPa) and the clean membrane TMP (approx. 40 kPa), one can than calculate the fouling rate (20 kPa/d). Figure 2 than gives the flux, which can be applied: for the ARC a flux of 85 Vin’h and 75 Linh can be reached for ‘operation with and without ferric chloride addition respectively. For the BP-water, the flux fr stable operation i around 73 Vin?h.Water Quality and Treatment ar ‘The results shown if Figure 2 indicate that a stable operation with the microitration can be achieved at relatively high fluxes and with low chemical demand. With ultrail- tration the same flux levels could be achieved, but with chemically enhanced backwashes every 3 hours and membrane disinfection every 24 hours ' mierofiltration however has higher energy consumption due to high pressure needed for the air back wash. Exact fig- ures forthe energy consumption however cannat be given on the basis ofthis study, be- cause the further optimisation of the air backwash time interval is than necessary’ Armeterdam Fine Canal Bethune Polder 10 100 z* a i g / g of © 2 Kd zx 2s we 2 ~ oe Fuxtimzn) Pectin) Figure 2. Fouling rate as function of flux and ferric addition for both water types. ‘Moperation without ferric addition; «a operation with addition of ferric chloride. 32 Water Quality 3.2.1 Treated water quality One of the major objectives ofthis stuby was to invest gate whether microfltrtion can be used for phosphate removal. Target value for phos- phate is 20 + 5 g/l P. The levels in the raw water expressed as tota-POs are 100-200 g/l P in the ARC-water and 50-100 ug/l P in the Bethune polder water. The phosphate in the ARC water is almost completely present as dissolved ortho-POs, This means that tan ferric chloride addition is necessary to precipitate the phosphate in order to remove it by the membranes. The optimum ferric dosage was determined by varying the dosing rate and while on-line following the phosphate content of the microfiltrate. Figure 3 shows the results. From this graph it can be seen that a ferric dose of 2.5 mg/l Fe is sufi- cient to remove the o-POs from the water. Further increase of the ferrie dose will not af- fect the phosphate concentration in the microfiltrate For BP water, the phosphate is almost completely bound to the natural ferric in the ‘seepage water. This means that no ferric addition is necessary for phosphate removal Moreover ferric addition did not yield a flux improvement on BP water.2 ‘Membrane Technology in Water and Wastewater Treatment Fane dose tm Fe) Figure 3. Ortho-phosphate concentration in the ARC microfiltrate as a function of the ferric chloride dose, Beside the phosphate concentrations, other water quality parameters were determined. ‘Average values are shown in Table 3. It is shown that microfiltration offers a very good water quality, as expected. Interesting point is that by the ferric addition, also DOC re- ‘moval in the ARC-water was possible. In the BP water no removal of DOC was ob- served, Table 3. Overview of average water quality parameters [Parameter —| Unit | Amsterdam Rhine Canal —| Bethune polder ] aw Filtate {Raw Filtate | Temperature °C $6 10-6 | DOC mg [5.0 45 10 10 Fe me [08 <0. 3 <0. Ma mg —|0.18 0.14 0204 0.1.03 As wet ou nd. ad. PO, nel —|200.300 20450 100-200 <20 0-POs nel =| 140-160 20-40? | 100-150 <10 Turbidity = PTY |12 “05 20 <05 SRC! 'N/100 ml_|700 <1 800-1000 <1 ‘Spores of sulphiteredveing cosidia 2 Depending of ferric dosage rate 3.22 Back wash water quality Water quality of the back wash water and the Clean In Place (CIP) solution after cleaning was determined as well. For treatment of Bethune polder water iron, manganese and arsenic were important. Also DOC and AOX levels ‘were measured because these parameters are required for disposal of the backwash and CIP solution. For the Amsterdam Rhine Canal, also heavy metals are important, since they are present inthe raw water at significant levels.Water Quality and Treatment Table 4. Overview of average back wash water and CIP quality 8 Paramcter | Unit Amsterdam Rhine Canal — Beane polder {Back Wash CIP Back Wash CIP [rarbiaiyy FTO ]150 650 100330" Suspended mg/l [65 65 160 220 Solids \Fe mgt [27 209 30 170 Mn mpl [0.44 03s los 1.02 As wel 2 7 | B |tPO, nglP OL 195 5.6. 215 o-PO, nglP |08 321 {0.02 120.5 Doc myc |145 530 69 320 A0x neta | 140 15 180 SRC N/100 mt 4800 16000 ‘Table 5. Heavy metals in back wash water and CIP of ARC [Parameter [Unit [amsterdam Rhine Canal { ‘Aluminium | {Barium | Cadmium {Chromium | Copper Lead {Nicker | Vanadium {zine ‘Table 4 and Table 5 indicate that the back wash water contains a large amount of the fl- tered contaminants. However the CIP’s, using a caustic surfactant solution, show that even more contaminants are released from the membrane surface. Because the CIP's ‘were able to restore the original clean membrane flux itis assumed that they were effec tive, 4 INVESTMENT AND OPERATING COST Based on the above results an estimation of investments and operation and maintenance costs was made for a extension of the River-Lake Water Works with 37 Mm°/y. The total investment cost for this plant are estimated as 24.8 million Df (MBuro 11.3). This re- sults in an O&M costs of Dfl 0.13 per m’ (Euro 0.06 perm?) 5 CONCLUSIONS From the results it is concluded that microfiltration offers good possibilities for phos- phate and turbidity removal from seepage and surface water in a cost effective way. Compared to ultrafiltration and conventional coagulation technology only limited amounts of chemicals are used, mainly due to a very effective air back wash system. A4“ ‘Membrane Technology In Water and Wastewater Treament stable operation could be achieved during one month of experiments on each source. Ad- dition of ferric chloride of 2-4 mg/l to the ARC water resulted in factor 2 reduction of the fouling rate of the membranes. For optimisation for a full-scale design more and long term research is necessary. The results of this study will be used in a multi-criteria analysis for optimisation and extension of the River-Lake Water Works of Amsterdam Water Supply. References 1. J.AMH, Hofinan, MM. Beumer, ET Baars, P. van det Hock, H.M.M. Koppers, Enhanced surface water treatment by ultrafiiration, Desalination, 119 (1998) 113 125 2. JAMH. Hofman, C.A. Groot, 1P. van der Hoek, Utrafitration as alternative for coagulation and sedimentation for surface water treatment, Proceedings of Wter- symposium’99, Breda Apri 13th 1999. (In Dutch)REUSE OF FILTER BACKWASH WATER AS A SOURCE FOR DRINKING WATER PRODUCTION: PILOTING AND IMPLEMENTATION OF A FULL-SCALE ULTRAFILTRATION PLANT A. Brigger, K. Vossenkaul, T. Melin, R. Rautenbach Institut fiir Verfahrenstechnik, Rheinisch-Westfilisch Technische Hochschule Aachen ‘Turmstrabe 46, 52062 Aachen Germany B. Golling, U. Jacobs, P, Ohlenforst Stadtwerke Aachen AG, Postfach 500155, 52085 Aachen Germany L INTRODUCTION ‘The drinking water treatment plant Hitfeld of the Stadtwerke Aachen AG (STAWAG), Germany, processes 2.7 Mm? reservoir water per year. After flocculation and pH adjustment the raw water passes recirculation filter and a dual-media sand filter. Backwashing of these filters consumes about 10 % of the treated water. Since the processing capacity of the treatment plant exceeds the annual contingent of raw water from the dam, standstlls of the plant may occur in dry years. In order to achieve a better plant utilisation and to avoid the costs of the wastewater, recycling of the filter backwash ‘water is highly desirable. Before recycling, the filter backwash water is stored in a basin for sedimentation. Until 1997 the clear water from the settling tank was then directly returned to the raw water. In compliance with a recommendation of the Drinking Water Commission of the ‘Umveltbundesamt (Ministry for the Environment) the reuse of this water as raw water ‘was stopped and the clear water was drained. Today the reuse of filter backwash water requires the separation of solids and moreover the complete removal of microorganisms and parasites. ‘The required secure retention of microbiological parameters cannot cost effectively be guaranteed by means of conventional technologies. The discussion of reusing the Dackwash water today often results in considering the use of membrane technology’. ‘Therefore on behalf of STAWAG a pilot-scale ultrafiltration plant was operated by the Institut fur Verfahrenstechnik (IVT) in co-operation with Rochem UF-Systeme GmbH, Hamburg (Rochem). The promising results of the pilot study led STAWAG to decide to treat the backwash water with a full-scale ultrafiltration plant. Here, results from the pilot study as well as first experiences with the full-scale plant will be presented.6 Membrane Technology in Water and Wastewater Treamnent 2 PILOT-SCALE EVALUATION 2.1 Research Objectives From the middle of June until the middle of August 1998, a pilot plant was set up in the drinking water treatment plant in Hitfeld in order to investigate the operational possibilities and the efficiency of ultrafiltration for the treatment of filter backwash water. ‘The pilot plant was bypassed to the clear water disposal. Two full-scale modules treated corresponding tothe incorporation of the future full-scale plant (Figure 1, the clear water. cite ran | Maer ttn 4 oF Figure 1 Incorporation of the membrane plant into the drinking water treatment process Only working under real process conditions, including water quality and deily standstils caused by the necessary time for filter backwash and setling, give reliable results to the questions regarding the possibility of the treatment of the filter backwash water with the Use of membrane technology. The aim ofthe plot study was to provide results about the efficiency of a membrane plant, the optimisation of the plant operation and the ‘establishment of design data for a full-scale plant. Therefore, the following questions had to be answered: Which membrane material is most suitable for this application and can guarantee the production ofthe desired quality of filtate in the long term? Which flux can be obtained ‘ona long-term basis? ‘What are the optimum conditions forthe operation of the plant and what recovery rate ‘can be obtained? How can the module flushing be fitted on to this application and how often are chemicals required for cleaning or disinfection? ‘How does the daily standstill ofthe plant influence its operation? 2.2 The Pilot Plant Rochem developed the FM module system applied for the piloting in co-operation with the IVT. The principal elements of the FM module are membrane cushions, eachWater Quality and Treatment a consisting of two rectangular membranes; two intemal permeate spacers, and an incorporated carrier plate, which are welded by an ultrasonic method on the outer edges (Figure 2). A membrane element consists of these membrane cushions, which are stacked ‘one on top of the other supported by two pins with a gap in between. These cushions are enclosed in two halves of a shell. In addition to fixing the membrane stacks, the guide ins in combination with comespnding bore oles inte halael elements, cary of he permeate’. Yara offedchana 4-3 mm Figure 2 Rochem FM Module System Inthe pilot plant eight of these membrane cushions elements are connected in a series ‘and installed in a PMMA pressure tube. The pilot plant consists of two modules each fitted with about 7m? membrane area. Both modules can be operated separately and are installed vertically to allow for effective air flushing. In order to optimise the plant operation the pilot plant can be run in manual or automatic mode, where values for the ‘main process parameters can be chosen separately for each module, Since solid matter concentration in filter backwash water — especially after sedimentation — is quite low, dead-end mode was chosen for this application. This means that during filtration the water is flowed perpendicularly through the membrane. The ‘module is run as a two-end-module. The retained solid matter forms a fouling layer on the ‘membrane surface and, gradually, builds up a flow resistance, which adds to the flow resistance of the membrane. To avoid a decreasing permeate flux, the pressure difference hhas to be increased during the filtration cycle. ‘When the filtration cycle is complete the fouling layer has to be removed, i. e. the ‘module has to be cleaned. The long-term stability of a dead-end filtration process relies on the efficiency of the cleaning procedure. The configuration of flat channels in the FM ‘module allows a very efficient cleaning method consisting of a combination of feed sided air bubble flushing and back washing of permeate through the membrane. The air injected into the raw water at the base of the vertically installed modules induces high shear stresses and removes the fouling layer from the membrane surface. The subsequent short period of cross-flow flushing cartes the detached particles out ofthe module. 2.3 Results of the Pilot Study The pilot study was begun in early June 1998. At the beginning, both modules were88 Membrane Technology tn Water and Wastewater Treatment quipped with polyaramid (PA) membranes. At the end of July polyacrylonitrile (PAN) ‘membranes replaced the membranes in one of the modules. Both membrane materials had cut-off of 50D. ‘The investment costs of dead-end membrane plants make up for the biggest part ofthe total specific treatment costs. Thus, increasing the plant capacity by enhancing the permeate flux means a substantial reduction for the total treatment costs. Therefore, the ‘most important aim of the pilot study was to find out the highest possible flux at stable conditions with optimised process parameters. The parameters, which needed to be optimised, were the filtration time and all parameters related to the module flushing. ‘The start up of the pilot plant occurred ata relatively low flux performance. The fx could be raised by gradually adjusting the module flushing parameters to this application ‘Figure 3). Reasons for drops of the transmembrane pressure-difference (TMP) at constant fluxes were usually due to extra module flushes with air and water which were carried out in order to optimise the module flushing. As can be seen in Figure 3 it was possible to increase the performance ofthe PA membrane to over 150 Vin?h. ‘After membrane replacement the flux had to be reduced due to a rapid increase of ‘TMP. Obviously the PAN membrane material can only be run at a somewhat lower level of flux. Applying a relatively high TMP a flux of about 115 Veh was possible at stable conditions with the PAN membrane. During dead-end filtration the amount of produced filtrate equals the content of feed. Only the filtrate used for backwashing and the feed used for the cross-flow flush reduce the recovery rate. After a short start-up phase the plant reached a recovery rate of about 96 %. Recovery rate as well as energy consumption and the net permeate flux depends on the filtration time. At the end of the pilot study the filtration time for the PA membrane ‘was 120 min and for the PAN membrane 100 min, altogether resulting in a high net flux and recovery rate at ow energy consumption (Table 1). It must be mentioned at this point, that during the three months of piloting, no chemicals were used for backwash enhancement or for chemical cleaning. The high fluxes ‘were achieved only by adapting and optimising the module flushing procedure for this application, “The membrane material chosen to equip the full-scale plant was the PAN membrane. This was due to doubts as to whether the membrane cushions made from PA could ‘guarantee long term integrity. Especially the welded edges of the PA. membrane cushions ‘were susceptible to damages. It was observed during the pilot study that the performance of the PA membranes regarding permeate quality was inferior to the PAN membranes. ‘The quality of the permeate was determined by continuous particle counting and by plate counting. Although both membranes always showed drinking water quality and low particle counts there was a difference between the two materials. The retention of CFU for PAN membranes was about one log-stage better, and particle counts (1 jum) for PAN were always below 0.5 particles per ml while the counts for the PA membrane were between 0.5 and 1.5 particles per ml.Water Quality and Treatment » Figure 3 Membrane performance in the pilot study ‘Table 1 Operating data ofthe pilot plant Permest tow rte tzaon um) 128000 Mombrane mati 1 PAEDKDPANSOIO Tempera of am war (average) rq 0 Dxeaton of aon ee (ny t20 100 acaton of tush (ony SA Memorane area perme (1 er 72 Transmembrane pressure diterence oa) oa 175 Pormano fax umn) 190 18 Recovery Dae > i863 sar Nt permeate Bux rey tat 14 Enegy consumeton iweranry 08 08 3 DESIGN AND COST CALCULATION FOR THE FULL-SCALE PLANT 3.1 Design Data ‘As a consequence of the positive pilot results STAWAG decided to employ the membrane filtration technology to recycle the filter backwash water. It was agreed that STAWAG. would buy, ata fixed price, the filtrate treated by the plant owned by Rochem UF-Systerhe GmbH, and operated for further optimisation by the IVT. The plant capacity was fixed to ‘be 10.000 m? per month, ie, about 350 m? per day. The daily operating time was assumed to be about 18 hours. ‘At the time when the plant design took place, the manufacturer of the PAN membranes ‘was offering the same material used in the pilot plant with higher cut-off values. Thus, an.0 “Membrane Technology in Water and Wastewater Treament optimistic value forthe permeate flux forthe plant design seemed to be realistic and it was assumed that a flux of about 100 Vim'h could be achieved (Table 2). As a result the required membrane area is 200 m? installed into 20 modules. ‘Table 2 Design data for the full-scale UF- plant Daly retest ty a, 380 praing hous par doy mwa 1° ‘capacity (mm 184 Membrane ares por mode ro 10 Number ot made 8 20 Ente marbrane area ima 200 ipod etme @ 4 Permatex ‘un or Firaton cycle tri » Recovery rs 96 Energy consumption owe 02 ‘operating pressure (oar 10 3.2 Cost Caleulation ‘The calculation of the specific costs for the ultrafiltration treatment of filter backwash ‘water is based on a time of depreciation of 10 years. The characteristic value of investment ‘of 1900 DM per m? installed membrane arca represents an average value of comparable plants offered by different manufacturers. The specific costs for chemicals and ‘maintenance are based on empirical coefficients. ‘The total specific costs are about 0,6 DM/m’ (Table 3). These costs are relatively low for the treatment of filter backwash water, relying on realising the high assumed permeate flux. According to this calculation the depreciation contributes about 70% of the total costs. This is characteristic for membrane plants operated in dead-end mode. 4 FIRST EXPERIENCES WITH THE FULL-SCALE PLANT In February 1999 the full-scale plant was set up and operation was started. The plant consists of two blocks with ten modules each. At start-up the plant was equipped with about 200 m* of the PAN $0 KD membrane. The permeability of this membrane forthe first four month in operation is shown in Figure 4. ‘A breakdown of the permeate backwash pump after 24 days resulted ina sharp decline in membrane permeability. For one weekend the efficiency of the module flushing was negligible. Since one ofthe modules was equipped with a PMMA pressure tube it was possible to see thick fouling layers on the membranes which built up over the weekend, ‘Attempts to recover the former membrane permeability by extra module using failed. Even further improvement of the module Mushing by adapting the procedure to theWater Quality and Trearment on conditions of the full-scale plant could not recover the membrane permeability to the former level. Table 3 Estimation of cos for the full-scale plant a Ti of precon a 10 iret va 7 capt te ra tae ew otcpacy oo % acy er or 127000 (oat err pcre win 0 ety cost paar oz Investment cost Choc vie otivetmert owing 00 over cost pa set000 Dejan pay 06 aa Bue one ‘Chol, nonce oe aeisie) OM]. Replacement of matvanes Phim oar Speci ovr cost hit osto. ‘Thus, after about 3 month in operation a chemical cleaning was carried out. Preceding tests at the IVT led to the decision to use an alkaline cleaner and NaOCl as cleaning agents. Permeability after chemical cleaning was lmost equal to the start-up value. Applying the improved module flushing it was possible to achieve a stable permeability of ‘30 Vim*h bar for the PAN SO kD membrane. In parallel to the full-scale plant the pilot plant was operated with a new PAN. membrane with a cut-off of 200 kD. This membrane not only showed a very high permeability but also the same reliable retention for particles (>1 jm) and CFU as the 50 KD membrane. Therefore, in one block of the full-scale plant the 50 KD membranes were replaced with 200 kD membranes. This membrane was operated at a flux of about 105 \Vimh with a permeability of nearly 80 Vin*h bar. With this membrane itis now possible to achieve the plant productivity as required (Table 4). 3 SUMMARY Secure retention of microbiological parameters and high fluxes, which allow economical plant operation, are essential for successfully applying ultrafiltration for recycling filter backwash water. High fluxes on a long-term basis in this case were achieved by improving. and adapting the module flushing procedure to this application. As a result, the use of any chemicals is restricted to a chemical cleaning every 3-4 month.2 ‘Membrane Technology in Water and Wastewater Treament er ar nis Figure 4 Membrane permeability in the full-scale plant ‘Table 4 Development of productivity of the full-scale plant 11.03.09 akdonn pate backs puro fos ee ‘.0360° ~ Rassain of pest basa pire 3 ‘8 mong) pened mode ng « x0 240599 Chania aig 6 20 0899 rowed nadie iting ra 8 19.0780 Nw 20D nates nee Bock 19 28 ‘07.89 Radutin feat tie: 0 x References 1. P. Lipp et al., Desalination 119 (1998) 133-142 2. R. Rautenbach, K. VoBenkaul, Proceedings "2. Aachener Tagung Sedlungswasser- wirtschaft und Verfahrenstechnik: — Membrantechnik in der ffentlichen Wasseraufbereitung und Abwasserbchandlung”, Aachen, 1998, A18 3. T. Peters, Umwelt Technologie Aktuell International Edition, 4 (1998)IMPROVED PERFORMANCE OF DRINKING WATER MICROFILTRATION WITH HYBRID PARTICLE PRE-TREATMENT T. Carroll and N. Booker CSIRO Molecular Science, Private Bag 10, Clayton South MDC, VIC3169, Austral L INTRODUCTION Microfiltration is particularly effective forthe removal of particulate contaminants, such as bacteria, algae and protozoa from drinking water. However microfiltration is ineffective for the removal of dissolved contaminants such as natural organic matter (NOM). Residual NOM jin drinking water is a cause of colour, disinfection by-products, and microbial regrowth", Residual NOM in microfiltration processes is a cause of membrane fouling’. ‘The conventional approach to removing NOM from drinking water is chemical coagulation and flocculation, This approach is now routinely used in Australia as a mierofiltration pre-treatment to enhance colour removal and to reduce the rate of membrane fouling. However, there is considerable scope for further improvement by ‘combining chemical coagulation with adsorption of NOM onto a solid particle. The adsorbent could be selected on the basis of low hydraulic resistance to filtration, and could be used under conditions in which it acts asa site for flocculation. The hybrid adsorbent- coagulant particle formed in this way would then be removed by microfitration in the conventional way, depositing a more permeable filter cake on the membrane and producing a higher quality drinking water than coagulation alone. ‘Magnetite is a commonly mined mineral in Australia with potential as an adsorbent in a hybrid microfiltration pre-treatment. Magnetite is a magnetisabe iron oxide used alone or in conjunction with alum or polyelectrolyte in an existing treatment process to remove particulate and some dissolved contaminants fom drinking water’. It has @ low surface area (23mg) and therefore requires low contact times (< 6) bu relatively high doses (> Ig/L). In addition, particles are heavy (&.. ~ 5.1), and can be magnetically flocculated to settle rapidly (> 30mv/hr). The adsorbent is easily regenerated by a caustic wash. These properties would facilitate in-line adsorbent dosing in an existing microfiltration plant. ‘Removal of the flocculated adsorbent is simple and rapid, reduces any additional particle loading to the membrane, and aids recovery for reuse. The potential of magnetite as an adsorbent particle for use in combination with alum and polyelectrolyte as a ‘mierofiltration pre-treatment for Australian surface water is investigated inthis paper.9 “Membrane Technology in Water and Wastenater Treatment 2 METHODS ‘The raw water source was the Mooraboo! River near Anakie, Australia. This water has a relatively high total organic carbon content (8-10mg/L) and relatively low turbidity (3.9 NTU). The water was treated in standard jar tests with various combinations of magnetite (Fe,0,), alum (AL(SO,),.18H,0), and LT-20 polyelectrolyte (a high molecular weight, non-ionic polyacrylamide). The jar test procedure was to adjust the pH of 1.00kg of raw water from 7.5 to 6.0 with sulphuric acid. The water was then dosed with magnetite, alum, and polyelectrolyte as required. After each chemical addition the water was stirred for 60 at 250RPM. The pH was maintained at 6.0 throughout the procedure with sulphuric acid or sodium hydroxide. Once dosing was complete the mixture was stired for 90s at 1OORPM. The treated water was analysed for dissolved organic carbon (O/l Analytical 1010 Wet Oxidation TOC Analyser). All samples for water quality were filtered through a 0.45jm filter (Selby-Biolab HPLC-certfied) prior to analysis. In hybrid magnetite-alum treatment cases, the extent of magnette-alum attachment was measured from the residual aluminium content in the treated water by inductively-coupled plasma (Jobin-Yvon JY24 ICP spectrometer) after the jar contents were allowed to settle for 150s. Microfiltration experiments were performed on the single-fibr filtration rig shown in Figure 1. The membranes were polypropylene hollow fibres with a nominal pore size of 0.2um, an intemal diameter of 250m, an extemal diameter of 550um, and a length of 1.2m. Fibres were sealed at the open ends with a silicone septum, wet with ethanol and flushed with pure water. The feed water was pressurised with a peristaltic pump, and forced through the hollow-fibre membrane to emerge as permeate from the open ends of the fibre. The permeate was collected in a vessel mounted on an analytical balance. A pressure transducer measured feed pressure. The signals from the analytical balance (WT) and the pressure transducer (PT) were processed to calculate the permeate flowrate as a function of permeate throughput. The feed water was treated using the standard jar test procedure described above, although the 100RPM stirring regime was omitted. The feed suspension was stirred at 250RPM and the pH was controlled at 6.0 throughout the ‘membrane fouling experiments, which took 2-3 hours. In some cases, the feed suspension ‘was allowed to settle, and only the supematant was pumped onto the membrane. Figure 1 Single-fibre microfiliration rig.Water Quality and Treatment 9s 3 RESULTS AND DISCUSSION 3.1 Magnetite-Alum Pre-treatment: NOM Removal and Membrane Performance ‘The dissolved organic carbon (DOC) concentration of Moorabool water treated with various doses of magnetite and alum is shown in Figure 2. The optimum alum dose (at Og/L magnetite) for NOM removal was approximately 3.2mg/L as Al”. The DOC removal at this dose was 44%, and further addition of alum did not increase DOC removal. The combined magnetite-alum pre-treatment improved DOC removal,over alum alone at alum doses below the optimum 3.2mg/L Al and magnetite doses of I/L o higher 100. : 9.0) 2 io i a 00gL ovis 70 ae Log: 60. —O— l0gt. 00. ao 10 20 30 40 50 60 Alum dose (mg. A") Figure 2 Dissolved organic carbon of Moorabool water as a function of alum and ‘magnetite dose. ‘When magnetite was added before alum at pH 6, the alum floes formed on the magnetite particles. This attachment occurred at relatively low magnetite doses, as shown in Figure 3. At an alum dose of 3.2mg/L Al, a magnetite dose of 0.01g/L was sufficient to attach 34% of the alum flocs formed (as AI). At the optimum alum dose, and magnetite doses ‘of O.1g/L. and above, the concentration of unattached alum floes dropped below the ‘concentration of aluminium inthe raw water. Magnetite can therefore serve as a simple but effective carrier particle for alum floc thus reducing fouling ofthe membrane. ‘The membrane filtration performance ofthe untreated, alum-pre-reated, and magnetite- slum pre-treated water are shown in Figure 4. Microfiltration alone removed 11% of the DOC but the rate of fouling was relatively high; 35g of water was filtered before the permeate flowrate declined by 25%. Alum pre-treatment (3.2mg/L Al) inereased the DOC removal to 44% and reduced the rate of fouling; 110g of water was filtered before the permeate flowrate declined by 25%. The rate of fouling was not further reduced by magnettealum pre-treatment at magnetite doses below 0.Ig/L. However substantial9% Membrane Technology in Water and Wastewater Treatment reductions in the rates of fouling were achieved for magnetite doses above L.Og/L. At a magnetite dose of 1.0p/L, 270g of water was filtered before the permeate flowrate declined bby 25%. Ata magnetite dose of 10g/L, 460g of water was filtered before a decline of 25% 4 faction of alum and magnetite 00s @ Ole A O0lgL y LOgL. e @ let Too Permeaie throughput (2) Figure 4 The decline in permeate flowrate as a function of throughput for Moorabool ‘water treated with 3.2mg/L of alum (as AP’) and various doses of magnetite. Although the rate of membrane fouling was reduced by a combined magnetite-alum pre- treatment, there was stil a substantial decline in permeate flowrate with throughput during ricrofiltation. Two possible causes of this decline are:Water Quality nd Treatment a 1. Residual NOM which is not removed by the hybrid pre-treatment subsequently fouls the membrane. 2. Particles and flocs generated in the course of treatment offer a hydraulic resistance to filtration when retained by the membrane. To distinguish these two contributions, the alum-pre-treated and magnetite-alum_pre- treated waters were allowed to settle before microfilration and the supematant was pumped onto the membrane. The membrane filtration performance of the settled pre- treatments are shown as dashed lines in Figure 4 for alum only, and alum with 1.Og/L and 10p/L of magnetite. There was no reduction in the rate of fouling in any of these cases relative to the corresponding stirred pre-treatments indicating thatthe decline in permeate flowrate must be due to suspended or dissolved material in the treated water. This material could be residual NOM or unattached alum pin flocs. Fouling by unattached alum flocs should be higher in stired than in settled pre-treatment, and this was not observed. Furthermore, from Figure 3, the concentration of aluminium from which these floes were constituted was relatively low (less than 0.5mg/L Al) after magnetite-alum pre-treatment and setting. ‘The reduction in the rate of membrane fouling achieved using a combined magnette- ‘alum pre-treatment is compared to alum pre-treatment alone in Figure 5. The basis for comparison is the permeate throughput corresponding to an arbitrary 25% decline in permeate flowrate (as an indicator of the throughput which could be achieved before intervention to recover performance would be required). On this basis, there was io reduction in the rate of fouling over alum for hybrid pre-treatments with doses of magnetite up to Olg/L. However at a magnetite dose of 1g/L, a 2.5-fold increase in throughput was achieved, and at a magnetite dose of 10g/L, a 4-fold increase in throughput ‘was possible. 5 : ] 3.2mg/L alum (as AP*) Figure 5 The increase in relative permeate throughput for Moorabool water treated with 3.2me/L of alum (as AP") and various doses of magnetite. 3.2 Magnetite-Alum-Polyelectrolyte Pre-treatment: Membrane Performance ‘The membrane performance was also investigated when a polyelectrolyte was added in addition to alum and magnetite. A polyelectrolyte may become necessary in full-scale processes to reduce magnetite-alum detachment and break-up under ubiquitous shear98 Membrane Technology in Water and Wastewater Treatment forces. The rate of fouling was unaffected at polyelectrolyte doses below 0.6ma/L but increased at higher doses. The reduction in the rte of fouling achieved using a combined ‘magnetite-alum-polyelectrolyte pre-treatment is compared to alum pre-treatment in Figure 6. There is no improvement in the rate of fouling over alum alone for hybrid treatments with doses of magnetite up to 0.1g/L. However at a magnetite dose of g/L, a 3-fold increase in throughput was achieved, and at a magnetite dose of 10g/L, a 4.5-fold increase in throughput was possible. ‘eam D00gCV.0IeL* Cig LoeL* Tost. “Magpetite dose Figure 6 The increase in relative permeate throughput for Moorabool water treated with 3.2mg/L of alum (as AP”), 0.2mg/L of polyelectrolyte, and various doses of magnetite. 3.3 Magnetite-Alum-Polyclectrolyte Pre-treatment: Cost Benefit Analysis, ‘The increased costs of the magnetite-alum pre-treatment are compared to the reduced cleaning costs for a microfiltration plant using only alum pre-treatment in Figure 7. The comparison was based upon the assumption that the benefit of higher throughput is reduced chemical cleaning frequency. Backwash frequency (and the associated downtime, reduction in product water yield, and energy consumption) was assumed to be unaffected. ‘The capital and operating costs of the magnetite recovery plant were not considered. The cleaning costs are taken from a 6MLiday microfiltration plant operating with alum pre~ treatment at Cresswell, near Melbourne, Australia. The additional treatment costs include sodium hydroxide for magnetite regeneration, polyelectrolyte for magnette-alum flocculation, and magnetite to replace unrecoverable losses. From Figure 7 a doubling of ricrofitration membrane throughput is required before the savings in cleaning costs balance the increased cost of the magnetite-alum treatment. From Figure 6, it can be seen ‘that this throughput increase was achieved with a magnetite dose of 1.0 g/L at an alum dose of 3:2mg/L Al.Woter Quality and Trearment 9 g 80. Bol ae ss i: g 7 ‘Cleaning saving i 0 Throughput (eave to alum) Figure 7 The comparison between reduced cleaning costs and increased treatment costs associated with a hybrid alum-magnetite pre-treatment, 4. CONCLUSIONS. A significant. improvement in the laboratory-scale performance of drinking water rmicrofiltration was achieved by combining a magnetite treatment step with conventional ‘lum treatment at pH 6.0. The permeate throughput at which the permeate flowrate declined by an arbitrary value (25%) increased by up to a factor of four for magnetite doses up to 10g/L, although the water quality after magnetite-alum treatment was not substantially beter than that achieved with alum alone. This improvement would translate into a four-fold increase in water production before the criterion for membrane cleaning or ‘other remedial action is reached. The improvement in membrane filtrtion performance ‘was independent of whether setling preceded filtration, or whether the treated water was filtered directly as a suspension. This may not be the case for an alternative water source, although settling was rapid, allowing a reduced particulate loading to the membrane if required. The rate of fouling after magnetite-alum pre-treatment was also unaffected by dosing with trace quantities of LT-20 polyelectrolyte, a flocculant aid which may be necessary to bolster the shear resistance of the magnetiteaum particles in a larger-scale ‘treatment process. The hybrid alum-magnetite pretreatment process. is presently undergoing pilot trials on a 40kL/day microfiltration pilot plant at Yerring Gorge, near ‘Melbourne, Australia References |. IJ. Rook, Wat. Treatment Exam., 1974, 23, 234. 9. D. van det Kooij, J. AWWA, 1992, 842), $7. ‘A.B. MacCormick in Modern Techniques in Water and Wastewater Treatment, eds. LO, Kolarik and A.J. Priestly, CSIRO Publishing, Australia, 1995, p. 45. 4, N.A. Booker, L.O. Kolarik and R.B. Brooks in Modern Techniques in Water and ‘Wastewater Treatment, eds. L.O. Kolarik and A.J. Priestly, CSIRO Publishing, Australia, 1995, p.25.RIVER TRENT ON TAP - COMPARISON OF CONVENTIONAL AND MEMBRANE ‘TREATMENT PROCESSES BEE. Drage and J.E. Upton P. Holden and J.Q. Marchant ‘Severn Trent Water Ltd, Anglian Water Ltd, St Martins Road, ‘Thorpewood House, Coventry, ‘Thorpewood, UK Peterborough, UK 1 ABSTRACT Severn Trent and Anglian Water serve over 13 million customers in central and eastem England. Following revent droughts both companies see the River Trent as a potential ‘new drinking water resource capable of safeguarding existing supplies in the East ‘Midlands and Lincolnshire. ‘The River Tren, hitherto an ‘untapped! resource for potable supplies, contains the treated sewage from 4 million people together with the influence of industrial discharges from the Midlands means that the river quality can be highly variable. In particular high levels of ‘pesticides, flame retardents, boron, bromide and nitrate in the Trent render it difficult to ‘reat using conventional water treatment processes. Altemative treatment options such as ‘ultrafiltration and reverse osmosis have been identified and the results of preliminary trials used to develop a programme of pilot plant studies undertaken as a joint venture between Sevem Trent and Anglian Water. Results to date compare the treatability of key risk ‘compounds by conventional and membrane treatment processes and the effect on the formation of disinfection by-products at the elevated bromide levels often found in the ‘Trent, 2 INTRODUCTION The River Trent drains the large urban catchment of the Midlands. The industrial and domestic waste from Birmingham, Nottingham and Leicester, in addition to run-off from highways and agricultural activities, all contribute to the contamination of the River Trent, Historically the river has always been considered to be ‘too polluted’ for drinking water abstraction but, due to the improvement in quality of its tributaries and the tighter standards and compliance of sewage effluent, the river has recently undergone a marked quality improvement. In the light of the above quality improvements and the projected need for security of water supplies in the East Midlands and Lincolnshire, the River Trent was identified as aWater Quality and Treament 101 potential source of potable water in early 1996. Severn Trent Water being imerested in Utilising Trent water at their existing works situated between Nottingham and Derby (Church Wilne WTW) and Anglian Water for a new works close to Lincoln. There was then a need to identify a suitable water quality monitoring and treatment strategy capable of producing final water to meet all curent and potential future drinking water standards This paper describes a series of pilot plant studies designed to compare conventional treatment of River Trent water with treatment using state of the art membrane technology. 3. DESIGN OF THE ADVANCED WATER TREATMENT PLANT The Advanced Water Treatment Plant (AWTP) was designed following preliminary pre- treatment and Reverse Osmosis (RO) studies. Spiking trials on single RO elements were carried out to compare_ the rejection of Trent risk compounds (boron, bromide, nitrate and isoproturon) achieved by both High Rejection (HR) and Ultra Low Pressure (ULP) RO membranes, There was some deterioration of rejection achieved by the ULP RO (55 to 40%) for boron removal but all other spiking compounds were removed by >95%. ULP RO clements were therefore selected for the AWTP in view of the energy savings when compared with running HR elements. In pre-treatment pilot trials the performance of self cleaning, fine screen filters was compared with a self cleaning sand filter (Dynasand) prior to UF. Incoming turbidites between 1 and 120 NTU were experienced, with only the Dynasand filter producing filtrate of sufficient quality in all events ‘The AWTP was commissioned in May 1998, and comprises a 350 mid “Conventional” stream (based on the flowsheet of Church Wilne WTW) running parallel to 2350 md “Membrane” stream (Figure 1). Both streams can also make use of ozone and powdered activated carbon (PAC) and the modular nature of the individual processes allows them to be interchanged. Thus, the initial direct comparison will be followed by the optimisation of a process for the River Trent, possibly by combining processes from. the two flowsheets Figure 1 Schematic of Advanced Water Treatment Plant02 ‘Membrane Technology in Water and Wastewater Treatment The plant is monitored and to some extent controlled using an ‘In-Touch’ SCADA package, which also provides historical data trending and can be accessed remotely. Weekly analysis and spiking trials with problem determinands is testing the performance ofthe various processes, allowing an optimum treatment process forthe River Trent water to be developed, based on final water quality and operating costs. Feed waters to the Advanced Water Treatment Plant (AWTP) are available from four Afferent sources: 3 abstraction from the River Trent + Witches Oak Waters (River Trent from bankside storage) ~ Church Wilne Reservoir impounded Derwent) = River Derwent (pumped from the Church Wilne WTW reservoir intake line) ‘The plant can treat a single feed water, ora blend of Trent and Derwent waters, mixed in the inlet break tank. This is important to both Water Companies involved as Severn Trent ‘Water intend initially to treat blended water from the two rivers and Anglian Water will take the Trent downstream of the TrenvDerwent confluence. 341 Conventional Treatment Processes ‘The conventional treatment stream is based on the Church Wilne WTW flowsheet and comprises clarification, followed by four rapid gravity sand filters (RGF) and four ‘granular ectivated carbon (GAC) adsorber. ‘The cone shaped upflow Clarifier is designed to process 15m'h of water from the inlet break tank, and uses ferric sulphate coagulant and a starch based polyelectrolyte as a coagulant aid. A height adjustable sludge cone collects the floc blanket and the clarified ‘water overflows under gravity to the RGF header tank. Here it is distributed equally to the four 4.4m (height) x Im (diameter) filters, each of which contains sand and anthracite on a gravel support. The RGF are backwashed every 48 hours or on head loss. The filtrate from all four filters enters a common line and flows to a break tank, which is used both to backwash the RGF and feed forward onto the GAC via a relift pump. The four 6m (height) x Im (diameter) GAC columns are fed from a common header tank and designed to process a combined flow of 14m, the depth of Chemviron F400 activated carbon in cach adsorber resulting in a designed empty bed contact time of 30 minutes with all four units in use. 32 Membrane Treatment Stream ‘The membrane stream comprises a self cleaning sand filter (Dynasand), an ultrafiltration skid and a reverse osmosis skid Raw water is pumped to the bottom of the DST-15 Dynasand and filters up through approximately 3m of 1.2-2mm sand, the filtrate flowing over a weir atthe top. The sand is cleaned and recycled by means of an air li pump running from top to bottom through the centre of the Dynasand. Air, injected at the lowest point of the unit, causes a ‘water/sand/dirt mixture to rise to the top of the unit where it emerges in an open bottom vessel. Filtrate flows upwards into this vessel and escapes over a weir to waste, carrying the dirt with it, the sand meanwhile sinks back down on to the top of the sand bed. The ‘Dynasand filtrate overflows to a break tank from which the Ultrafiltration skid draws its feed.Water Quality and Treatment 103 ‘The Norit MT Ultrafiltration plant (containing X-flow Magnum hollow fibre membrane elements) is designed for a net daily production of 350m’, at a flux of approximately 80lm 2". The unit operates in dead-end mode and is taken off line automaticaly at pre-set time intervals to backwash or chemical clean (using sodium hypochlorite or nitric acid) to maintain a given trans-membrane pressure at >90% recovery. The filter material is polysulphone based with a nominal molecular cut off of 150-250kD and reduces the turbidity and suspended solids in the permeate to below the limits of detection. The ‘permeate quality is therefore consistent, and any breach in the integrity of the unit can be identified by a detection of turbidity, in particle counts or an increase inthe particle index. The UF permeate is collected ina 10m’ break tank, which is used to backwash the UF and feed the reverse osmosis skid. ‘The Reverse Osmosis skid contains fluid systems polyamide, spiral wound, TRC-ULP 4” membranes and operates in ether 4:2:1 ot 43 configuration. The unt is designed to treat 240m’/day at 80-90% recovery, and is dosed with antiscalant to prevent scaling on the membrane surface. Computer software is used to set the RO plant configuration, antiscalant dase and to assess its performance. 33 Ancillary Processes ‘The ozone plant has been used prior to UF and GAC treatments and is capable of introducing up to 6mg/l of ozone into a flow of 1Sm’/h with 3 minutes contact time. ‘The Powdered Activated Carbon plant has to date been used prior to clarification, dosing up to 40mg/l of Norit SA Super PAC. Use of PAC prior to UF willbe included within the agreed programme of work. 4 RESULTS AND DISCUSSION The programme of work at the AWTP includes routine sampling and analysis to optimise performance of each treatment process relative to the feed water quality, regular spiking trials of key risk compounds to ascertain cost effective removal and special projects such as disinfection by-product control. Figure 2 shows the removal capability of GAC and RO treatments for high levels of key risk compounds actually found to date in River Trent feed water. Boron being such a small element is unusual in that i is only removed to the same extent by RO as conventional treatment, approximately 50%. This could be important if the current UK limit of 1000ug/l was reduced to WHO guide level of 300ug/1. To date up to {60% removal has been seen as a result of pH adjustment to 8.5 prior to RO and this will be investigated further together with the effects of reducing pH. Bromide was detected in River Trent feed water at concentrations between 150 and S40ug/l. It is removed by RO to its limit of detection but has a mean concentration of 280ug/l (maximum 450ug/l) post GAC. Bromide is not included as a regulated parameter ‘within drinking water standards but it plays a significant role in the formation of trihalomethanes and bromate resulting from disinfection with chlorine and ozone ‘treatment respectively. Nitrate levels in the Trent can at times exceed the 50 mg/l limit which can not be reduced using the conventional processes currently available at Church Wilne WTW. To108 ‘Membrane Technology In Water and Wastewater Treatment bromide soproturon | N00%= —Astug—S58pT REG tt a ZT oat Figure 2 River Trent AWTP Results - Occurrence of Key Risk Compounds reduce nitrate levels in the Trent to drinking water standards would require either for it to be blended 1:1 with River Derwent water (the current raw water source), or for it to be treated with another process such as Reverse Osmosis (RO). A combination of Ultrafiltration (UF) and RO (treating apart of the flow) can be considered as an alternative to blending and conventional treatment. UF offers a ‘total barrier’ method of disinfection when considering removal of cxyptosporidia, giardia, bacteria and viruses. Chlorine addition after UF is therefore only required to give a protective residual rather than full disinfection as is the case for the conventional stream. Micobiological data to date show the complete absence of bacterial ‘or virus penetration of UF. The reported 3 log removal for eryptosporidia and giardia (cocysts and cysts) by UF will be checked within the work programme. RO treatment achieves approximately 90% removal of Alkalinity and Total Hardness (raw feed waters being typically 300 and 150 mg/l as CaCO; respectively). This yields a slightly acidic, weakly buffered product requiring stabilisation by remineralisation or blending with UF or another treated supply to be suitable for public supply. For cost effective treatment to drinking water standards a. minimum percentage of the flow should be treated by RO and the need for remineraistion be avoided if possible. Severn Trent have established a very advanced on-line monitoring unit on the bankside of the River Trent, to establish the necessary river intake protection to the proposed water supply scheme, The on-line monitors includes a SAMOS-LC system using a conventional HPLC technique on-line for the analysis of organic pollutants. This monitor originally developed for the analysis of triazine and phenylurea herbicides in the River Rhine system is unique in this River Trent installation in that it includes a modification to allow the detection of several acid herbicides and phenolic compounds known to be frequentlyWater Quality and Trearment 10s present in the River Trent. Detection of three such acid herbicides in July 1998 in the River Trent at approximately 0.3 ug/l resulted in sampling of RO permeate and GAC product waters. The complete eliminetion of these acid herbicides to below analytical limits of detection by these treatment processes was observed. Figure 2 shows the removal capability of GAC and RO treatments for isoproturon and mecoprop when treating 0.1 and 0.27 ug/l concentration respectively. ‘A wide range of herbicides (triazines, urons and acid herbicides) have been spiked at 2 ug/l with GAC treatment of 15 minutes empty bed contact time (EBCT) and RO achieving < O.}ug/l in all cases. This matches the capability of the on-line pesticide monitoring facility which alarms when selected herbicides from each group exceed 1.0 ug/l. Ozone doses of 2 mg/l applied reduced the pesticide loading on to GAC or RO by between 25 10 ‘50 % across the range of herbicides added although ozone residuals after 3 minutes contact had to be avoided to prevent the formation of bromate. PAC (at 20 mg/l) has been seen to reduce 4 ug/l by 80% for most of the above range of pesticides. ‘The flame retardents TNBP, TCEP and TCPP have been regularly spiked at 2.0, 5.0 and 10 g/l respectively in line with the maximum concentrations typically found in Trent and Derwent derived feed waters. TNBP and TCEP are removed to <0.1 jg/l by GAC (15 ‘minutes EBCT) and RO. TCPP is reduced to <0.5 and 90% >) high frequency of acid washing of the membrane packet (once a day, during the ‘washing phase, since the plant had one lin only, no treated water was produced). The high percentage of water recovery, calculated as the ratio between water produced ‘with alow nitrate concentration and the untreated water being fed in, obviously lead to the production of a concentrated discharge with high percentage of ion NO3 ‘Therefore, during the second experimentation phase, substantial modifications were ‘made tothe plant conditions in order to cut consumption down: © polarity reversal cycles every 20 minutes ‘* acid washing using 2% HCI solution once week ‘© sulphamic acid proportioning with conductivity set on the electrolyte at 10500 uS/em.Woter Quality and Treatment ns “Poses SEEEEE ISS SOI LISI ESTE ALSIP MOVER SEITE Figure 6 kWh/me and % water recovery126 Membrane Technology in Water and Wastewater Treatment ‘The application of these conditions - especially the reduction of the number of acid ‘washings - led to a considerable reduction in the consumption of additives, while keeping the abatement throughput, 4.1 The quest for compatibility of the concentrated discharge. ‘The tests aimed at obtaining a discharge compatible with the limits set out earlier were run ‘working progressively on the reduction of the percentage of water recovery: practically speaking. It was decided to increase the quantity of water drawn in by-pass from the input and delivered to the brine tank to dilute it; this led to an increase of the quantity of discharged brine. This operation was performed in stages, lowering the yield from 94% step by step to 1%. Measurements of the concentration of NO3” ion for discharge showed the following general trend: ‘© with recovery around 94%, there is a concentrate discharge between 500 and 600 mg/l. with recovery around 90% there is a concentrate discharge between 300 and 330 mg. + _with recovery around 80%, there is a concentrate discharge between 200 and 250 mg/l. On the basis of incoming concentrations of untreated water of around 50 mg/l of NO3~ jon, one can make a rather accurate estimate on the basis of which one can say that in onder to reach the limit set previously. It will be necessary to achieve recovery values around 70%, ‘The important factor which emerged from the tests that were made is that - despite the concentration of recirculating brine having been heavily reduced (from around 7000 1S/em to around 2500 yS/em), the percentage of ion nitrate abatement on the water produced is always near the previous levels, .c. 70%. Therefore, the decision to preserve high percentages of water recovery was dictated by reasons of opportunity (especially in financial and management terms) and not by any technical limits ofthe EDR technology. 42 Assessing management costs On the basis ofthe technicel and management data gathered during the experimentation, ‘we can draw up a sufficiently accurate estimate of management costs on the basis of the size of the plant. The following data cannot be applied directly to plants with a larger ‘throughput, but itis quite probable that costs will diminish, since some will be spread out ‘over larger production (additive consumption, labour, and analyses) Total operating costs, on the basis of the experimental data, are lower than the 337 {tf/m3(ltalian Lire) of the data submitted by the Milan CAP which refer to a Reverse Osmosis treatment system, fora similar application’ Pilot Plan Throughput m3/year [About 205,000 m3 (average NOS” ion [concentration of less than 15 mg/l) [Elecite Power Rtn (estimate 250 10k Wh) | 100- ‘Running Consumption (additives) i/m3_—_[T [Routine Maintenance itf/m3 10s Personnel itfim3 __ Bs ‘Chemical Analysis 127m 36, =| TOTAL OPERATING COSTS Timi [280Water Quality and Treament vw 5 CONCLUSIONS ‘We can definitely say thatthe results of the experimentation made by A.M. together with the builders of the plant (FRAME S.p.A. of Ozzano Emilia, inthe province of Bologna, Italy) are positive. “The data gathered on about 5000 hours of operation of the plant allow us to draw several ‘important conclusions: © during the last 8 years (the previous experimentation was in 1990), EDR technology for the abatement of nitrates from underground water has met with interesting developments, especially in terms of the performance of the miembranes, and this leads us to expect further future developments; © EDR technology has tumed out to be very well focused as an application for treating the removal of nitrates in underground water, especially in terms of preservation ofthe ‘ste and mineral balance features of the original water; ‘+ EDR technology has absolutely competitive features in terms of management and investment costs when compared to high-tech processes such as Reverse Osmosis. ‘Compared to other technologies also applied in treatment of the removal of nitrates from underground water (lon Exchange and Biological Denitrification), EDR definitely involves higher investment costs, but still offers an interesting margin of competitiveness in terms of management costs, especially if we consider how much ‘more flexible itis than a biological system (difficulty in controlling the process and critical plant start and stop phases). + EDR technology, together with Reverse Osmosis (technologies with a higher degree of innovation) certainly present the best features in terms of flexibility and adaptability (at least for the specific application refered to here), especially when we consider that it was already possible during the planning stage to formulate tailor-made solutions, depending on the results which were desired also, once the plant has been built, it ean be easily adapted to new requirements without any special complications in terms of planning and/or management; ‘© Application of EDR technology for the treatment of underground drinking water presents no special problems in terms of adaptation, since this technology has already been used widely and for a long time in the food industry. Indeed, in the water distribution field conditions are far less critical than in other applications (concentration of the fluid being treated, temperature and pH conditions, et.) so the life of the main components. (membranes and electrodes) is. certainly’ longer. ‘Therefore, the amortisation costs of the plant ar les. References 1 P, Coté: INTERNATIONAL REPORT "State of the art techniques in reverse osmosis, nanofiltration and electrodialysis in drinking-water supply - Anjou Recherche, Centre de Recherche de la Compagnie Generale des Eaux, Chemin de la Digne, BP76, 78603 Maisons Lafitte Cedex, France. 2 American Water Works Association: Manual of water supply practices - Electrodialysis and Electrodialysis Reversal - M38 (1995). 3G. Gariboldi, S, Moriggi: "Esperienze applicative di tratamento dei nitrat -trattamenti con Osmosi Inversa" - XI Corso Residenziale sullacqua, CISPEL Lombardia, Milan, May 21%-22, 1996.NANOFILTRATION FOR DRINKING WATER TREATMENT FROM A EUTROPHIED LAKE IN TAIWAN ‘Houan-Hsien Yeh, Sheng-Hemg Lin, Shan-Jhen Kao, and Grace T. Wang Department of Environmental Engineering, National Cheng Kung University, ‘Tainan 70101 Taiwan 1 INTRODUCTION Cheng Ching Lake Water Works, located in southem Taiwan, is the main supplier of domestic water for the Greater Kaohsiung Area, the second largest metropolis in Taiwan ‘with a population of over two millions and the location of major heavy industries. The ‘water works drew its raw water from a nearby lake, a man-made off-line reservoir storing surface water pumped from the Kao-pin River. Owing to upstream discharge of farming, industrial, and domestic wastes, the water of the Kao-pin River is polluted. Therefore, the lake also becomes eutrophied. Although the treated water from the water works, which employs conventional treatment processes, including prechlorination, coagulation, sedimentation, filtration, and disinfection, can meet the current water quality regulation, the taste and odour causing compounds from algae can not be completely removed, and the hardness ofthe water is high. Therefore, the consumers have lots of complaints about the water quality. Recently, the use of membrane technology in water treatment has developed quite idly. Membrane processes have the advantage of occupying less landspace and are ‘flexible in expansion. Nanofiltration (NF), one application of the membrane technol is reported being able to reject higher than 90% of the divalent and trivalent ions’, However, its rejection for the monovalent ions is only 30~60%, NF was first successfully used for ground water softening in Florida, USA? Furthermore, it also has been ‘employed forthe removal of dissolved organics including the precursors of halogenated ‘organics. Is rejection for THMFP and AOXFP can be higher than 90%. In order to upgrade the water treatment processes, a pilot scale testing, including nanofiltration, has being conducted to trea the raw water from the Cheng Ching Lake for simultaneous organic and hardness removal. In addition to water quality monitoring, the efficacy of various pretreatment strategies used to prevent membrane fouling was also studiedWoter Quality and Treament 19 2. MATERIALS AND METHODS 2.1 NF Membrane Pilot System ‘The core of the NF membrane pilot system is a two-stage membrane unit. The unit was ‘configured with two pressure vessels; in the first stage feeding the concentrate to a single pressure vessel in the second stage. The permeate from each stage was collected and combined into a single product stream. Each pressure vessel is consisted of three clements in series, with a total of nine elements. The element used in this study was the spiral wound NF70 thin-film composite membrane (FilmTec, Dow Chemical). antiscalant [| al ee ghee | mim I intuent OVO TO. | weste AC tank sum cari fer A permeate stage 1 Figure 1 Schematic Diagram of the NF Membrane Pitot System Bach element was 10.2 em (4 inch) in diameter by 101.6 em (40 inch) long, with 7.62 m: (82 f0) of active membrane area. In order to meet the membrane specification of minimum flow, recycle stream was routed from the final concentrate stream back to the suction side ofthe booster pump blending the recycle with the membrane feed stream. ‘The NF membrane pilot system also includes an influent pump, a GAC tank, dosing pumps for acid or antscalant, a Sym cartridge filter, and a high-pressure feeding pump (Figure 1). 22 Pre-treatment Scheme Before the water was fed into the NF system, the raw water from the Cheng Change Lake ‘was subjected to various pretreatments. The purpose is to compare the efficacy of these pretreatments on membrane fouling control. The pre-treatment schemes tested in this study included: (1) coagulation, sedimentation, and granular media filtration (called conventional treatment process, CTP); (2) preozonation and CTP; (3) CTP and UF; (4) preozonation, CTP and UF. For the CTP, conventional paddle type mixing tanks were used for both rapid and slow ‘mixing. The coagulant used was liquid aluminium sulphate (7.5% AO), with a dosage between 50-100 mg/L. Tube settler was utilised for sedimentation with an overflow rate ‘varying from 18.8 to 69.8 m?/-day. Dual media filters, wth 35 cm depth of anthracite ES. 0.89 mm) laying over 25 cm of silica sand (ES. 0.51 mm), were used to further ‘reat the settled water. The filtration rate was between 100 to 200 m"/m*/day. The goal of ‘operating the CTP, either with or without preozonation, was to control the turbidity of the10 Membrane Technology in Water and Wastewater Trecoment effluent from the granular media filter to be lower than 0.3 NTU over 95% of the time, ‘The UF membrane unit (Microza UF system, Asahi Chemical Industry, Japen) was consisted of two hollow fibre membrane modules (LGV-5210), made of polyacryloniti (PAN), in parallel. The length and diameter of the module were 2227 mm and 140 mam, respectively, while the inner and outer diameters of the fibre were 0.8 and 1.4 mm, respectively. The effective surface area of each module was 41 m. The molecular weight cut-off (MWCO) ofthe membrane was 13,000. The filtration mode was outside to inside 23 The Evaluation of NF Membrane System Performance In order to control inorganic salt scaling, the pre-treated raw water was either subjected to pH adjustment to about 62 by adding sulphuric acid (or hydrochloric acid) or antiscalant application (Hypersperse™ AF200UL, Argo Scientific) before flowing through the 5m cartridge filter and finaly to the NF module. ‘The evaluation of the NF performance under various pretreatments is based on the decline of water mass transfer coefficient (MTCy) with time. MTCy is determined by ing the water flux of each membrane element by the net driving pressure: ‘Water flux of each membrane element (LMH) = Net driving pressure (kg/em*) NDP is defined as: glen’ 100mg ‘TDS, TDS¢ and TDS- are the total dissolved solid concentrations of inlet, concentrate and permeate, respectively. Once MTC, for each set of monitored pressure and water flux is determined, the MTC, for the entire operation period is plotted as a function of operation time (or ccurulative permeate volume). The slope’ of the graph can then be obtained using ‘Statstical regression techniques. The slope is called the fouling rate of the membrane, ‘When MTC, decreases to 85% of its original value, the membrane needs to be cleaned. ‘The interval between cleaning is called the cleaning frequency (CF), noted by days. Daring this testing, the influent, permeate and concentrate of the UF and NF systems ‘were collected periodically for water quality analysis. The parameters analysed included turbidity, alkalinity, hardness, total dissolved solids (TDS), conductivity, UV absorbance at 254 nm (A254) and nonpurgeable dissolved organic cazbon (NPDOC), TDS, )x0.0708 ¢Warer Quatityand Treatment 1 3 RESULTS AND DISCUSSION Sixteen NF runs were carried out during the period from January 1999 to September 1999, Table 1 lists the operation conditions of all these runs. The recovery forthe whole system was set at 75%, except Run 12, which was 60%. Figure 2 shows the plot of the MTC, values obtained from Stage 1, Stage 2, and the whole NF system, respectively, versus the cumulative permeate volume for all these runs. The fouling rate and CF for ‘each run are listed in Table 2 In the first 9 runs, no matter whether the pretreatment was CTP with or without preozonation, the CF of NF system was always less than 7 days, with relatively high silt density index (SD) of the influent (about 5.2). At that point, it was assumed that the influent to NF system contained a high level of colloid and/or particulates thus fouled the NF membrane quickly. AS a matte of fact, some of the runs were terminated due to clogging ofthe Sm cartridge filter by algae, causing rapid headloss accumulation. This also indicated that some algal cells passed through the granular media filter, despite the turbidity of the filtrate had been controlled to be less than 0.3 NTU over 95% of the time. ‘Based on the assumption stated above, the UF unit was introduced to the membrane system from Run 10 (March 29, 1999). The filtrate from the granular media filter was further treated by UF before it entered the NF unit. As shown in Table 3, UF proved to be very effective for colloid and particulates removal, as SDI decreased from 5.2 to 0.8, and turbidity from 0.21 to 0.03 NTU, There was about 17% removal of NPDOC, probably by adsorption on the fibre surface. However, removal on dissolved species, such as TDS, conductivity, and hardness, was not observed. The CF for Run 10 was stil very short (3 days only, as shown in Table 2). The over-all efficacy on NF membrane fouling control by applying UF as pre-treatment was minor. This suggested the colloid and particulates probably were not the primary factors causing membrane fouling. ‘Another possible factor, the soluble species, was then taken into consideration. To investigate the major composition of the foulant, the spent cleaning solutions were analysed. In the first four runs, the cleaning agent used was hydrogen peroxide solution, However, after the cleaning, the original MTC value (7.48 LMH/kg/em’) could not be reestablished. The cleaning efficiency of hydrogen peroxide solution was doubtful From Run 5 to 9, the NF membrane was frst washed by sulphuric acid solution (pH-2) and then followed by 45% sodium hydroxide solution wash. It can be found in Figure 2 thatthe MTC, value after the acid-base wash resumed its original value. In Run 10, only acid wash was employed for membrane cleaning and the restoration of MTCy, was just the same as that of acid-base wash used in Runs 5 to 9. This probably indicates that {inorganic scales mainly caused the fouling. From Run 1 to 10, the depression of pH level of the influent was achieved by adding H,S0,. However, the concem of adding H;SO, was that the fouling may be caused by the deposition of sparingly soluble sulphates salts, such as CSOs, $180s, or BaSOs. Therefore, since Run 11, the acid employed was switched from H2SO¢ to HCI. Nevertheless, the improvement in fouling control ‘was minor, since the CF only inereased to 5.7 days. Later, when the water quality analysis data ofthe feed and the acid-wash waste were available (Table 4), it was surprising to find a tremendous increase in aluminium concentration in the acid wash waste, as compared to that in the influent, This suggested that ‘Al might be the major cause for the membrane fouling-Menbrane Technology in Water and Wastewater Treatment im yg afpunes wrissanisdt3 “9 {¢92ejonwoo Hd wopnneo> + sayy spa unl § + 4+ LD S| yy s8punmo uml ¢ + a9 + woubberoany “Z soy a8pus unt ¢ + (Goeeo014 yuan |BUOUDAGED) ALD “T ae _— a. moe = 9 SL ESRF GNHTORTIN a) G6IFI-G6IBM OT moporg + veor ' sk ssa GoiLI~66ILIE $1 s st 99-8 1 s st s5-st fl s 09 ss-stt zi + St $69-S6% u € SL so1~so+ 66Irie-65Ii6z OL HON "08H 1 se teat GsIeicc~soresz HOON+ 1 SL 99-S6% 8 HOON +05 1 st 199s L HON +S 1 SL £195 9 HON +05 z sk ees $ 1 SL L69-595 ' z 99-8 € z SL 69595 z 1 SL_sci-aw's 1 Pr ody Suzy joauog 1d nf roy “sig ciowony Nad ang wy “#tUaIsS IN ays Jo SuojIpUO) WoHDIedD TAAL,Water Quality and Trearment 1 Figure2 Plots of MICw versus Cumulative Permeate Volume ‘MTCw (LMEVke/era) MTCw (LMH/kglert) (Came pees vue (a) In order to allow the least amount of Al into the membrane system, the operation of rapid ‘mixing tank of CTP was controlled at pH 6.8 by adding hydrochloric acid in Runs 13 and 14, since the solubility of Al salt is least at this pH level’. Although these 2 runs seemed to last longer than the previous ones, the improvement was limited. In the following run (Run 15), sulphuric acid was once again used to control the pH level of coagulation in CTP and the influent to the NF system. It was found that the performance of the NF system did not deteriorate. Therefore, at this stage, it could not be determined whether the fouling of NF was caused by sulphate precipitation,134 Membrane Technology in Water and Wastewater Treament Table2 Fouling Rate and Cleaning Frequency (CF) of NF Operation ‘Run Fouling Rate (LMEkgiem®) CF day) T Ft 2 45 3 69 4 3 5 26 6 10 7 19 8 22 9 13 0 30 " 37 R 50 B 37 4 34 15 98 16, 534 “MTC a7. 48 LMEgfom? Table3 Water Quality of UF System — ‘ed stream Tub NT Hardness (mg/L as CaCO3) 57 Alkalinity (mg/L as CaCO3). 54 NPDOC (mg/L) 89 TDS (mg) n Conductivity(usiem) 2 SDL 2 Before starting Run 16, the NF membrane was thoroughly cleaned by two proprietary formulated cleaners. The membrane was first washed with a low pH liquid formulation (Bioclean 103A, Argo Scientific) for one hour, then followed by a high pH liquid formulation (Bioclean 511, Atgo Scientific) for another hour. Based on the pressure required to maintain certain flux after each cleaning as well a the water quality analysis (eg, Al, Si, Ca, etc.) on the cleaning sotution waste, it can be noticed that acid cleaning ‘was more effective than alkaline cleaning. The performance of these cleaning products is superior to that of the H,SO, /NaOH solutions used in previous runs. In Run 16, a phosphonate and polymer based antiscalant (Hyperspersc™ AF200UL, Argo Scientific) was added continuously into the influent of the NF system at a dosage of 3 mg/L. The pH of coagulation, prior to the Sjim cartridge filter, by addition of acid was suspended. The result was very promising and the CF was extended to about 53 days. So far, the application of antiscalant provided the best NF performance, Table 5 lists the average water quality of the influent and permeate of the NF system. ‘More than 80% of total hardness, TDS and conductivity were removed. The rejection for ‘organic parameters, such as NPDOC and A254, is 75 and 72% respectively. ‘The ‘organoleptic quality ofthe NF permeate was also superior to that of the tap wate from ‘Cheng Ching Lake Water Works, according to blind tests by consumer groups. One of theWater Quay and Treament Bs advantages of the membrane process is its stable permeate quality from these 16 runs, the permeate quality was consistent over the operation period, ‘Table 4 Water Quality Analysis of the Feed and Acid-wash Waste ‘obtained from Ru 10 0 0. 1107 32.56 ‘Table Water Quality of NF System Feed stream — Permeaie Turbidiiy NTU) 0.06 0.03 Hiardness (mg/L as CaCOs) 233, 20 Alkalinity (mg/L as CaCOs) 121 2 A254 (an!) 163 045 NPDOC (mg/L) 0.675 017 ‘TDS (mg/L 47 Conduetivty(uSlem) 4 SUMMARY This pilot scale testing, including NF membrane unit has led to a successful upgrade of the treatment processes of a water works which drew its raw water from an eutrophied lake. The NF membrane process was found to produce high quality permeate, with over 75% of NPDOC rejection, 90% of hardness rejection, and excellent organoleptic quality. However, pre-treatment should be carefully chosen to prevent fouling. Several pre~ treatment schemes were studied, including conventional treatment process only, ‘conventional treatment process with preozonation, conventional treatment process with ‘UF, and antisealant application. Although conventional treatment process with or without preozonation could reduce turbidity to less than 0.3 NTU, the SDI was still high. When UP was added, the SDI reduced to 0.8, indicating that most colloidal and particulate foulants were removed. However, scaling caused by inorganic precipitation still can not be solved. So far, the application of antiscalant provided the highest efficacy on NF fouling control. 5 ACKNOWLEDGEMENTS ports provided by National Science Council, Taiwan, ROC (Project No. NSC #7-991-8-006-010), Taiwan Water Supply Corporation, and China Stel Company axe greatly appreciated, The assistance provided by Dow Chemical Company, Asahi Chemical Industry Co, and BeteDearbom Company during this stidy is also acknowledged136 Membrane Technology in Water and Wastewater Treatment References 1, AWWARF, LAE, and WRC, Water Treatment Membrane Processes, McGraw-Hill, New York, 1996. 2. WJ. Conlon and S.A. MeClellan, Jour: Am. Wat, Wks Asso. 1989, 81, 47. 3. TJ. Blau, JS. Taylor, K.E, Morris, and L.A. Mulford, Jour Am. Wat. Whs Asso. 1992, 84, 104. 4. §. Kawamura, Intergrated Design of Water Treatment Facilities, John Wiley & Sons, New York, 1991.Fouling and CleaningMEMBRANES AND MICROORGANISMS - LOVE AT FIRST SIGHT AND THE CONSEQUENCES H.-C. Flemming Department of Aquatic Microbiology University of Duisburg Geibelstrase 41 D-47057 Duisburg, Germany 1 INTRODUCTION The treatment of water by membrane technology intrinsically implies the contact of very large quantities of water withthe membrane surfaces, This water isnot sterile. In drinking ‘ater, the numbers of cells actually present as demonstrated by microscopic quantification usualy range between 10" and 10° cells mL’. These cells have a tendency to adhere to surfaces; in oligotrophic systems, this is considered e survival strategy". There is virtually no surface material which cannot be colonised, even under extreme conditions, regardless of hydrophobicity or hydrophilicity, smoothness or chemical composition? - surface conditions and materials will simply select for colonising species among the spectrum of organisms in a given water volume. In a membrane system, adhesion to the membrane surface is facilitated by the vertical transport vector which is given by the water flow through the membrane - this can be described metaphorically as "love at first sight", because there will always be some organisms which prefer to settle on the given membrane material, be it hydrophobic or hydrophilic. Once the organisms colonise the surface, they will inevitably multiply and form biofilms. All membrane systems, which are not operated under absolutely sterile conditions, will cary biofilms”. ‘Not all of the systems carrying biofilms suffer from biofouling - “biofouling” is an operational term, applied when the effects of biofiims exceed a certain threshold, or tolerance level, which is individually set for different systems* In membrane systems, however, biofouling is the “Achilles heel” of the process, because al other fouling components, such as organic and inorganic dissolved substances and particles can mostly be removed by efficient pre-treatment; however, microorganisms are particles which can multiply, Thus, if they are removed to 99.99 %, there are still ‘enough cells left which will grow at the expense of biodegradable substances in the water. Microorganisms arc ubiquitous in any technical system unless it is Kept sterile by enormous and continuous effort. The types of microorgenisms and the nutrient concentration represent the biofouling potential40 ‘Membrane Technology in Water and Wastewater Treatment Biofouling leads to considerable technical problems and economic loss’. Not only the feed side of the membranes is concemed but microorganisms can also pass the membrane’, although they are believed to be too large to penetrate @ reverse osmosis ‘membrane, As membrane technology will be of increasing importance in meeting the rising demand for treated water, biofouling will be an increasing problem. Biofouting is a biofilm problem; itis important to understand basic biofilm processes and properties in ‘order to design rational countermeasures 2. ADHESION AND COHESION Within minutes of contact with non-sterile water, the first microorganisms will adhere’ Primary colonisation is strongly influenced by the concentration of cells in the water phase. Among the spectrum of various microbial species, a clear preference for given ‘membrane materials can be observed; one example is the preferential colonisation of polysulfone by Pseudomonas diminuta and Staphylococcus warneri as investigated by Flemming and Schaule". In order to investigate such preferences, samples of the mixed population of a mature biofouling layer were selectively removed from the bottom (cell- ‘membrane interface) and the top of the layer. Various membrane materials were exposed to suspensions of the two populations and the adhesion kinetics were measured. While the bottom population showed a clear preference forthe material from which it was isolated, the top population did not. This makes sense as the population at the water-biofilm interface is separated from the membrane material by many layers of microorganisms and extracellular polymeric substances (EPS). Thus, manipulation of the membrane material in order to reduce microbial adhesion most probably will select for a species that adheres to that material. With time, this species will cover the surface and mask its effect. Similar observations have been made with ant-fouling costings on ship hulls, in particular, with copper plating It was shown that dead cells of P. diminuta adhere at the same rate to the surface as living cell’. Ths indicates that these cells already carry the "glue in suspended form, and ‘the material which mediates adhesion as well as cohesion is the EPS. These are composed of polysaccharides, proteins, glycoproteins, lipoproteins and other macromolecules of ‘microbial origin”, They form a slime matrix, which sticks the cells to the surface and ‘keeps the biofilm together. Any cleaning measure has to overcome the overall binding ‘energy of this system. This energy is not provided by covalent chemical bonds, but by ‘weak physio-chemical interactions. In general, they can be divided into electrostatic interactions, hydrogen bonds and van der Waals interactions (fig. 1). The average binding energy ranges between 0.1-10 % of that of a covalent C-C bond, depending on the respective conformations of the macromolecules, the water content, pH value, ionic strength, temperature and other parameters. The weak binding energy of the individual ‘bonds is increased by the fact that the EPS molecules possess many functional groups capable of interaction. Ifa macromolecule has 10° possible binding sites and only 10 % of ‘them are interacting; the binding forces ofthe weak interactions are multiplied by a factor of 10, resulting in a considerable stability. The mechanism by which cleaners disintegrate biofilms is based on interference with these interactions. They do contribute, but not always in the same proportion, to the overall binding forces - they vary, according toFouling and Cleaning 1 surface properties and EPS composition. Cleaners have to overcome the cohesion forces ‘and have to address all forms of weak physio-chemical interactions. Fig. 1: Primary adhesion of P. diminuta to polysulfone membrane material ‘Surfactants are a major constituent in many cleaning formulations. They will interfere with van der Waals interactions and influence the so-called hydrophobic interactions. Van der Waals interactions can be dominant in systems in which cells adhere from water to hydrophobic surfaces, Schaule!” has shown that the adhesion of P. diminuta to polysulfone membranes is performed with significant participation of van det Waals interactions. This could be demonstrated by the influence of surfactants on the adhesion rate: however, in preliminary experiments, the cohesion of alginate, an extracellular polysaccharide of Pseudomonas aeruginosa, was not affected by surfactants, but was strongly influenced by electrostatic interactions’. These examples demonstrate that in ferent systems, different binding forces can dominate Phosphate, citric acid, salts, other ionic compounds and complex-formers will interfere with weak electrostatic interactions, Many of these substances are components of cleaners. thas been shown that electrostatic interactions, which are important in cohesion of EPS molecules!2, do not prevail in adhesion of P. dimimuia to hydrophobic polysulfone ‘membrane surfaces" So-called chaotropic agents, such as urea, tetramethyl urea, guanidine hydrochloride, ‘and others which are known from protein chemistry interfere with hydrogen bonds. They literally cause a chaos in water structure by rapidly binding water molecules, which are ripped from hydration, water surrounding proteins or polysaccharides. Hydrogen bonds represented a dominant kind of force in both adhesion and cohesion systems as described above; however, chaotropic agents are usually not constituents of cleaners. In cohesionma Membrane Technology in Water and Wastewater Treaoment and adhesion experiments using alginate as an EPS model, guanidine hydrochloride showed significant effects!" ‘The entanglement of the macromolecules provides an additional factor in biofilm stability. Ths i addressed by the use of enzymes, The problem, however, is thatthe EPS macromolecules are of highly variable composition and structure, and enzymes mostly are too specific 0 act on the entire variety. This is one reason why enzyme treatment frequently yields disappointing results in practice. In membrane systems, it has been shown that cleaners can improve the permeability of the fouling layer without reducing the fouling layer. In an experiment, an agar gel layer ‘was taken as a model for a hydrogel!” Even though the layer thickness is not changed by the application of the cleaner, an almost five-fold increase in the layer permeability was seen, Thus, cleaners can improve (and decrease) the permeability of the biofilm, although this is a transient effect. The results suggest that optimisation of the fouling layer is possible even if it is not removed. This, however, is not true for every given cleaner. Correct selection and tailoring of conditioning agents has been shown to significantly alter the hydraulic resistance of biofilms. Increasing the permeability is desirable (ic. decreasing the resistance or specific resistance). Table 1 shows some examples of changes in fouling layer permeability demonstrated with model fouling layers of filter cakes from bacteria and activated sludge. Inthe case of formaldehyde, a commonly used disinfectant, the effects ofits application is to reduce the permeability of microbial layers substantially, that is not very surprising, as formaldehyde is used as a fixation agent in microscopy. This is an experimental example with artificial model biofilms, which helps understand why the performance can deteriorate aftr the use of cleaners. Table 1. Alterations in specific hydraulic resistance, 1 and permeability, Ly of microbial layers due to chemical conditioning (after {13))___ of bacteria Filter cake OF bacteria Activated sludge 3 DETECTION In fouling cases, it is important to distinguish different foulants in order to design effective countermeasures. Usually, the membrane system will respond to biofouling problems by an increase of APfeed/brine AMV/OF APmembrane: however, this response is very non-specific and can be caused by other fouling mechanisms as well. The diagnosis “biofouling” will be given if countermeasures, which are usually taken against non- biological foulants, fail. Theo, a microbiologist is called and will most probably takeFouling and Cleaning 143 ‘water samples, count colony forming units or cell numbers, and maybe will identify a few of the microorganisms found in the water. In most cases, these data cannot be related to ‘the location of biofilm growth, as the cells in the water can be released from any site in the system, including the membranes, As cells do not detach from biofilms at a continuous rate but randomly, water phase cell numbers cannot be related to the extent of biofilm srowth either. ‘Thus, sampling on surfaces is mandatory. A good preliminary sampling protocol for ultrapure weter systems is given by Sematech'. Unfortunately, there is still no technology available to take biofilm samples non-destrucively from an operating membrane Interesting approaches are under development, such as fibre optical sensors that can be integrated in a membrane module and report the development of deposits on surfaces by an increase of reflected light", but this sensor still has to be adapted to common ‘membrane module designs. In practice, cither a bypass membrane device is used from which membranes can be removed and investigated destructively, or other accessible and representative surfaces, such as cartridge filters are sampled. In order to evaluate biofouling problems fatal to a ‘module, an autopsy is performed 3.1.1, Standard procedure for module autopsy A standard procedure for module autopsy as caried out in our laboratory isthe following 1. The module arrives in the laboratory (preferably cooled, not preserved, not dried, and immediately after being removed from the plant), is opened mechanically and unfolded 2. Optical inspection gives first information about the colour, the thickness and the consistency of the fouling layer. The colour can is the participation of humic. substances and iron compounds in the material; it also can be caused by microbial pigments", which can be misinterpreted as inorganic foulants. Scratching on the surface ‘gives an indication about the consistency of the layer. 3. A part of the membrane is inspected microscopically. This gives an indication about the prevalence of microorganisms and ebout the structure of the fouling film. In special cases, a cryosectioning procedure is carried out in which the membrane and the adjacent fouling layer are embedded in a water-soluble resin, which can be frozen and cut into slices of 2-50 jim. This procedure allows the exact determination of the thickness of the fouling layer" 4. A defined surface area is scraped off thoroughly. Water content and incineration loss are determined. A high water content and incineration oss are indicative for biofilms. 5. A part of the material is suspended in a defined volume of sterile water. The following parameters are determined in relation to the surface area: = Cell number (by epifluorescence microscopy; it yields the maximum number of cells present in the sample). = Number of actively respiring cells can be determined (by using a redox dye such as S-cyano-2,3-dtolyltetrazolium chloride, CTC", which is used as an electron acceptor by the microorganisms) = Number of colony forming units (on various nutrient media, which gives information about the minimum number of living cells, With regard to the particular fouling case history, the number of anaerobic or autotrophic bacteria, or ‘other specific organisms, can be determined and related to the surface area, = Content of proteins (determined after Lowry et al.'*)a Membrane Technology in Water and Wastewater Treatment = Content of polysaccharides (by a modification of the method of Dubois et al.") = Content of ironic acids (indicative for acidic polysaccharides; determined after ‘Blumenkrantz and Asboe-Hansen®) = Humic substances (after Feotund et") 3.1.2 FTIRespectroscopic analysis: The use of FTIR spectroscopy in the analysis of ‘microbial aggregates has been described in detail by Nichols et al* and Schmitt and Flemming”, The attenuated total reflection (ATR) mode offers a further possibility to investigate smooth surfaces of various materials directly. This can be achieved practically without sample preparation. With respect to biofilm research, it offers the significant advantage thatthe sample can be investigated relatively undisturbed. Thus, a biofilm does not have to be removed from its support, which would alter its structure considerably from its native form. The ATR method allows one to measure in aqueous media and to investigate the development ofa biofilm in situ, directly atthe interface, non-destructively and in realtime” 4 COUNTERMEASURES ‘The far most widespread approach against acute biofouling problems is the application of biocides. This originates from a medical kind of thinking: if microorganisms cause the problem, itis wise to kill them; however, this approach has some severe drawbacks in practice because it is applied 10 2 technical system and not to a living organism, ‘Therefore, a biocide may kill the biofilm organisms, although it is known that biofilm organisms display a significantly higher resistance to biocides", but it usually will not remove the biofouling layer. As the physical properties of this layer, ie. the hydrodynamic resistance, is the problem, it does not help very much to kill the organisms and leave the biomass where it is (fig. 2). Cleaning should be understood as a two step process: (i) Weakening of the fouling layer (performed by cleaners which interfere with the dominating physio-chemical interactions), and (ii) Removal of the fouling layer, usually performed by shear forces. It is very important to assess the efficacy of the cleaning process, as it may be much less effective than expected. It is evident that the ‘remnants of the fouling layer will invite new biofouling, 5 PREVENTION The conventional anti-fouling strategy is to dose continuously with biocides™, Various ‘membrane producers and vendors of water treatment provide dosing programmes systems; however, they do not always fulfil expectations, which is the reason for continuous research on additional and alternative strategies against membrane biofouling. Continuous biocide application generates. wastewater problems. The cost of the treatment of this ‘wastewater can exceed the savings gained by using the membrane technology. Tehas been demonstrated that practically every membrane system operating with water carries biofilms, unless it is operated under absolutely sterile conditions’, As a consequence, the biofilm participates inthe separation process, although in an undefinedFouling and Cleaning us way, however, not every membrane system has biofouling problems. This indicates that biofouling is an operationally defined phenomenon - it only occurs if biofilm development ‘exceeds a certain threshold of interference which may differ from one system to the other and is individually defined. Thus, a suitable anti-fouling strategy could be to tolerate biofilms below the threshold and curb the excess of biofilm growth above the threshold. Fig. 2: Scanning eleciron micrograph biomass left on fouled membrane surface after extended disinfection As indicated earlier, the biofouling potential comprises the ubiquitous microorganisms and the availabilty of nutrients. These must be considered as potential biomass. In technical systems, most biofilms will have reached the plateau phase of biofilm thickness ‘The extent of biofilm accumulation in this phase will depend on different factors that contol he eqlibrim level of the plateau ‘Nutrient concentration, type and availabilty + Shear forces + Mechanical stability ofthe biofilm matrix, as influenced by = oxidising agents (biocides, ete.) = temperature = type of microorganisms = physiological activity (gas production) = structure and physical strength of the EPS network ~ grazing organisms146 ‘Membrane Technology in Water and Wastewater Treatment ‘The most important factors are nutrient concentrations and shear forces. A factor of minor importance seems to be the cell density in the water phase. In the plateau phase, the further adhesion of cells to the biofilm does not contribute significantly to biofilm accumulation”; however, these cells are targeted by an effective biocide treatment, while the nutrient concentration isnot decreased. In some instances, the nutrient concentration may even be increased, when the biocide reacts with recalcitrant organic molecules, increasing their bioavailability” Biofouling can be regarded as a biofilm reactor in the wrong place, as exactly the same natural processes occur in both situations. Thus, if a surface-tich area is offered ahead of & system, which should be protected against biofouling, biofilms will form there and consume degradable matter from the water stream. This will decrease the extent of biofilm development in subsequent compartments. Bott investigated biofouling in heat ‘exchangers and has elesrly shown that decreasing the nutrient concentration could reduce the extent of biofilm accumulation. Biological filter systems such as those used in drinking water treatment might provide useful tools in reducing the nutrient concentration and thus, in prevention of biofouling. To date, biocides are worldwide the only answer to biofouling. Facing increasing difficulties in the application of biocides in both ‘effectiveness and environmental regulations, membrane technology might be well advised to develop biocide-free anti-fouling strategies. The optimisation of nutrient limitation techniques could provide new possibilities. The role of assimilable organic carbon (AOC) will be very important inthis approach. This approach would utilize biofilms in the right place, where they can be handled easily, in order to minimise the extent of biofilm formation at sites where biofilms need 10 be limited, Clearly, this concept cannot be applied to any given separation plant; however, if i is applied where appropriate, it might save large amounts of biocides. In addition, the concept needs reliable biofilm monitoring devices, Le. representing biofilm accumulation on membrane surfaces. Sacrificial module elements, as already proposed” will be highly useful. Considerable reseasch efforts will ‘be required to put this concept into reality, but it might help to solve fouling problems in ‘general. A practical example has been carried out in a water treatment system for a heat exchanger’. River water treated by flocculation and sedimentation was used as cooling water. A sand filter was integrated as a biofilter. Before and after the filter there was a reverse osmosis test cell installed, one operating with water before, the other with water after the biofilter (parameters given in table 2). Uronic acids were measured as indicative for EPS, The protein-uronic acid ratio before the sand filter was 7.7:1 (wiv) while it was after the sand filter 3.6:2.3, This indicates thatthe biofilm contains much more EPS under conditions of nutrient depletion. The BDOC is the biodegradable fraction of the dissolved organic carbon (DOC) in water. Interestingly, the thickness ofthe biofilm was reduced to tenth of the original size, and the structure of the biofilm was different; it seemed that ‘the biofilm formed under nutrient depletion was much less dense than the biofilm that formed in raw water, It was acceptable to reach a BDOC of 0.125 mg Lin order to stay below the threshold of interference. Such a value can be reached easily with common biofilters, The data on microbial concentration in the water phase show that the bifilter does not cause additional microbial contamination but rather removes cells from the process stream,Foulig and Cleaning ur Table 2: Biofilm, water and performance parameters before and afier sandfilter {after (6). ee = i @az3nmm——Gosaaem Protein content 3.6 wg cm* ‘Carbohydrate content 2.6 wg cm? ‘Uronic acid content 23 wg em? we ee . one Microbial content 1x 10" cfumL” 1.2.x 10° cfu mL Wetton eum iene strategy should reflect the properties and dynamics of biofilms. Fundamental aspects include sampling on surfaces, removal of the biomass, regarding nutrients as part ofthe fouling potential, and the integration of monitoring techniques for carly recognition of fouling problems and for optimisation of countermeasures'* ", Thus, biofouling cannot be overcome by application of some wonder-chemical but by a better understanding of the entire process and by interfering in the most effective way and on a scientific basis. This can include the "love at first sight” indicated inthe title of this paper, because it is possible to live with biofilms as long as their effects remain below the ‘threshold of interference. The combination of nutrient removal and monitoring represents a sustainable strategy for the future, References 1. Marshall, K.C. (1996): Adhesion as a strategy for access to nutrients. In: Fletcher, M.M. (cd.): Bacterial adhesion, John Wiley, New York, 59-87 2. Flemming, H.-C. (1991): Biofilms as a particular form of microbial life, In; H.-C. Flemming and G.G. Geesey (eds.): Biofouling and Biocorrosion in Industrial Water Systems. Springer, Heidelberg; 3-9 3. Flemming, H.-C., Griebe, T. and Schaule, G. (1996): Anti-fouting strategies in ‘technical systems -a short review. Water Sci, Technol., 517-524 4, Ridgway, HLF. (1988): Microbial adhesion and biofouling of reverse osmosis mem branes. In: Parekh, B.S. (ed.): Reverse osmosis technology. Marcel Dekker, New York, Basel; 429-481 5. Flemming, H.-C., G. Schaule and R. McDonogh (1993) How do performance parameters respond to initial biofouling on separation membranes? Vom Wasser 80, 177-186 6. Griebe, T. and Flemming, H.-C. (1998): ree antifouling strategy to protect. RO membranes from biofouling. Desalination 118, 153-156 7. Characklis, W.G. (1990): Microbial fouling control, In: W.G. Characklis and K.C. ‘Marshall (eds ): Biofilms. John Wiley, 585-6338 ‘Membrane Technology in Water and Wastewater Teaiment 8. Ghayeni, S.B.S., Beatson, P.J, Fane, AJ. and Schneider, RP. (1998): Bacterial passage through microfiltration membranes in wastewater applications. J. Membr. Sci. 3972, 1-12 9, Flemming, H.-C. and G. Schaule (1988): Biofouling on membranes - a microbiologi- cal approach. Desalination 70, 95-119 10. Wingender, J, Neu, T. and Flemming, H.C. (1999): What are bacterial extracellular polymeric substances? In: Wingender, J, New, T. and Flemming, H.-C. (eds): Bacterial extracellular polymeric substances. Springer, Heidelberg, Berlin, 1-19 11Schaule, G. (1992): Primaradhésion von’ Pseudomonas diminuta an Polysulfon- Mombranen. Dissertation Univ. Tubingen 12. Mayer, C., Moritz, R, Kirschner, C., Borchard, W., Maibaum, R., Wingender, J., and Flemming, H.-C. (1999) The role of intermolecular interactions: studies on model systems for bacterial biofilms, Int, J. Biol. Macromol, 26, pp. 3-16 13. MeDonogh, R, G, Schaule and H.-C. Flemming (1994); The permeability of biofouling layers on membranes. J. Membr. Sci. 87, 199-217 14,Sematech (1992): Sematech provisional test method for determining the surface associated biofilms of UPW distribution systems. Technology Transfer Number 920109S8B-STD 15. Flemming, H-C., Tamachkiarowa, A., Klahre, J. and Schmitt, J. (1998): Monitoring, of fouling and biofouling in technical systems. Wat. Sci. Technol. 38, 291-298 16, Wasel-Nielen, J. and Nix, N (1990): Kesselspeisewasser - Erzeugung aus Fluwasser durch Fonenaustausch und Umkehrosmose, Erste Betriebserfahrungen. Vom Wasser 75, 127-141 17.Schaule, G., H.-C. Flemming and HLF. Ridgway (1993): The use of CTC (5-cyano- 23- ditolyl tetrazolium chloride) in the quantification of respiratory active bacteria in bio ilms. Appl. Environ, Microb. 59, 3850-3857 18,.Lovry, O.1, Rosebronkh, NJ, Farr, AL. and Randall, RJ. (1951): Protein ‘measurement with folin phenol reagent. J. Biol. Chem. 193, 265-275 19. Dubois, M., etal, (1956): Colorimetric method for determination of sugars and related substances. Anal. Chem, 28, 350-356 20.Blumenkrantz, N. and Asboe-Hansen, G. (1973): New method for quantitative determination of uronie acids. Analyt. Biochem. 54, 484-489 21. Frolund, B., Griebe, T. and Nielsen, P-H. (1995): Enzymatic activity in the activated sludge floc matrix, Appl. Microbiol, Biotechnol. 43, 755-761 22.Nichols P., Henson M, Guckert J, Nivens J, and White D.C. (1985); Fourier ‘ransform-infrared spectroscopic methods for microbial ecology: analysis of bacteria, bbacteria- polymer mixtures and biofilms. J. Micobial Methods 4, 79.94 23. Schmit, J. and Flemming, H.-C. (1998): FTIR-spectroseopy in microbial and material analysis Int. Biodet. Biodegr. 41, 1-11 24.LeChevallier, M.W., Cawthon, C.D. and R.G. Lee (1988): Inactivation of biofilm bac teria. Appl. Environ, Microbiol, $4, 2492-2499Fouling and Cleaning 9 25.Ridgway, HLF. (1988): Microbial adhesion and biofouling of reverse osmosis membra nes, In: Parekh, B.S. (ed.): Reverse osmosis technology. Marcel Dekker, New York, Basel; 429-481 26.Bott, T:R. (1990): Bio-fouling. In: Bohnet, M. (ed.): Fouling of heat exchanger surfa «es. Conf. Proc., VDI Ges, P.O.Box 1139, 4000 Dusseldorf I; 5.1-5.20 27. LeChevallier, M.W. (1991): Biocides and the current status of biofouling control in water systems, In: H.-C. Flemming and G.G, Geesey (eds.): Biofouling and Bio ‘corrosion in Industrial Water Systems, 113-132; Springer, Heidelberg 28, Flemming, H-C., Schaule, G., Griebe, ., Schmitt, J. and Tamachkiarowa, A. (1997): Biofouling - the Achilles heel of membrane processes, Desalination 113, 215-225OPTIMISING MEMBRANE PERFORMANCE = PRACTICAL EXPERIENCES LLY. Dudley, F. del Vigo Pisano, M. Fazel PermaCare International, Aquazur Limited, Windsor, Berkshire SL4 3D, UK \ INTRODUCTION Europe is an expanding market for reverse osmosis (RO) technology used extensively for treating brackish water, seawater and wastewater sources. Membranes are now seen as a viable option for future potable supply in some areas of the UK. Regulators within the European Community are imposing stricter water quality standards, To comply with these tougher regulations and discharge constraints, the use of membranes will significantly increase in this new Millennium, ‘The financial implications of operating reverse osmosis membrane systems below ‘optimum performance levels can be considerable. Many large municipal plant have been constructed on @ ‘build own and operate’ (BOO) basis and the water costs to the end user, normally a municipality or large industrial user is crucial, This cost is usually the primary decisive factor for selecting the water production process, In recent years membrane desalination has become a more viable option in many countries. Recently finished costs for BOO potable water from seawater have ranged from as low as $0.50 to $1.00 per cubic metre, Table 1 gives an example of the contribution to total operating costs for a municipal seawater system, ‘Table 1: Example of Seawater RO System Costs Percentage of Total Cost Capital Recovery _ 186% Scale Inhibition 3.8% ‘Membrane Replacement 38% { Cleaning Chemicals 1% — Tow Pressure Pumps 48% High Pressure Pumps 12.1% ——Fas2%Fouling and Cleaning 151 One the most significant contributors to the supply cost is expenditure on pumping, due to the requirement for high feedwater pressures. This however, can be minimised by the use of energy recovery systems and suitable pre-treatment to maximise the water production and minimise membrane fouling. Poor performance results in higher feedwater consumption and the need for frequent cleaning. This means increased energy and consumable costs; therefore pressure is always on the OEM to design systems to operate at optimum efficiency. There are computer packages available that can evaluate and compare the major factors contributing to total RO operating cost. These programs ‘compare chemical dosing and energy costs and assess their ability to increase overall plant efficiency. Figure 1: Process Optimisation Program fre- opine frovtact iescnae [merece | | ssc re al heey feeeel more ‘The major factors, which affect membrane performance, are condition of the raw water supply effectiveness of pre-treatment procedures system operating parameters degree of plant maintenance and continuous monitoring responsiveness of plant operators to significant performance changes the rate and degree of fouling.12 Membrane Technology in Water and Wastewater Treeiment ‘Membrane fouling is the most common reason for performance problems, the effects are often: - reduced membrane productivity, poor salt rejection characteristics and increasing pressure differential across the membranes. 2. MONITORING OF PRE-TREATMENT PLANT One of the most crucial aspects for ensuring trouble-free operation is the regular ‘monitoring of feed water condition throughout the pre-treatment system and inspection of associated equipment and pipe-work. Analysing samples of feed water at each treatment stage enables @ good assessment to be made of existing plant conditions. Careful inspection of pipe-work, dosing tanks and cartridge filters can identify the presence of biofilm slime requiring sanitisation procedures to prevent continued ‘widespread microbiological growth. If a poor scale inhibition programme isin use or the ppro-reatment has failed, visual inspection may reveal colloidal particles, iron or inorganic scale on the micron cartridge filters. These are always positioned immediately prior to membrane filtration. Closer plant monitoring could prevent many system failures. Regular monitoring should be routine with data taken and plotted either electronically or by hand, Graphs should be prepared, dated and filed appropriately with any unusual or erratic results highlighted with explanations and comments. Important parameters are - pressure drop, output, pH, temperature, conductivity, and TOC. Computerised data lozging, packages incorporate alarm systems to alert the operator when parameters have exceeded fixed Timits ‘Table 2: Pre-Treatment Failures TREASO? FOR] INDICATORS MONITORING FAILURE Poor scale Tncreased delta P due to | Check antisealant dosing scale formation on | equipment and monitor membranes, usually at the| changes in water quality, softener failure Poor sanitisation Monitor sand filters, GAC, planktonic micro counts in feedwater. Check biocide dosing levels and inspect dosing tanks for biogrowth Look for signs of pipe-work corrosion, ferric breakthrough from media beds or failure of media High iron conteat filters igh Organic Content | Organic _matter__on | Feed water composition, cartridge filter and | review floceul membrane inlet procedures, monitor feed water colour/humies & TOCFouling and Cleaning 153, REASON FOR | INDICATORS MONITORING FAILURE Collide Breaiirogh | Colloidal panics Fouling | Measure SDI& ibidiy, the micron filers and | check condition of earrdge membranes filters, eliminate media filter fines Fouling by GAC fines | Carbon fines foul micron] Check washing procedure to filters and membrane remove fines ‘Overdosed loceulant | Severe loss of flux, | Check dosing levels and cationic floceulant can | detect excess traces prior to ireversibly foul | membranes ‘membrane surface. Presence of residual | Membrane damage, high | Use Redox meters, check chlorine permeate conductivity and | bisulphite dosing levels and ‘sudden increase in flux | positioning of injection point, chlorine test kit ‘Sand fier breakthrough | Sand and colloidal fouling | Check wash procedures to of cartridge filters and | remove fines membranes Failed aid dosing] Rapid scale formation, fast | Check aad desen, ead iT] increase in delta P & increasing delta P. Inspect cartidge fillers for CaCO, CiimatcSeazonat High _aerobioogial | Panktonic microbiological Change loading, biofilm slimes, on | counts, look for evidence of cartridge fies bots on filters & atthe ‘membrane inlet 3. CHEMICAL PRE-TREATMENT TO INHIBIT SCALES Feed water quality determines allowable plant operating conditions, which will dictate the optimum system recovery and overall production rate. It is necessary to make a full and detailed water analysis to identify all major anions and cations, which contribute to scale formation and general fouling. The major scaling/fouling ions are calcium, ‘magnesium, bicarbonate, sulphate, silica, iron and barium. ‘Many natural waters will deposit calcium carbonate on the membrane surface if untreated. Calcium carbonate scaling potential is determined by the Langelier Saturation Index (LSD), or the Stiff & Davis Saturation Index (S&DSI) for high ionic strength ‘waters. The risk of other scalants, such as calcium sulphate and silica are determined by ‘measuring their ionic concentration against their known solubility products (Ksp values). Scaling results in increased pressure drop and the need for greater feed pressure 10 ‘maintain constant product water output, Dosing of chemical antiscalant reduces the risk of scaling and allows elimination of acid dosing, while maintaining plant efficiency and optimum conversion rates. The use of an effective antiscalant will allow plant recovery to be increased to a brine LSI of up to +2.6 compared to a limitation of LSI +1.0 when using a commodity158 Membrane Technology in Water and Wastewater Treatment antiscalant such as SHIMP or zero in an untreated system. Computer p are available from some chemical suppliers to calculate scaling potential of a range of ‘water sources. The software allows accurate recommendations for antiscalant addition The effect of antiscalant on scaling potential is illustrated below for a brine of LSI +42.01.Te graph for the treated brine conditions indicates thet the plant recovery of 70% could be increased further if plant design limitations allow. At 70% recovery 3.4ppm ‘Antiscalant A is required to be dosed to the feedwater. No Treatment ‘Treated with Phosphonate Antiscalant A Figure 2: Scaling Potential of RO brine with and without Antiscalant A In the example (Figure 2), dosing of the phosphonate based antiscalant allows safe ‘operation of the RO plant with brine concentration of scaling species considerably in excess of normal solubility limits. 4 MINIMISING FOULING 4.1 Biofouling control ‘Most RO systems suffer biofouling to some degree, although this does not always severely affect performance as some biofilms remain within tolerable levels or have a high level of porosity, therefore not severely affecting the permeate flux. Biofouling potential should always be anticipated and measures taken to prevent and control biogrowth. This may require maintenance cleaning with @ non-oxidising biocide or for non-potable applications intermittent biocide ‘shock dosing’ on-line. Laboratory analysis can be used to characterise the fouling and propose the appropriate methods of control Biofilm material can be scraped from fouled membrane samples for microbiological analysis consisting of basic identifications and enumeration of bacteria, fungi and yeast’s. Most membrane biofilms contain both bacterial and fungal species. The physical structure of biofilms found in membrane systems can be ‘gel like or ‘slimy and adhesive’ with some consisting of a large ratio of polysaccharide slime to viable miero- ‘organisms. Membrane biofilms investigated in our laboratory often contain between 10° and 10* colony forming units (cfu) of bacteria per cm* of fouled membrane.Fouling and Cleaning 155 Biocide Sensitivity Tests (BST’s) have been used to evaluate the performance of selected ides on sessile micro-organisms isolated from membrane foulant and determine 1um conditions for use. A quantitative suspension testis used to determine biocidal efficacy against bacteria and fungal species. Samples of the sessile organisms are obtained by swabbing the foulant from the membrane surface and spot or pour plate ‘counts used 10 determine the efficacy of each biocide, The biocidal performance is expressed as percent kill for a known concentration and contact time, Non-oxidising biocide formulations are preferable due to the limited tolerance of polyamide to ‘oxidising products such are chlorine or peracetic acid ‘The following table outlines the multi-purpose use of a non-oxidising Biocide A as a periodically shock dosed biocide, sanitising cleaning agent or for membrane preservation ‘Table 3: Membrane Biocide Treatments Function Conditions of Use Tntermittent ‘shock dosing” on-line | Dose 60 - 80 ppm to feed water for 4-ohrs/day Non potable applications only ‘Sanitising Cleaning Agent Recirculate 0.3% solution for & 10 hrs recede and follow with alkaline surfactant ‘Membrane Preservativel Biosiat | Preservation period up to 7 days: 200ppm up to 6 months: SO0ppm 4.2 Maintenance cleaning ‘Maintenance cleaning will ensure optimum membrane lifetime and permeate production. Routine cleaning of membranes should always be carried out at a lower transmembrane pressure (TMP) than that used for water production. It is recommended that an operating pressure of less than 4 bar with minimal permeate flow is maintained for cleaning ‘operations. Cleaning practices should include periodic soaking of the membrane and the use of ‘warm cleaning solutions up to 30°C. Membrane manufacturers’ guidelines should always be followed regarding product compatibility and pH limits. Cleaning of RO systems typically takes between 4 - 12 hours to perform, depending on the severity of fouling and plant size. Cleaning durations of up to 24 hours incorporating overnight soaking may be necessary if heavy biofouling is suspected. Frequency of cleaning may range from monthly cleaning cycles to an annual maintenance clean. There are many alkaline surfactants, acidic formulations and sanitising agents available in the ‘marketplace. The complexity of the clean and number of products required for optimum cleaning conditions is wholly dependent on the composition and quantity of foutant.156 Membrane Technology in Water and Wastewater Treciment 5 CASE STUDY 5.1 European Paper Mill Site Application: Production of boiler feed make-up and process water Details: The RO plant treats town mains supply to produce 500m'/d product for industrial use. Polyamide 8” brackish water membranes are installed, Prior to our investigations, pre-treatment included - sand filtration, acidification (to pH 5.5), S micron cartridge filters, polymer Antiscalant X and dechlorination with sodium bisulphite. The system operates at 75% recovery. Problem: The plant was suffering from fouling and the membranes required cleaning every 2 weeks to maintain the required treated water volume. Fouling Investigations: Membrane autopsy revealed an orange/brown foulant ‘covering the membrane leaves and plastic spacer. Chemical composition of the major foulants was determined as 60% organics, 11.4% calcium carbonate and 15% iron oxide. Microbiology Results: Microbiological enumerations and identifications were performed. Bacterial Counts Fungal Counts (cfu/em?) Membrane 30x10" 20 Plastic Spacer 88x 10° 2 Product Water Carrier = 4.5.x 10° " ‘The following were identified as predominating inthe foulant: Bacteria: rod shaped bacteria, Arihrobacier Fungi: Trichoderma Biocide tests evaluated a fast acting non-oxidising Biocide B st 200ppm and 400ppm ‘concentration with a 30 minute contact time. A 100% bacterial and fungal kil rate was achieved at the higher concentration ‘Cleaning Tests: Crossflow cleaning tests using fouled membrane samples demonstrated that an alkaline clean followed by an acidic clean would successfully remove the organics, biofilm iron and inorganic scale, Water Analysis and Antiscalant Proposal: The feedwater supply was of good quality containing negligible quantities of iron, Iron was detected in the feedwater to the RO and the planktonic counts were 1.6 x 10° efuiml, Inspection of the pre-treatment plant revealed corrosion of some pipework and the inside of the sand filter vessel. A computerised scaling prediction programme calculated the brine LSI using the non- acidified feed water at 20°C, pH 7.4 and 75% recovery as +1.79. ‘Conclusions: ‘+ Corrosion was due to prolonged acid dosing and poor selection of materials ‘©The antiscalant in use was not inhibiting scale ‘+ ‘There was insufficient microbiological control atthe site Recommendations: ‘+ It was proposed that the polymer Antiscalant X was replaced by a phosphonate ‘Antiscalant A. This product was to be dosed at 278 ppm without acid adjustment. ‘+ It was recommended that acid dosing should be ceased to eliminate the isk of further corrosionFouling and Cleaning 137 * The following cleaning programme was proposed: Step 1: Alkaline Surfactant A Seep 2: Biocide B Step 3: Alkaline Surfactant A removes biofilm and ‘other organic debris Step4: Weak Acid CleanerD removes iron oxide and inorganic scale Outcome: All of our recommendations were followed by the site. The plant is now ‘operating well with no indication of severe biofouling, scale or corrosion on the cartridge filters or membranes. Cleaning frequency has been reduced to every 4 months. 6 CONCLUSIONS 1. The selection of appropriate proprietary chemicals and their use in conjunction with ‘good pre-treatment design will ensure cost-effective operation and optimise product water quality and membrane lifetime. 2. This paper demonstrates how better monitoring and laboratory investigations can identify and anticipate the cause of poor performance and allow recommendations to be made to optimise future operation. This can result in significant cost savings to the operator end end-user.FOULING CHARACTERISTICS OF MEMBRANE FILTRATION IN MEMBRANE, BIOREACTORS MH Thomas', $ J Judd? and J Murrer* ‘Barth Tech Engineering Ltd, *School of Water Sciences, Bamsley, Cranfield University, South Yorks Beds S75 3DL 1MK43 OAL, “Technology Group, ‘Anglian Water, Peterborough, PE3 6WT 1 INTRODUCTION ‘The combination of membrane filtration with a biological reactor is known as a membrane bioreactor (MBR), in which the membrane normally replaces the sedimentation stage of a conventional biological process. An MBR may have the membrane module submerged within the reactor (integrated) or as a separate unit (sidestream), and Key aspects of these have been reviewed elsewhere ', but can be summarised as follows: ‘© Small footprint, Membrane modules required to perform the separation occupy @ smaller land area than the sedimentation tanks required to treat the same flow. The reactor can operate at a higher mixed liquor suspended solids (MLSS) concentration leading to a smaller volume needed to treat the same waste, leading to a further ‘eduction in overall footprint for the system. ‘© High quality efluent. The membranes used have a low pore size (typically 0.1ym for ‘microfiltration), which means the effluent suspended solids (SS) content is very low. ‘The reduction in microorganisms is much greater than for conventional techniques. In an earlier study using an ultrafiltration membrane, it was found that the effluent contained no heterotrophic micro-organisms’. ‘Better control over biological conditions. As the solid/liquid separation is complete, all sludge can be recycled to the reactor. This means that the sludge age is independent of hydraulic retention time, giving more control over the biological process. This can lead to reduced sludge production and greater contaminant removal’. The high shear ‘environment found in some sidestream MBRs can lower the average particle size in the biomass’. This size reduction is thought to aid mass transfer in the biomass, offering a possible explanation for improved nutrient removal rates’ ‘A sidestream MBR relies on a recirculating pump to provide pressure to drive the filtration. The membranes are operaied in erossflow mode, and the ability to generate aFouling and Cleaning 159 high trenamerbrene pressure coupled with high Reynolds numbers yields a correspondingly higher permeate flux than that attainable from the submerged configuration. The increased energy consumption is thus partly offset by the decreased ‘membrane area requirement compared to the submerged configuration. On the other hhand, te higher fluxes mean that the sidestream configuration is more prone to fouling — the accumulation of particles and soluble species at the membrane surface causing an increase in resistance to permeate flow. The propensity of membranes to foul is dependent on the feed solution, membrane type and operating conditions. The MBR should be designed to minimise fouling, as either physical or chemical cleaning adds to the cost and complexity ofthe process. The major limitations of using membrane processes for this application are economic. The capital cost of membranes is high and their useful life short compared to conventional separation techniques. The driving force for membrane separation processes is pressure; hence operating costs are also high. For membrane bioreactors to be economically viable, the membrane separation step must be designed for optimum The flux, J, of clean water across a membrane with no materials deposited on the surface ‘can be described by ’: o where = permeate flux (m'm?s") AAP = pressure drop across the membrane (Nm?) {1 ~ absolute viscosity of the water (Ns) = hydraulic resistance of the membrane 12 Membrane fouling Fouling may arise from particle deposits on the membrane surface, macromolecules adsorbing onto the surface or into the bulk membrane material, or pore blocking. The increase in membrane resistance is manifested as a decline in the permeate flux, as predicted by Equation (1), The solutions being filtered in an MBR contain high levels of suspended and dissolved material. The presence of these species has a significant influence on permeate flux. Earlier studies have shown the build up of membrane foulants to occur in two stages”: ‘ Initial flux dectine due to concentration polarisation: the presence of dissolved substances in the solution causes an accumulation of solutes on the retentate side of ‘the membrane producing a layer less permeable to water than the membrane on which itresides. + Long term fouling due to solute adsorption and particle deposition (gel layer and cake formation if extemal, pore blocking if internal): high concentrations of solutes at the ‘membrane surface may cause precipitation to form a ge! layer; particles in suspension are transported to the membrane surface and form deposits which reduce the hydraulic permeability and so permeate flux.160 ‘Membrane Technology in Water and Wastewater Treatment Convective transport of particles and solutes to the membrane surface is partly balanced by diffusion back into the bulk solution, possibly induced by shear. The characteristics of the system determine the balance of forces to and from the membrane surface. 1.3 Factors affecting membrane fouling ‘The three principal factors affecting the rate and extent of membrane fouling comprise” 8) Membrane type: the membrane material, pore size & distribution and module configuration. b) Operating conditions: factors such as pressure, cross flow velocity and turbulence. ©) Solution characteristics: the nature of both solvent and solute, concentration and nature of the bulk fluid. Published studies into the effect of increasing solids ‘concentration of the filtered solution have reported a decline in permeate flux""™. 2. MATERIALS AND METHODS A pilot scale rig was purpose built to test the performance of commercially available ‘membrane modules. The mixed liquor used in the study was drawn from an existing aerobic biological reactor. A schematic of the rig is shown in Figure 1, and an analysis of the mixed liquor at different concentrations is shown in Table 1. Table 2 details the three ‘membrane modules used during the trials. (Ceanog soliton Figure 1 Schematic of experimental rigFouling and Cleaning 161 Table {Sample analysis of mixed liquor used during tials “Measurand — | Mixed liquor infeed tank (met) a 5S BOE IT Hw = 295 FAIO TT Ta900= ST BOD, 15602324 19004574 2810536 2960+ 295 cop 2960401 «47002907 75701000 15600912 NEN 64 a4 1545 m1a5 TKN 1246 2152106 82417 res boc 72406 ie4 20210 1526 ‘Table? Specifications of membranes used daring tials | Material — Configuration MICO (RD 7 Pore Total membrane size (m) Mi PVDA 12mm tubular 250 M2 PVDF —12mmtubular 250 M3. PVDF 12mm tubular _0.1 3. RESULTS & DISCUSSION 3A Permeate water quality ‘The permeate water quality was consistently high. The low pore size of MF (0.05-10 m) and UF (0.001-0.1 m) membranes and the dynamic membrane formed by surface deposits ensure a high rejection of suspended solids". The permeate was analysed for key contaminants during all experimental runs (Table 3). The measured valves for SS and BOD, in the permeate: were normally below the limit of detection. In previous studies ‘using comparable membrane types analysis has shown that heterotrophic microorganisms and viruses are also rejected’. A reduction inthe dissolved organic carbon content shows that the membranes were rejecting large organic molecules as wel as particulate matter. ‘The rejection of high molecular weight compounds and their return to the reactor are ‘thought to explain the improved mineralisation of organic matter in MBRs". 32 Flux decline with time Initial fouling of the membrane surface caused a severe exponential decline in flux to a semi-stable low value after a few hours. The semi-stable flux value was taken as being that after 24 hours continuous operation. Fouling also generated an increased pressure drop through the module. To enable direet comparison of the resistance due to fouling, flux, the permeate182 Menbrane Techaology in Water and Wastewater Treatment Table3___Permeate analysis results summary - ‘Removal, Contaminant MI ma , 3S > oo $99 > 99 BOD, >99 >99 >99 cop >99 >99 >99 NH,-N 60-96 50-98 63-93 TKN 15-98 83-98 85-98 Doc 65 30-72. 16-70. 33 Crossflow velocity ‘The convection to and diffusion from the membrane surface determine the rate of fouling, The rate of convection to the membrane is a funetion ofthe permeate fux, and the diffusion away is linked to the degree of turbulence. An inerease in the crossflow velocity directly increases the degree of physical scouring at the surface and hence {improves back transport into the bulk sotution. ‘The specific flux after 24 hours was recorded under the range of cross flow velocities and feed solution concentrations for each membrane listed (Table 2). The resistance of the fouling layer was found by subtracting the resistance to clean water flow from the total hydraulic resistance measured. The reduction in resistance asthe velocity inreases ‘© approximately 3 ms" (Figue 2) can be attributed to increased shear induced back diffusion. The increase in resistance above 3 ms" is likely to be due to a more compact fouling layer and increased pore plugging duc to the higher pressure associated with a higher feed flow rate. Published studies" have indicated that turbulence promoting baffles and other devices can reduce fouling. The use of these methods can increase back diffusion without increasing the trans-membrane pressure. This offers the advantages of increased cross flow velocity without the penalty of higher pressure and associated coma s MI B6 a Es| — gal i ie} i he 21|— Ze tory} Figure 2 Resistance of fouling deposits afler 24 hours (feed concentration 2500 mgt)Fouling and Cleaning 18 34 Feed solution MBR offer the ability to support a high biomass concentration inthe reactor. However, solution with a high concentration of suspended solids exhibits different behaviour to that of pure water. The solids increase the density and viscosity of the bulk fluid. This influences the flow regime at the membrane surface, and hence the propensity for fouling. More importantly, solids directly cause fouling through cake formation and pore blocking ‘The effect of suspended solids concentration on the final specific flux (atthe optimum cross flow velocity of 3.1 ms") is illustrated in Figure 3. The data shows that fouling increases with feed solution concentration for a given operating regime. This direct comparison also illustrates the influence of membrane characteristics on fouling behaviour. The differences seen between the membranes tested in this study can be largely attributed to the surface properties of the membranes, as operating conditions ‘were replicated as closely as possible. Published studies of membrane fouling show various relationships between stabilised flux and increasing solids concentration (Table 4. i am | aa ne bie = 8 is meoeBe be ° | © 2000 4000 60008000 10000 12000 14000 16000 ‘Suspended solids concentration (mg/l) Figure 3 Final specific flux change with feed suspended solids concentration 38 Energy consumption ‘The energy consumption of the recirculation pumps was calculated from the measured ‘values of flow and pressure through the system using equation (2) and (3): Power (W or Nm") = Flow(m’s')x Pressure (Nm?) @) ‘The energy consumption per unit permeate is found using: Energy consumption (kWhm®) = __Power (W)/ 1000 ® Permeate flow rate (mh)168 Membrane Technology in Water and Wastewater Treatment The results showed the minimum energy consumption per unit permeate produced to be 1.75 kWhm?, as achieved by membrane M3 at the lowest cross flow velocity employed (23 ms"), This illustrates the influence of the pressure drop through the system, which increases in proportion to the square of the velocity, and the higher permeability offered by the larger pore size. The energy consumption of a conventional activated sludge plant is around six times lower than the value found for membrane filtration “®. ‘Table 4 Comparison of reported dependence of flux on SS “Vemibrane type | Solution type and Pressure Veloclty Specific flux. J, Ref and pore size |S concentration (bar) (ms!) (Imi*s"bar'!) vs SS (mgt) cone. (mgt!) MI erable mixed 04 J = 12.0 In Fiquor +133, 2,500 - 15,000 M2 Aerobic mixed 04 31 J, = -47.4 Inxs) liquor #505 2,500 - 15,000 MB Aerobie mixed 04 34 J, = -73.4 Ins) liquor +763 2,500 - 15,000 Ceramic 4mm | Anaerobic mixed 0.35 2.0 ‘J, -0.003(8S) + 15 tubular 0.2m — | liquor 1416 2,500 - 22,000 Ceramic 4mm | aerobic mixed 0.5 3.0 “54.1 In(SS)_ 3 tubular liquor from + 621.8 300 kDa synthetic WW 2,100 - 15,400 Polymeric UF | Aerobic mixed 0.3 . 94.9 In(SS) 16 50KDa liquor + 1090.9 5,000 - 15,000 4 CONCLUSIONS 1, Permeate water quality is consistently high for all membranes tested. The membrane with the highest pore size (M3) proved the most economical to operate with no detectable deleterious effect on effluent quality. 2. The resistance of the fouling layer is minimised at a cross flow velocity of approximately 3ms" for the membrane modules tested. 3. The semi-stable permeate flux after 24 hours continuous operation shows a logarithmic decline with increasing suspended solids concentration. The relationship between semi-stable flux and suspended solids concentration of the feed solution has ‘been compared for the membranes used in this tral and other published reports. The wide differences between findings indicate that membrane surface properties play large part in the fouling characteristics. 4. The lowest energy consumption for the fltrtion was 1.75 kWhm? and was achieved atthe lowest cross flow velocity using the membrane with the largest pore size.Fouling and Cleaning 6s 5. The high energy consumption of a sidestream MBR must be justified by a need for the particular benefits offered by the provess. This need could arise from strict effluent control, high land costs, prohibitive sludge handling and disposal costs or high strength wastewater requiring specific biological conditions. References 1. T.Sato and Y. Ishii, Wat. Sci. Tech., 1991, 23, 1601 2. K-H. Choo and C-H, Lee, Wat. Res., 1996, 30, 171 3.__N, Cicek, H. Winnen, M. T. Suidan, B. E. Wrenn, V. Urbain and J. Manem, Wat Res,, 1998, 32, 1553. 4, E,Trouve, E., V. Urbain and J, Manem, War, Sei, Tech., 1994, 30, 151. 5. A.D. Bailey, G. S. Hansford and P. L. Dold, Wat. Res., 1994, 28, 297. 6, American Water Works Association Research Foundation (AWWARF), Lyonnaise des Eaux (LB), Water Research Commission (WRC) of South Africa, Water Treatment ‘Membrane Processes. New York: McGraw-Hill, 1996. 7. M,H. Ab-Malack and G. K. Anderson, War. Res., 1997, 31, 3064. 8G. Belfort, and F. W. Altena, Desalination 1983, 47, 105, 9. A.G. Fane and C. J. D. Fell, Desalination 1987, 62, 117 10. J. A. Howell and S. M. Finnigan, in Effective Industrial membrane processes - Benefits and Opportunities, Elsevier Science, 1991, p49. LL, M. C. Portes, Microfitration. Synthetic Membranes: Science, Engineering and Applications, D. Reidel, 1986, p225. 12. K. Brindle and T. Stephenson, Biotechnology and Bioengineering, 1995, 49, 601. 13. R. W. Field, D. Wu, J. A. Howell and B. B. Gupta, J. Membrane Sci. 1995, 100, 259. 14, T. Ueda, K. Hata and Y. Kikuoka, War. Sei Tech., 1996, 34, 189. 15. A. Beaubien, M. Baty, F. Jeannot, E. Francoeur, and J. Manem, J: Membrane Sci 1996, 109, 173. 16. Y. Magara and M. Itoh, M., Wat. Sci. Tech., 1991, 23, 1583.CLEANING OF MEMBRANES IN WATER AND WASTEWATER APPLICATIONS Ralf Krack, Technical Manager Membrane Cleaning / Water Treatment Henkel-Ecolab GmbH & Co. OHG, P.O. Box 130406, 40554 Disseldorf 1 INTRODUCTION For more than 25 years membrane filtration processes are used as state of the art technology in many diferent applications e.g. seawater desalination and applications in food processing industry and biotechnology. With increasing costs for water and ‘wastewater a rapidly increasing tendency of using membranes can be observed in Central Europe. LA What is water? ‘Water in the definition of the terms of this topic can be sea water with high amounts of sodium chloride, surface water with varying amounts of minerals and organic ingredients like humic acids but as well normal potable water already treated water from the water works. 1.2 What is wastewater? Wastewater can be the effluent water from an industrial company discharged to the sewer, this is know as “end of pipe” treatment. Here, a wide range of compounds are present in varying concentrations. Wastewater can also be water treated using biological processes. Here, bacteria, residues of flocculants and sewage sludge can present. ‘Wastewater can also be a part of a production plant e.g. a washing machine or botle ‘washer. In this case an influence on the composition of the wastewater can be made and ofien not only the water can be reused but useful products, which can have an added value for use in the production process again. 1.3 Which membrane processes are in use? Depending on the feed and quality demands of the filtrate any filtration process ranging from microfiltration up to reverse osmosis involving a wide variety of membrane materials can be found. Ifa laundry or a car producer wants to meet the consent limits for the discharge of wastewater a simple microfiltation or ultrafiltration unit may beFouling and Cleaning wr appropriate, If they want to re-use the water a nanofiltration or a reverse osmosis plant is needed. The composition of the wastewater in terms of viscosity, turbidity, fibres and abrasive materials determines the type of membrane c.g. tubular, hollow fibre, flat sheet or spiral wound. The permeate quality, stability of the membrane, lifetime and amount of investment could determine ifa ceramic or polymer membrane is chosen, 2. CLEANING 2.1 Why is cleaning of membranes essential? ‘Membrane surfaces - for potable water production - have to be cleaned like any other surface that are in contact with food or pharmaceuticals to guarantee a high quality product as well as hygienic standards. in addition a good cleaning procedure is also essential forthe functionality and eapacity of the membrane plant, even if bacteriological requirements are as low as in most wastewater applications. 2.2 What are the requirements for ficient cleaning? ‘A sufficient cleaning can only be guaranteed, if the applied cleaner is optimised for removing the soil and does not adversely influence the membrane characteristics. As a ‘simple example, the use of a caustic soda applied in highest concentration will not be able to remove a mineral sealing ‘Additionally the physical aspects of membrane cleaning like temperature, mechanical forces, time and last but not least the chemical activity are points of major concem in ‘membrane cleaning processes. Even the most important “solvent” in membrane cleaning the water - should be of the best quality; otherwise the cleaning itself could become & cause for fouling, 2.3 Physical aspects of membrane cleaning, Pree ot coarky aoe ‘The “Sinnersche Kreis", Figure 1 and 2, is one of the most important and well known descriptions for cleaning in general that can be used for membrane cleaning. To get the ‘optimum cleaning result, time, temperature, mechanical force and the right chemical168 Membrane Technology in Water and Wastewater Treatment activity is needed, Ifone of the parameters is deeeased atleast one ofthe others has to be increased in many cases two or ll thee have to be increased. 23.1 Temperature Its well known that an increase of the temperature increases the reactivity of processes. This is the same for cleaning processes. As a major value it ean be stated that an increase of 10°C doubles the cleaning activity. Nevertheless depending on the different foulants @ minimum temperature might be necessary. Many polymer ‘membranes unfortunatly allow only a certain temperature. Any exceeding of these limits can cause sever damage of the membrane. For th usc itis complicated to understand the reasons for different limitations even if the membrane material applied is the same igure3) [gare Even if the membrane mattals are the | SiRrsa na phase same, the spacer material, the support material, and the glue may influence the tmodule stability. For that reason iis important to have all dta available 0 as to make good cleaning recommend Especially in case of water and wastewater treatment plans there i. no option to heat up the cleaning solution That is one of the major differences compared 10 food application where. on very large plants witha cleaning volume ‘of 3000 lites a possibility for heating is State of the at. If there is a wastewater treatment plat producing a latex emulsion at 40°C the pre rinse should already be 2 done at the same temperature. Ifthe pre Le RS rrinse can be done only with cold water the | carps eqn dees retare emulsion may break down and wil ead to coarse ‘much more dificult leaning conditions. Ce In some eases cleaning might be impossible. Comparable may happen if a wastewater ‘with high amounts of proteins i treated. High concentrated proteins can be liquid at high ‘temperature but can become solid if they are cooled down. 2.3.2 Mechanics The relationship between flow and cleaning results are well known from different models and practical experiences. Higher flow rates lead to higher turbulence and to better emulsifying and dispersing properties. The increase of flow rates is limited by pump capacity, mechanical stability and of course in the economies. During ‘membrane cleaning itis important to watch the pressures both parallel to and vertical to the membrane surface. Therefore it is necessary to illustrate the different types of fouling depending on the different membranes. On ultra- and nanofiltration membranes as well as on reverse osmosis membranes mostly the fouling or scaling layer is directly on the surfaces (figure 4). Using microfitration membranes as well as surface fouling an inner porous fouling appear (figures). During cleaning of UF, NF, and RO plants the filtering capacity should be as Jow as se itis possible that a secondary membrane builds up during cleaning. is the best to clean first with a low pressure to take away theFoullng and Cleaning 169 loose cake on the surface then discharge this first dirty solution and finally change to a higher filtration capacity to clean the pores of the membranes. igo = meen Feutng of UF «membrane Hee ‘An often discussed cleaning procedure is backflushing (many membranes and modules, cannot be back flushed). In practice we have leamed that back flushing is very effective in improving membrane capacity wien a hard cake layer is present, although a totally clean situation will never be achieved. By using backflushing, as a method of removing crude surface fouling, care must also be taken that the monomolecular and oligomer fouling is removed, The removal of monomolecular and oligomer fouling is independent of the direction of flow; chemical aspects are most responsible for the clean. Another method is a cleaning from the permeate to the concentrate side. This may be an altemative process on membranes which allow this type of cleaning but it will only function well if the dirt is discharged and not recirculated. If the cleaning solution is recireulated then emulsified fouling will come from the feed side (dirty side) to the ‘permeate side (clean side) and may foul the membrane from the permeate side. However even if the permeate side is not fouled, an additional risk is the possibility of contamination with bacteria, 2.3.3 Time Generally it is stated thatthe effect of an increased time increases the cleaning efficiency. That is valid for many cases. Especially in membrane cleaning the practical experiences have shown also different effects. During cleaning our membrane system is still able to filter out some only emulsified or dispersed soils. “The exact time fora cleaning can only be given by the practical experience or by direet control of the cleaning process. 2500 2000 +500 =c00, +1000 500 °. ° 2 40 6 80 Figure 6, Effect of cleaning time170 Membrane Technology in Water and Wastewater Treatment 24 Chemical aspects of membrane cleaning Generally the chemicals applied for membrane cleaning are as important as the water quality. Many users are not aware ofthe problems that can appear if the wrong chemicals or water quality is applied, 24.1 Water quality Water itself is a very good solvent and also a good cleaner. Water used in membrane cleaning should be of good quality. Any impurities can be filtered out during cleaning and may block the membrane rather than cleaning them. Prefilters should remove any suspended solids. The bacteriological contamination should be ata very low level to prevent biofouling, Several metals form salts, which precipitate especially iron and manganese, these; together with silicates can form insoluble salts The normal deposition of non-silicate sats of iron and manganese may be removed but silicates can only be removed by hydrofluoric acid. This is not a viable proposition as, using this, most membranes and nearly every plant would be destroyed, not including ‘ny human risks. Water hardness is not a problem. With formulated detergents, both acidic or alkaline hardness can be complexed or removed. Very often demineralised ‘water is used, being produced by ion exchange, RO plants or evaporators. Water from evaporators can sometimes cause problems on membranes, as antifoamers are often used in defoaming evaporated solution. These antifoamers can escape into the vapours Contamination with even very small amounts ean decrease the membrane capacity with time. Measuring the silt density index (SDI) can easily test the water quality. There are published values for water quality from which it can be determined whether or not they can be treated by RO plants. Here the SDI should not be higher than 5. For cleaning of ‘membrane plant the SDI should be less than 3, otherwise problems may occur. 24.2 Influence of the soil The soil itself has a direct influence on the chemicals, which should be used for cleaning the membranes. For membrane cleaning itis necessary to determine the type of soil present much more than for usual stainless steel surfaces. 24.2.1 Fat, oil and other hydrophobic fouling The removal of hydrophobic fouling like fat from hydrophobic surfaces like organic polymers is more difficult than from steal ‘or glass. This is depending on the hydrophobic characteristic of fat that adsorbs more strongly on the membrane surface. The removal can be done with surfactant based products above the melting point of the fat or grease in aqueous solution. Below the melting point itis nearly impossible to remove fat and grease residues without using @ real solvent like ethanol. The surfactants must be chosen precisely, because they have to bbe compatible with the membrane, the spacer and the support. Surfactants and defoamers used in normal CIP detergents or washing powders are often not compatible with membranes even if temperature and pH limitations are suitable. Some membrane ‘manufacturers have approved normal washing powders in the past and could reach quite good cleaning results. However a normal washing powder is usually changed at least ‘once a year and of course nobody will take care thatthe old version was approved on a ‘membrane system, 2.4.2.2 Proteins Alkaline detergents remove proteins best, this is well known from practical experience. The higher the pH the faster the protein hydrolysis and the better the solubility. Protcin solubility is poor in the neutral pH range. Ata pH of 4-5 milk proteins for example are denatured and precipitated. The initially used pH-sensitive membranes like cellulose acetate led to the development of enzymatic detergents. Many of today’s commonly used membranes are pH resistant. Initially, only inorganic ceramic membranes were pH stable. Latterly also organic polymers like polysulfone,Fouling and Cleaning m polypropylene and polyvinyldfluoride with high pH stability were developed. The exact specifications vary from manufacturer to manufacturer. The best cleaning process is not achieved by attaining the correct pH alone. A pH of 11-11.5 is very often the limit; even ‘a small amount of caustic may be enough to reach this value but not sufficient to clean well. Additionally to the alkalinity there is a need for dispersants, emulsifiers, soil carrying agents, stabilisers for hardness salts, buffering systems and available chlorine or ‘oxygen as cleaning boosters. Enzymatic cleaners are usually applied if the pH limitation is ator below 10 oF if high level of dirt is present. 2.4.2.3 Minerals and salts Most ofthe mineral scale can be easily removed by an acid cleaning with organic or inorganic acids. Alkaline products ‘containing complexing agents can remove some others. In cases of silicate and sulphide residues itis best to prevent than to clean. 2424 High molecular polysaccharide, EPS extra cellular polysaccharide Polysaccharides, in particular, cause severe problems if they appear on non-oxidising stable membranes. Here two typical phenomena appear. The one is that EPS is directly fixed on the membrane surface and lead to a strong decrease of the capacity. The other especially in spiral wound membrane systems appearing is thatthe EPS seftle down in the spacer material and rise the difference pressure. Both can lead to a loss in the performance of the membrane system. The cleaning process in such cases has to be adapted individually. One successful process to decrease the difference pressure is the enzymatic cleaning followed by an acide sanitation and another enzymatic cleaning step Figure 7). Development of difference pressure with two different ‘leaning processes (Figure 7) ‘Whereas with the old cleaning procedure a cleaning became necessary every 7 to 10 days, the new enzymatic based process could lengthen the cleaning cycles to around 30 days. Trials that were done during optimisation of the plant performance with additional disinfection steps could kill most of the biofilm producing bacteria but did not remove them from the membrane and spacer material, So the dead bacteria became ideal food for ‘new bacteria and speeded up finally the developments of new colonies. 24.2.5 Cleaning agents from industrial companies Many wastewaters contain cleaning agents fom the factories that may influence the behaviour of the membrane system. If only part of the total wastewater of a factory is treated then it is easier tom ‘Membrane Technology in Water and Wastewater Treatment control what detergents are used. Within sereening test it can be controlled what influence the detergent has on a membrane system. If there is a decrease in capacity ‘mother detergent fr the factory cleaning has to be applied. The producer of the cleaning products should be able to offer a product, which fulfil as well the cleaning in the factory as the treatment by a membrane unit In some eases the pH during recycling is of major interest. There might be a cleaning agent that can be recycled at a pH of 8 without any problems but if decrease the pH to value ofS a blocking ofthe membrane may appear. I is well known that the temperature has a strong inTuence on the capacity of membrane systems. Usually by increasing the temperature the capacity will increase as well. If a wastewater with some cloud point depending antfoamer is treated the reaction may be the opposite. Meanwhile the product is clear and good soluble at 25°C it becomes insoluble at 45°C and decrease the capacity ‘of the membrane system or even block it. 3 CONCLUSION Because of the costs and inconvenience incurred when changing damaged or blocked ‘membranes in large filtration plants, much care should be taken. In any case it is recommended to do pilot studies before using a large membrane plant. Such studies should include the membrane cleaning aspects as well as any changes that might be necessary to give excellent results. Today's membrane technology is in most of all cases able to produce clean and healthy water independent of the feed quality. To make the processes beneficial to the user, the membrane manufacturer, the plant supplier and the producer of detergents these all have to work together from the very beginning of new projects.Water ReuseWATER REUSE FOR THE NEXT MILLENNIUM - MEMBRANE TREATMENT AT ‘THE MILLENNIUM DOME J.H. Khow!, A. J. Smith!, A. Rachwal’, A. Donn” and C.V. Meadowero "Thames Water R & D *Leopold - PCI Membranes Spencer House, Laverstoke Mill, Manor Farm Road Whitchurch Reading RG2 OIN Hants RG28 TNR 1 INTRODUCTION ‘As we move into the new millennium, increasing population and water usc will lead to increased demands on water resources world wide. South East England is no exception, ‘with water demand increasing by approximately 2% per year. The long term trend of climate change towards hotter summer temperatures and drier winters will challenge conventional surface and groundwater resources which require the winter rains to top up reservoirs and groundwater levels. Sustainability of water resources is a key factor in ‘maintaining supplies to our customers. ‘Thames Water's water resources strategy is based on the management and enhancement of conventional resources, the investigation and use of altemative water sources where appropriate, and water conservation through the application of new technology nd the “understanding of customer demands and behaviour. In accordance with tis strategy, Thames Water, in association with the New Millennium Experience Company (NMEC), have implemented the first major in-building recycling scheme in the UK. The wide ranging objectives of the scheme are to promote sustainable ‘water use, demonstrate and research water recycling technologies, evaluate water efficient appliances and investigate public attitudes to water recycling initiatives. “The water recycling system has been constructed at the Millennium Dome, Greenwich which is to be the focus of the country’s millennium celebrations. At a cost of £758 zillion, the Dome is the largest building ofits kind in the world with a perimeter of 1 km and height of 50 m at its centre, An estimated 12 million people will visit the Dome during the year 2000', ‘As part of the Dome’s environmental and water management strategy Thames Water's ‘Water Reclamation Plant will supply 500 m* per day of reclaimed water to flush all of the WC’s and urinals onsite. Water is reclaimed from three sources, greywater from the hand ‘washbasins in the toilet blocks, rainwater from the roof of the dome and groundwater from the chalk aquifer below the site,176 Membrane Technology in Water and Wastewater Treatment 2 WATER SOURCES AVAILABLE FOR RECYCLING 2.1 Grey Water ‘The Dome has six core buildings that house the toilet blocks. Greywater will be collected from the hand basins in these blocks. The expected 35,000 - 55,000 visitors each day and. staff are predicted to use, on average, 120 m’/d of hand basin water which will then be ‘treated and reused. This figure is the best estimate based on available information, including the Chartered Institution of Building Service Engineers (CIBSE) design recommendations’ and studies of hand basin water demand in domestic situations™. 2.2 Rainwater ‘The 100,000 m* surface area of the Millennium Dome roof provides the opportunity to capture large quantities of rainwater run-off during rainfall events. Rain runs from the roof through specially designed hoppers that feed into the surface water drainage system. ‘Although the potential volumes are high, the constraints of the site mean that the area available for storage is restricted, A maximum flow of 100 md is pumped from the main surface wate collection system into the surface water treatment system. 23 Groundwater London, like many other Metropolitan cities, has a problem with rising groundwater. Historical over-pumping for industrial and commercial supply from the chalk had reduced natural groundwater levels by over 50 m. However, since 1970's decline in these pumping rates due to the changing industrial base has allowed a rapid recovery of water levels. ‘Groundwater levels have already risen by over 35 m over the last 20 years with rates of rise under the City of London and Westminster of over 2 m per year. GARDIT (General Aquifer Research and Development Investigation Team), a multi organisational grouping set up to ook atthe problem of London’s rising groundwater has identified the Greenwich peninsular as a potential borehole location where pumping from ‘the aquifer would be beneficial. A 110 m deep borehole has been drilled on the Millennium Dome site to provide the 600 m'/d required and also to contribute to the de- ‘watering of the aquifer. 3 PROCESS SELECTION AND PILOT PLANT STUDIES Due to the tight deadlines the project required fast tracking specific pilot trials for process selection, The treatment processes were chosen with additional regards to security of supply and future flexibility. 3.1 Grey Water Treatment ‘The greywater feed quality was established from previous work" and trials carried out using the handwash basins in the Thames Water R&D office complex. A survey of soaps and detergents! manufacturers confirmed that all handwash soaps are significantlyWater Rewse i” biodegradable and that the majority of modem soaps supplied for large scale public applications include synthetic surfactants’. The synthetic feed quality used for pilot trials is shown in Table 1. Table 1 Synthetic greywater feed quality Determinants | units | Synthetic Greywater Toial BOD. mgt 39 Soluble BOD mel 37 Total COD. me 133 Soluble COD, mei 109 TOC mg/l 34 ‘Suspended solids mg/l 18 ‘Anionic surfactants | mg/l 6 ‘Non-ionic surfactants | mg/l 21 NH, -N mg/l 20.05" Total Kjeldaht W mg/l 38, Total P mg 02" Fiardness as CaCO, [melt 2 (C= limit of detection ‘A literature survey showed that world-wide recommendations for total BOD for flushing purposes ranged from 20 mg/l to 2 mg/I*®, From Table 1, it can be seen that greywater requires further treatment prior to reuse, even for toilet lushing, The relatively high BOD concentration will lead to biological growth in the reclaimed water distribution system, with the added risk of odour formation. The nature of the source of the greywater also sives a high risk of contamination with pathogenic micro-organisms’ Considering previous work, the proposed treatment for the hand basin greywater is biological treatment using a Biological Aerated Filter (BAF) followed by membranes". In onder to ascertain the optimum configuration of these processes a range of pilot trials were undertaken with specific emphasis on BOD removal’. The performance of a pilot seale BA, followed by a variety of membranes were investigated using the synthetic preywater. ‘The BAF comprised of two downflow columns, each with a diameter of 150-mm, a height of 2 m, and a total bed volume of 0.036 m’. Lytag pulverised fuel ash media was used in ‘both columns. The membrane test rig used for treating the BAF effluent consisted of six single tubes, with a total membrane area of 0.22 m?, It was operated in a batch mode with a variety of PCI’s ultrafiltration (UF), nanofiltration (NF) and reverse osmosis (RO) ‘membranes. The results in Table 2 showed that BAF is capable of achieving a total BOD of between. 20-25 mgf/lin the effluent from an initial value of 60 mg/l, yielding a reduction of 58-67% tolal BOD. All of the membranes tested achieved permeates containing between 2.3 to 10,6 mg/l total BOD, within the 2-20 mg/l recommended range, In order to minimise biological growth in the distribution system, tight UF membranes, type 2 and 3 were probably sufficient for hand basin greywater treatment. However, before a final decision ‘on membrane type was made, the quality of the other source waters had to be considered78 Membrane Technology in Water and Wastewater Treasment ‘Table 2. BOD reduction following BAF and membrane treatment Membrane types ropes —UFtigt NF RO a+ greys BAF 123 4 influent effluent Total BOD (mg/l) 60 20-25 106 51 63 3 230 48 “Boluble BOL ‘610 244. sejetion characteristios MWCO 200,000 Daltons FP 200 1 2 ‘modified polyethersulphone MWCO 6,000 Daltons EM 006 3 polyethersulphone MWCO 4,000 ES 404 4 polyamide fim 75% CaCI rejection AFC30 5 polyamide film 80% NaCl rejection ‘ARC 30 6 cellulose acotate 90% NaCl rejection CDA 16 3.2 Rainwater Treatment ‘As the Dome has not been constructed atthe design phase the roof runoff quality was not available. However as an indication of water quality to expect, the composition of roof runoff from buildings at a major European city has been reported!” and is shown in Table 3. The rainwater analysis highlighted a significant amount of hydrocarbon and heavy metals such as cadmium, copper including lead, probably attributed to vehicle exhaust pollution, The wide variation between the minimum and maximum concentration of the ‘measured parameters is probably due to the intensity, duration and period between rainall event, Due to the potentially high concentrations of contaminants especially after a long dry spell, the precaution of avoiding the higher contaminated “first flush” was ineluded in the design. Constructed wetlands are commonly used for surface water treatment. Micro-organisms that exist in the roots of the reeds break down micro-pollutants efficiently while heavy metals are absorbed into the root system. The process can cope with flow surges, is economical to operate and maintain and will not need to be replaced over the duration of the project. Reed beds were also chosen for their aesthetic value and to demonstrate the performance of a ‘low-tech’ natural treatment. ‘Table 3. Composition of Roof Runoff in Paris Parameter Units [Min | Max__| Median Suspended Sods | mg/l 3 308 25) cop ml 5 318, 31 ppoDs___| mgt i—_|_ 77 4 Hydrocarbon ug | 37 [eas 108, Cadmium wel_| 01 32 13 Copper uel 3 247 37 Lead ul | 16 214 _[ #93Water Reuse 19 33 Groundwater Treatment Analysis ofthe groundwater is shown in Table 4. High concentrations of sodium chloride, hhardness and hydrocarbons were found in solution, with hydrogen sulphide gas also detected. The inorganic and organic water quality results suggest thatthe borehole wate is 1 mixture of chalk groundwater and ‘Thames tidal river water together with contamination ‘from industrial pollution, possibly from the industry previously located on the sit Hydrogen sulphide can be removed from water by aeration, chemical oxidation and adsorption onto carbon. Chemical oxidation was selected because of the advantage of converting the sulphide to sulphate which is innocuous. Hydrogen peroxide was chosen as the oxidant as it does not form any by-products. The two possible oxidation reaction reported are: HS +0, = Vx $,+2H,0 where frequently x=8 (1) HS'+4H,0,~S0,'+41,0+11 @ ‘A stoichometric excess of hydrogen peroxide is required to effect reaction 2 within 15 minutes”. Following chemical treatment the groundwater is passed through granular activated carbon columns to effect removal of hydrocarbon. The remaining salinity, hardness and residual pollutants would be removed by RO. ‘Table 4 Analysis of Groundwater Quality Parameters Units Minimum | Maximum | Average it 73 73 [74 ‘Conductivity uSfem 4090 4665 4342 Colour Hazen [1 5 aT Ti Fonnaain_| 13] 308 | “ee TOC mg | 28 32 3 TDS, 180°C mg/l 2419 2943 2705 Hardness as CaCO, mg/l 74 962, 783, “Tal ron wet —| tel | 223 [200 Dissolved Iron m7 1 72 1 Manganese wel a a [39 Calcium mg/l 149 159 155 Magnesium mgl_| 87 35 30 Sodium mg/l 680 369 756. Alkalinity as CaCO, mg/l 275 324 293 ‘Sulphate mgt 165 29, 27 ‘Chloride mg/l 1208 1470 1309 Nitrate as N mg/l 1 7 T Fluoride Tae | 85 [Toluene er 72 € 76 Xylene neil 147 16a 15.6 Benzene nel 20, 2 205 Biiyibenzene ugh z a [82 Sulphide ngt[ 1 z 16180 Membrane Technology in Water and Wastewater Treatment 4 TREATMENT PROCESS FOR THE DOME 4.1 Deseription of Overall Treatment Process ‘The final treatment processes selected is shown schematically in Figure 1. Grey water is collected in a balancing tank prior to biological treatment through a BAF at a hydraulic loading rate of approximately 2 m’/m‘/d to reduce the BOD load. Figure 1 Process schematic ofthe grey water recycling treatment process Groundwater is pumped out of the borehole, dosed with hydrogen peroxide and held for 4.15 minute contact time to oxidise the hydrogen sulphide and the dissolved ferrous to ferric iron, The water then passes through two granular activated carbon (GAC) absorbers ‘arranged in parallel to remove organic contaminants. Each absorber contained 5 m° of carbon and was designed to handle a hydraulic flow rate of 10 m°/m*Mh. ‘The rainwater is passed through a series of reed beds and a lagoon which form part of the landscape design. The first reed bed is designed for storm water treatment at 5.0 /m’/d. Following this is storage lagoon of approximately 400 m*. The second reed bed will perform a tertiary treatment function. Each reed bed has an area of 250 m? and is designed for a maximum flow of 100 m'/d, All flows beyond this will be discharged to the River Thames. The beds are approximately 0.6 m deep and have a 0.5% gradient. The media is washed river gravel of 5-10 mm, planted with the common reed, Phragmites ‘australis, and grown from an appropriate seed to ensure sat tolerance™ All three influent process streams combined in a single balance tank where they are mixed and pumped to the UF plant. There is a hierachy for the treatment of the water sourees. Greywater is treated in preference to rainwater and groundwater is the least preferred source. UF was selected to remove pathogens and particulate materials so that the RO plant can operate optimally. RO is required due to the salinity and high level of hardness of the sroundwater. Both the produet streams from the BAF and Reed Beds are treated through the UF/RO membrane systems to remove the potential for biological growth to a greater degree and provide extra security within the reclaimed water distribution system within the Dome.Water Reuse 18 The RO product is disinfected with a residual chlorine dose, re-hardened to reduce its corrosion potential on plumbing fitings and stored before being pumped tothe Dome for flushing the WCs and urinals when required. The water quality will exceed curent draft UK standards for flushing" 4.2 Integrated Membrane Treatment 42.1 Ultrafiltration The fully automatic UltraBar UF hollow fibre membrane plant ‘was built by PCI Leopold. The schematic of the UF process is shown in Figure 2. The plant is designed to treat a feed flow of 700 md and to achieve a recovery of 85%. The membranes have a nominal pore size of 0.01 micron and are made from hydrophilic polyether sulphone/polyvinyl pyrrolidone blend. The single stack unit consist of fourteen standard RO pressure vessels fitted with twenty-cight hollow fibre membrane elements cach with an area of 21m. ‘A programmable logic controller (PLC) automatically executes a pre-selected filtration, backwash and cleaning programmes. In the process mode, feed water is pumped at ‘typically 0.5 bar through a 100 micron backwashable sereen and fed to both ends of the pressure vessel in a dead end mode. Under hydraulic pressure contaminants larger than the ‘membrane pore size are rejected and are collected on the inside lumens of the fibres. The UF filtrate is collected in a balance tank, which also feeds the RO process. At the end of the filtration cycle a small volume of UF permeate is used to reverse flush the membranes to remove the filtered contaminants from the membrane modules. Periodically 100 mg/l of chlorine is dosed into the backwash to inhibit bacterial growth in the system. The UF plant is also designed to allow dosing with acid and caustic to control mineral scale or organic fouling. Due to the extremely low tolerance of the RO membranes to chlorine, the UF filtrate flow to the RO plant is dosed with sodium bisulphite for a period of time after each chlorine backwash to ensure complete removal of free chlorine. A redox meter is fitted on the RO plant upstream of the membrane modules to shut down the plant if any free chlorine is detected to prevent membrane damage. Baciovasn RO teed ur tank filtrate oro “aad 2 Sodium cate ————} | traBar uF membranes Bisulphite rmixed = prep g ser water tosewer ¥ 1008 Sackwashing fiter Figure 2 Schematic Diagram of the UF Process12 ‘Membrane Technology In Water and Wastewater Tearment detergent cleaner ‘acid cleaner leaning tank © Ba = 200-0 SIS ee acid — 93 array a me Figure 3 Schematic process for the RO plant 42.2 Reverse Osmosis The fully automatic RO plant was supplied by PCI Leopold. ‘The plant is designed to teat a feed flow of 600 m'ld and to achieve a recovery of 85%. ‘The schematic flow diagram of the RO process is shown in Figure 3. The UF filtrate is dosed with sulphuric acid to aciify the water in ord to control carbonate scaling during the concentration process. Houseman Permatreat 191 antiscalent is dosed as a scale Inhibitor to provent precipitation of insoluble salts. The feed water is then prefilter through ‘4 nominal 5 micron cartridge filter and the pressure boosted prior to filtration across the RO membranes arranged in a 5/3 array to achieve the optimum hydraulic efficiency of the system, The membranes used are eight inch diameter and sixty inches long thin film composite spiral wound membranes. ‘A clean in place (CIP) cleaning system is designed to enable cleaning of each stack individually. Argo Scientific Bioclean 103A and Bioclean 511 would be used to remove organic and inorganic foulant from the membranes. 5 CONCLUSIONS The project posed @ number of technical and logistical challenges during the design, procurement, installation and implement stages of the recycling scheme. The fast track project has culminated in the first in-building greywater recycling install in the UK. ‘The venve provides an ideal location to carry out on-going aeademie research into water recyeling, water effcieney and conservation and to communicate the water wise message toa wide audience. Research i being cated out to assess the role of membrane processes 25 suitable technologies for future reeyeling of greywater and other altemative water ‘sources in the 21* century.Water Reuse 183 References 1, The New Millennium Experience Company Ltd Annual Report and Financial ‘Statements forthe period ending 31 March 1998. 2. CIBSE, Chartered Institution of Building Services Engineers Guide, Volume B, CIBSE, London, 1986. 3. D. Butler, E. Friedler and K. Gatt, Wat. Sci. Tech, 1995, 31(7), 13. 4. JB.Rose Wat. Res, 1991, 25(1) 37. 5. D. Christove-Boal, R-E. Eden, and 8. McFarlane, Desalination, 1996, 106, 391. 6. J. Murrer, and G. Bateman, IChemE Conférence, London, 26% March 1998. 7. R.Birks, MSc Thesis, Cranfield University, School of Water Sciences, 1998. 8. J. Crook, Wat. Sci. Tech, 1991, 24(9), 109 9. N.M.Kayaalp, Desalination, 1996, 106, 317. 10. J-H. Tay and P-C. Chui, Wat. Sci. Tech. 1991, 24(9) 153 11, B, Jefferson, Wer News, 15 Sept, 1998. 12, Gromaire-Mertzet. al, Wat. Sci Tech, 1999, 39, 1. 13. DF Samuel and M.A. Osman, Chemistry of Water Treatment, Butterworth,1983. 14. PE, Cooper, GD. Job, M.B. Green and R.B.E. Shutes, Reed Beds and Constructed Wetlands for Wastewater Treatment, WRC Swindon, Wiltshire, UK, 1996. 15, Building Services Research and Information Association (BSRIA) Report, 1997.WASTEWATER RECLAMATION CASE STUDIES, THE BENEFITS OF OUTSOURCED MEMBRANE SYSTEMS David Threlfall Ecolochem International, inc. Hydrohouse Peterborough PE 2 6SE United Kingdom | INTRODUCTION Industrial and Domestic water consumption has increased significantly year on year Worldwide. As consumption has grown the costs of treated water have grown also. Municipal sewage is a low cost altemative that has become viable through the development of membrane water treatment systems. In the UK the recent shortages in domestic water supply have raised public awareness of recycling and reuse schemes. This paper will present three examples of outsourced membranes systems used for ‘municipal sewage reclamation. All these systems have been delivered on a Design Build Own Operate Maintain (DBOOM) contact by Ecolochem ine. in the NAFTA region and Ecolochem Intemational, inc. in the UK. Operational performance and experience of these systems is presented for each system. ‘Outsourced membrane systems offer significant advantages to the end user; minimal or investment, reliability of supply, and ‘hands off” system operation. The end a Tow cost alternative water supply without the concems of manning, cleaning and maintenance that a capital system would require. Short or long term trial systems, without capital expenditure, at sizeable flowrates can be used to judge the economic advantage of outsourcing. Treated sewage presents a serious challenge to the operation of an industrial water treatment system. The wide variety of organic and inorganic contaminants in the water process selection critical. Treated sewage will contain high levels of suspended solids, organic end inorganic colloids and biological material compared to usual well or surface ‘water supplies. Reverse osmosis treatment of the supply will not only reject ionic species but also form a physical barrier to colloidal, bacterial and organic matter. The choices for reverse osmosis treatment are presented along with comparisons to other membrane techniques.Water Reuse 18s Table 1 - Water Quality for Mexico site. Raw feed Treated water ‘TDS @ met 600 ~1000 20-60 Silica @ Sime/t 15~30 1-3 TOC @Cmg/ 5~20 40 NTU during operation. This was outside the design conditions and presented a severe challenge to the filtration system. Biweekly cleaning of the first pass membranes was required until the STW improved the quality of the effluent, when the turbidity fell to the design level the cleaning frequency fell to once per quarter. The level to < 10 ppm. 3.4 Alternative Systems ‘The development of ultrafiltration, microfiltration and membrane bioreactors are opening up other methods of wastewater treatment and reclamation. The systems described above utilise simple well understood technologies with an extensive track record. Currently combinations of MF or UF technologies with reverse osmosis (PA) can be used as an alternative to media filtration / (CA) RO. Micro and ultra-filters can be used to reduce the colloidal, particulate and bacterial contamination of the wastewater before polishing RO is used to reduce the ionic content. The membrane filters remove the common foulants from the feed water but leave dissolved nutrients. Both filtration systems are available in crossflow and dead end filtration modes, however they are liable to fouling from the rejected material and need frequent cleaning. The downstream PA membranes also foul relatively rapidly, over several days, as the nutrient content of the water feeds any biological growth in the system. Every membrane cleaning reduces the membrane life, and Teads to increased costs over the lifetime of the plant. ‘Membrane bioreactors produce high quality effluent, but operate at very low flux, this makes the system expensive, as a very large membrane area is required. As yet there is18 Membrane Technolog in Water and Wastewater Treaoment insufficient capacity, inthe right areas, to feed industrial needs. As the costs for the MBR systems fall then installations may become more attractive. ‘Other developments in the composition of polyamide membranes to give a ‘low fouling” membrane are of interest. However until the membranes are resistant to free oxidants, such as chlorine, they ate still likely to fouling several times faster than cellulose acetate. 4 BENEFITS OF OUTSOURCED MEMBRANE SYSTEMS ‘An outsourced system is designed to meet the customer’s demands for quantity and duality of treated water. To mect this end through the service contract and to minimise the ‘operating costs the design will be conservative and based on reliable technologies. ‘Competition between capital equipment manufacturers is intense and design integrity may be reduced to meet a cost. ‘Common techniques to lower the capital coat for equipment isto increase flux reducing the number of membranes, to specify low pressure polyamide membranes or fit high surface area membranes. Each of these techniques will lower the cost of the system but can lead to high operating costs and busy maintenance schedules. All these methods lead to high fouling rates, this may not be observed during the short commissioning period but will soon be apparent. Frequent cleaning of membranes is expensive and gives the added burdens of extra waste disposal and chemical handling. Without a flexible design, responding to changes in the feed water quality may be difficult. ‘An outsourcing contract is typically for 5 to 15 years, over this time the service company hhas to meet the contract specifications or payment may not be received. Long term ‘operational experience with a wide variety of systems gives the service provider a pool of {knowledge to draw on without resorting to expenses extemal consultants, An additional benefit is the available provision of mobile emergency equipment, this can be used in the event of the unexpected o to increase the supply of treated water over short periods or to cover plant outages ‘The removal of operating cost from the customer by the service company Jong term economic plans can be drawn up to compare the anticipated costs between capital or ‘outsourced systems, The ‘cost or ownership” for outsourced systems has been shown to be lower than capital purchase over the lifetime of the contract. ‘At anytime during the contract the service provider will offer the opportunity to increase cfficieney if technology is available. This option is available without additional expenditure, unlike capital purchase, and may reduce costs significantly. 5 CONCLUSIONS Industral feed water production from reclaimed wastewater will play a larger role in the fiture Industry has a great demand for water but is seen as a lower priority than domestic or agricultural users. The availability of treated sewage for exploitation is limited by the provision of pipelines to potential industrial users. Why put water to river discharge, at a ‘cost, that could be sold as a commodity? “The case histories presented have been chosen to illustrate the situations where outsourcing can delivery a low risk, economic advantage over a capital system. The choices made by the service company differ from those of the equipment supplier, and are made on the basis of long term objectivesWater Rewse 89 Outsourcing water treatment services is an increasingly common choice. The benefits of fixed costs, and guaranteed produetion over the lifetime of the contract are advantages that capital purchase cannot meet. References R.T. Taylor, in Proceedings ofthe 56 Annual Meting, International Water Confrence, Engineers’ Society of Westem Pennsylvania, 1997, p 442. 8. Mortis, Water and Environment Manager, 1999, March, p10 BB. Birkenhead, “Wastewater is now a viable economic water source for water companies and industry’, presented at Eurochem/ET, June 3°, 1998, Birmingham.COMPARISON BETWEEN DIFFERENT OUT-TO-IN FILTRATION MR/UF MEMBRANES FOR THE RE-USE OF BIOLOGICALLY TREATED WASTEWATER EFFLUENT Emmanuel Van Houtte, Johan Verbauwhede, Frans Vanlerberghe and Johan Cabooter Intermunicipal Water Company of Veurne-Ambacht (IW.V.A.), Doormpanne 1, B-8670 Koksijde, BELGIUM 1 INTRODUCTION ‘The Intermunicipal Water Company of Veurne-Ambacht ([WVA) and the Municipal Water Company of Knokke-Heist, producing and distributing drinking-water respectively at the western and eastern part of the Flemish coastal area, plan 10 re-use wastewater effluent for the artificial recharge of the existing dune water catchments. Artificial recharge would create a sustainable groundwater production preventing the ingress of salt water under the dunes. ‘Microfiltration was chosen to pre-treat the effluent; to remove salts and nutrients the ‘produced filtrate would be further treated with reverse osmosis, As a result of previous tests! using microfiltation and ultrafiltration membranes it was concluded that out-to-in filtration was the most effective way to treat wastewater effluent. It was decided to evaluate different systems with out-to-in filtration before tendering the full-scale plant. Three systems competed: two submerged systems (CMF-S from USF MEMCOR and ZeeWeed from ZENON) and one pressurised system (CMF from USF MEMCOR). Alll worked in dead-end mode. ‘This paper describes the results of these comparative tests. 2 DESCRIPTION OF TREATMENT SYSTEMS 2.1 Testing procedure and data acquisition The tests were done from June Ist until October 16th 1999, Both the CMF and the ZeeWeed installation performed during the full period. Tests with CMF-S started on ‘August 18th. The pressure and flow were logged continuously. The quality of the produced filtrate ‘was compared to the feed water (effluent) quality. The following parameters were controled on a periodic base: acidity (pH), conductivity, BOD, COD, total phosphorous, nitrate, ammonia, sulphate, total coliforms, faecal coliforms, faecal streptococci, heterotrophic plate counts (HPC) and UV-absorption. ‘The SDhis of all filtrates was controlled periodically and for the last month also particle counting was performed on the filtrate of the submerged systems.Water Reuse 1 Figure 1 Process scheme of comparative tests 22 Description of the systems 22.1 Effluent intake and pre-treatment “The raw effluent was taken from a reservoir with two submersible pumps both housed in Imm sereened PVC-tube. This pre filtered efluent entered a first Im’ reservoir. From this fist reservoir the feed was taken for the Zee Weed instalation. ‘The effluent flowed from the fist to a second Im? reservoir, but in between ammonia and chlorine (a mean of 1.4 mg free chlorine/}) were dosed. From this second reservoir the chloraminated effluent was pumped through an automatic backwashing SOQum strainer to feed the CMF and CMF-S installations. 2.2.2 CMF installation It was a 3MIOC unit from USF MEMCOR containing 3 modules; each housing 20,000 polypropylene (PP) hollow fibres with a nominal pore size of 0.2m; the total membrane surface is 45 m?. ‘The filtration, with a flux of 3,240 Vb, is from the shell (outside) to the lumen (inside) side of the fibre where the filtrate is collected. The contsminants aecumulate atthe surface of the membranes causing an increase of pressure with time. At the end of each filtration cycle, either at a preset time (20 minutes) or earlier ifthe rate of resistance to filtration is increasing rapidly, a gas backwash is initiated. This gas backwash — explosive release of compressed air from lumen to shell side ofthe fibre ~dislodges the contaminants from the surface of the membrane after which they are removed from the system by a cross-flow of food water. ‘As solids gradually accumulated over time a periodic chemical cleaning (CIP) was required when the transmembrane pressure (TMP) reached 130 kPa. A CIP consisted of an acid clean followed by a caustic clean and restored the TMP below 50 kPa.192 ‘Membrane Technology In Water and Wastewater Treaament 22.3 CMP installation It vas @ 1S10T wnit from USF MEMCOR. The module is equipped with the same fibres as the CMF unit. One S1OT module contains 13 m? of surface area. Inilly the filtration time was 15 minutes ata flux of 1,000 Vi; the clean water is drawn through the membrane wall by pump suction pressure into the centre of each fibre. Daring the backwash low pressure air is forced through the bundle of the membrane module (from bottom to top) creating a scouring effect on the extemal fibre surface. A shor filtrate flow (2,250 V/s from the lumen tothe shell side of the fibres supplements the air to further dislodge the contaminants that are drained from the tank atthe end of the backwash cycle. Afler the reservoirs refilled with feed water production can restart ‘Chemical cleaning is achieved by contact of the cleaning solution with the membranes in combination with a periodic aeration. The CIP is performed when the TMP was 85 kPa land restored the TMP to 20 kPa. On September 16th the flux was increased to 1170 Vs; on October Ist the filtration time was set to 20 minutes. 224 ZeeWeed installation ‘The ZeeWeed module (ZW-500) contained chlorine- tolerant polyvinylidene Muoride (PVDF) hollow fibres with a nominal pore size of 0.035pm and an absolute pore sizeof 0.1m; the total surface area was 42.8 mt. ‘The membranes operate under a partial vacuum created within the hollow fibres by the ‘operation of a centifugal pump. filtration period of 215 seconds at a Mux of 1,800 Uh is {allowed by a back pulse of 15 seconds. During a back pulse low pressure air introduced atthe bottom ofthe membrane bundle is combined witha reversed permeate flow (2,100 Uh). The contaminants are removed from the reservoir by a small overflow during this back pulse. ‘After a number of backpulses an extended back pulse is performed. Chlorinated filtrate ‘low (1,000 VM) was pumped through the membranes for 120 seconds followed by a short soak period of 120 seconds, always combined with air ‘Table 1. Characteristics ofall three out-to-in MF filtration systems used CMF CMES ZeeWeed "Membrane fibre —|0.2um pore size [0.2m pore size | .Tum pore size PP, symmetric and|PP, symmetric and |PVDF, asymmetric hydrophobic hydrophobic and hydrophilic non chlorine resistant | non chlorine resistant | chiorine resistant Driving force [Positive pressure via| Suction pressure via| Suction pressure via pumped feed centrifugal pump |centrifugal pump Backwash Pressurised air from|Low air pressure on|Low air pressure on| in-to-out and feed outside of membranes] outside of membranes ‘sweep _aongside| filtrate from in-to-out rate from in-to-out [membranes and draindown, with small overflow cr ‘Acid and caustic (4h) | Acid and caustic (4h) | Chlorine and| hydrochloric acid (2h) Biofouling |Chloramination of |Chloramination of Periodic back _pulse| orale (omen fein ao | ime YYWater Rewse 195 ‘The absolute pressure in the system varied from 1.00 atmospheric bar to 0.45 atmospheric bar. At that stage a chemical cleaning is performed by emptying the reservoir and filling it with chlorinated drinking water or filtrate. The system soaks for 60 to 90 minutes alternated with short cycles of back pulses and filtrate production. At the end hydrochlorie acid is added to remove scaling from the membranes; the last step isto drain the reservoir. On September Sth the module was replaced. The new module had 46 m’ of surface area, ‘The flux was increased to 1840 Vn. 3. RESULTS OF COMPARATIVE TESTS 341 Hydraulic results The ZeeWeed and CMF worked with constant fluxes respectively 40 and 72, Uhm’. The CMF-S initially had an intial flux of 77 Vin? but it was increased t0 90 Vn? 3.1.1 CMF installation "The mean recovery over the whole period was 87.5 % with a maximum of 95.4 %. The mean production period was 14 minutes. The frequency of IP's diminished as the temperature of the feed water increased: in June a CIP had to be performed every I or 2. weeks but in July, August and September a CIP was only needed ‘once in 3 or 4 weeks time. It should be mentioned that during the summer shutdowns of the CMF system occurred regularly due to clogging of the pump and the screened PVC tube in the effluent reservoir. After a shutdown the TMP of the CMF, but also of the CME-S, was always lower compared with the TMP before the shutdown (figure 2). With a constant flux of 72 Wham? the daily effective filtrate production varied between 60 and 68 Va’. 3.1.2 CME-S installation Because the BW-interval was constant the recovery over the whole period was constant: 88.7 %. The installation was only cleaned once during the 2 months of testing; the TMP increase was very slow. Even when on the CMF the BW-interval decreased below 15 minutes, the CMF-S performed very well on the same feed, with a constant BW-interval of 20 minutes and with a higher flux. Generally the TMP increase was 5 kPa during a production period, but it was noticed that the backwash was able to restore the TMP wien occasionally greater ‘TMP increases oveurred. With a constant flux of 77 Vim? the daily effective filtrate production varied between ‘57 and 65 Vh.m’; when the flux was increased to 90 Vian? it varied from 70 to 78 Vh.m?. Due to problems with the feed pump between October 6th and October 12th, the feed for the CME-S was changed to effluent that only passed the Imm screened PVC-tube. This did not have a negative effect on the performance of CMF-S soit could be concluded that 1000,m strainer would be sufficient on CME-S.194 Membrane Technology tn Water and Wastewater Treament A = € e . Figure 2. Performance of the compared MF out-to-in systems “from September Ist until October 15th 3.1.3 ZeeWeed installation It was not very easy to obtain constant recoveries, as the feed flux was not constant. After a flow controller was installed it proved possible to feed the ZeeWeed reservoir only while the system was in production mode; overflow only cccurred during the back pulse. In the last runs - after the flow controller was installed ~ recoveries of 87.5 % ware achieved with a weekly CIP (figure 2). At that time the system ‘already worked on raw effluent as one of the Imm screened PVC tubes around 2 submersible pump was broken (September 28th) and no strainer was placed between this pump and the Zee Weed reservoir. ‘After a chemical cleaning the pressure drop during a production period was 1 to 2 kPa. However as the pressure built up over time this TMP drop gradually increased (figure 2. ‘The mean daily effective filtrate production was between 30 and 34 Vm? with a mean effective production over the period June Ist ~September 30th was 29 V/b.m this includes CIP eycles and shutdowns in that period. Most of the shutdowns were due to a decline in feed flux before the flow controller was installed. Because no pre-screen was needed shutdowns due to clogging of the PVC-tube did not influence the ZeeWeed installation. 3.2 Quality Filtrate samples were taken on a weekly base for bacteriological control. Before sampling chloramination had been stopped during some hours. The SDl,s was controlled regularly and during the lat month ofthe testa particle counter monitored on lin both the filtrate from CMF-S and ZeeWeed. 3.2.1 CMF insiallation ‘The SDIis had a mean value of 1.9 with a minimum of 0.9 and a maximum of 3.4, Generally the SDI’s were lower in dry periods (no dilution of sewage with rainwater),Water Rewse 195 ‘The organic load of the effluent was partially removed by CMF (table 2). No bacteria were found in the filtrate and the log removal for HPC was around 3. Table 2. Quality of the wastewater effluent of the WWTP at Wulpen compared to the filtrate quality of CMF and ZeeWeed (Parameter Jetiluent JCMF ——JZeeWeed Number filtrate [filtrate [samples [Acidity @AY 7 {17 78 [26 [otal phosphorous (me Pt) os2 joao = foar_— fs |UV 254 absorption (em? jo2686 |o2426 |ozara fs [Total Organic Carbon (mg C/1) p42 figs fig fs [Biological Oxygen Demand (mg O2/})_ {8 B 5 is [Chemical Oxygen Demand (mg Oy) 39 33 3 is [Total coliforms (counts/ml) 0°-10% fo. jo 14 lFaccal coliforms (counts/mnl) ho fo lo 4 [Faccal streptococci (counts/ml) hot-108 Jo lo 14 |Log removal HPC 37°C (48 h) hoMio® pas festa ILog removal HPC 22°C (72h) poo B20 fsa fia 3.2.2, CMF°S installation The mean SDIys-value was 2.8 with a minimum of 14 and a ‘maximum of 4.3. However in the last weeks of the tests the SDI’s improved and were ‘comparable to those of the CMF and ZeeWeed filtrate. ‘The organic load of the effluent was partially removed; no bacteria were found in the filtrate (lable 3). The log removal for HPC was around 2 but it the counts inereased with time. This could be due tothe location of te installation. Unlike the CMF, both the CMF- S and the ZeeWeed installations were installed in open air near the aeration basin of the sewage plant. Small sewage droplets in the atmosphere could have contaminated the air (Compressor) and the filtrate reservoir. A sample taken while chloraminating the foed resulted in lower counts that proved the effectiveness ofthis method. 3.2.3. ZeeWeed installation In the first part of the tests the mean SDIys-value was 3.3, but after replacement of the module it decreased to 2.6 with a minimum of 1.6 and a maximum of 3.5. ‘The organic load of the effluent was partially removed and no bacteria were found in the filtrate (table 2). The mean log removal for HPC was around 3 but here also a small increase of counts was observed although not inthe same extent as on the CMF-S. ‘The pH of the filtrate produced with the ZeeWeed unit was slightly higher (table 2) presumably by stripping of carbon dioxide during the back pulse* (airflow). 3.2.4. Particle counting On September 81h a Met One 21SW particle counter was installed on both submerged systems. It continuously counted particles greater than 1m. Dae to the backwash it was difficult to get stable counts, especially on the CMF-S system ‘where the backwash cycle lasted longer. On the Zee Weed filtrate, the amount of particles never exceeded 20 counts/ml and most of the results were between 5 and 10 counts/ml; on the CMF-S filtrate it was slightly higher with a majority of counts between 0 and 20 counts/ml196 Membrane Technology in Water and Wastewater Treatment Table 3. Quality of the wastewater effluent of the WWTP at Wulpen compared to the Jiltrate quality of CMF-S Parameter [Efluent —[CMF-S Number of itrate [samples [Acidity (pH) 76 (7-7 7 [Total phosphorous (mg P/1) 0.57 10.37 3 UV 254 absorption (em 0.3119 0.2812, 8 [Total Organic Carbon (mg C’1) bes hi [Biological Oxygen Demand (mg Or/l) [8 + 3 |Chemical Oxygen Demand (mg Or/})_|39_ 35 B [Total coliforms (counts/ml) 10%-10° fo 7 |Faccal coliforms (counts/ml) 10° jo \7 }Faceat streptococe (counts/m!) o'-10° Jo r Log removal HPC 37°C (4h) otto |i.87 ILog removal HPC 22°C (72h) oro’ lisp | 4 CONCLUSION Three out-to-in microfiltration systems were compared with one another. ZeeWeed and CME-S were submerged systems; the CMF isa pressurised system. All three systems reached recoveries over 85 % and could produce for atleast one week without the need to be chemically cleaned. The filtrate produced with all systems was comparable and free of bacteria All MF systems that were tested proved capable of producing a filtrate quality ready to food a reverse osmosis installation ‘The submerged systems showed some advantages over the pressurised system. The footprints are smaller and the concept is less complicated compared to the pressurised air backwash of CMF; a SOQum strainer is no longer needed so coarser strainers could be sufficient. This could mean that investment and maintenance costs would be smaller. As the transmembrane pressures are smaller the intial pressure could be restored very easily. {A disadvantage is that submerged systems use filtrate forthe backwash that means that the risk of recontamination of those systems is higher. Care should be taken to prevent recontamination ofthe filtrate. However chloramination not only proved to enhance the performance of CMF® but also showed to be an efficient way to prevent recontamination of the filtrate, Using chlorine resistant membranes is an advantage asthe risk on damage is smaller and the chlorine and ammonia consumption for bio-fouling prevention could be reduced However the hydrophobic PP membranes seem to have a slower TMP increase which ‘means that chemical cleanings should be performed less frequently Standardisation of the submerged modules, which would make them interchangeable, could in the future benefit tothe customers, as the investment risks would be reducedWater Reuse 197 ACKNOWLEDGEMENT AQUAFIN gave the opportunity to do the tests om the wastewater effluent at Wulpen WWTP and actively supported the IWVAA in the project. It is important to mention the support from USF MEMCOR and ZENON. Their collaboration made that the trial work could be done under good conditions. We want to thank all the people from IWVA and Knokke-Heist who helped us with the tests especially Delphine, Luc P., Luc C., Pol and Sven. References 1. E. Van Houtte, J. Verbauwhede, F. Vanlerberghe, S. Demunter and J. Cabooter, Treating different types of raw water with micro- and ultrafiltration for further desalination using reverse osmosis. Proceedings “Membranes in drinking and industrial water production’ Conference, Amsterdam, Desalination 117, p 49-60, 1998. 2. B. Birkenhead, Microporous separation technologies. principles and economics. Proceedings Workshop on ‘Microporous membrane filtration technology improving 3.2.2 EDR Section. Key performance criteria here are water recovery, the number of stack cleans, salt rejection (percentage conductivity reduction) and product water quality From the results (Tables 5 and 6) itis evident that EDR treatment performed well. The objective of operating at a calcium sulphate saturation level of 250% was mostly met Excursions did however occur from time to time. The plant however, operates with acceptable levels of manual stack cleaning. The target of six stacks per month, which represents cleaning each stack twice per year, is generally met or slightly exceeded. The plant was cleaned in place on a 3 ~ 4 week interval, which is also acceptable, considering ‘the highly scaling operation conditions. Table 5 EDR Performance Criteria Target Average Recovery 78% 0% Salt Rejection 78-92% 16% CaSO, Saturation Max. 300 450 Average 250 262 ‘Manual Stack Cleans 6 1 Salt rejection was slighily lower than the target but this was sacrificed in favour of higher water recovery. Despite this and the variations in feed water quality, which were within the specified limits, (Table 6), the product water quality was relatively constant and well within the design specifications. Table 6 Mine water feed, EDR and SRO product qualities G Feed concentration EDR Product nee mg/l mgt ms 3994+ 786 1435 4.438 Calcium 42294 36215 Sodium 917279 358 £151 Chloride 828 238 121242 Sulphate 3254.5 842 7018487 Toc 212% Ld 19804 3.2.3 SRO Section. Key performance criteria (Table 7) show that plant performance ‘was acceptable, Flux was however lower than design due to biofouling which could not28 Membrane Technology in Water an Wastewater Treament be removed with standard CIP's. A CIP using EDTA (1%), tr-sodium-polyphosphate (1%) and sodium-dodecyl-sulphate (0.5%) at 35°C and pH 10,5 proved to be very efficient. The target flux of 25 Lm?.h was difficult to achieve due to this unanticipated fouling. The plant was found to run very well in the region of 18 é/m?h where long runs were achieved between CIP's. Biofouling is still considered as a serious threat and the development ofa treatment program and early waming systems are being investigated. ‘Table 7 SRO Performance Criteria Target Average Water recovery (%) 85 79416 Permeate conductivity (uS/em) 80 3349 CAP’s/train/month, 1 25 Flux - dm? 25 20,1432 Feed pressure (kPa) 1350 1350 Planktonic counts efvmé Log3 Log 6 ‘The lower flux of 18 é/m?-h results in a somewhat lower recovery. This is however not considered being a problem, since the brine is recycled to the EDR plantas brine make-up and does not affect the overall plant recovery of 76%. 4 OPERATING COST Average operating costs since start-up were approximately the same for both combinations of processes at ca R3.50/m of final product. 5 CONCLUSION Up to now Sasol’s experience with the treatment of wastewater using membrane technology, was positive. Plants conformed to expectations and the original objectives ‘were met. Where problems did occur, they could be resolved. In the Sasol water systems there is more potential to apply membrane technology and this will most probably realise in the future. References, 1 JG Nicuwenhuis, B. Steytle,1.W. van der Merwe, M.P. Augustyn, ‘The Water Institute of Southem Africa, Biennial Conference, Cape Town, 1998, Vol 3, Section 3E-1.RECOVERY OF WOOL SCOURING EFFLUENT UTILISING MEMBRANE BIOREACTOR (MBR) TECHNOLOGY AS PART OF THE ACTIVATED SLUDGE SYSTEM FOLLOWED BY TWO-STAGE REVERSE OSMOSIS (RO) MEMBRANE CONCENTRATION A.R. Bennett Membrane Systems Division ACWa Services Limited ‘ACWa House Keighley Road Skipton BD23 2UE United Kingdom 1. INTRODUCTION ‘The frst fully automated commercial facility inthe United Kingdom to incorporate & membrane bioteactor(MBR) enables upto 85% of water resoutees required on sitet be recovered for use inthe wool scouring process by using high prssute reverse osmosis (RO) membrane technology. “The ste uflises approximately 13 mk of clean water into the wool scouring process on 1 continuous basis A. zero eluent discharge consent has driven investment in the tflent treatment and recycling system with all wastes being concentrated and tankered aay for disposal, ‘The system incorporates the Zenon ZeeWeed™ microfiltration (MF) membrane system as part of the activated sludge process. Koch ~ Fluid Systems high pressure RO ‘membranes are included to conoentratefitered water from the MBR and recover 859% as permeate withthe remaining 15% waste containing concentrated contaminants The 159% oss is made up from the local mains water supply. 1.1 Development of MBR Technology Early work on MBR technology in the 1960's by Dorr-Oliver in the USA was not ccommercialised ~ it was the forerunner of Zenon Municipal Systems that developed a commercial system in the 1970's containing tubular ultrafilters based on a single studge aerobic — anoxic process (Husain & Coté, 1999). System sizes were small treating flows ‘generally below 100 m’day"' with plants used for treating and recycling wastewater for small companies. thas only been inthe lst ten years that MBR technology has moved fiom the small scale applications to flows in excess of 10,000 m'day''. The pace of development has increased more rapidly in the last five years, along with the increasing number of applications for MBR technology (Churchouse & Wildgoose, 1999). The number of plants and flows treated by the Kubota process provides a good example of the rapid surge in demand from the marketplace. In 1993 four Kubota plants had been installed, the20 ‘Membrane Technology In Water and Wastewater Treatment largest treating a flow of 125 mda". This had risen to 150 with a maximum capacity of 1907 m'day'" in 1998 and the total to date for 1999 in June was 237, The largest plant was now up in the mega litres per day (Mld") eategory at 12,700 m’day'' (12.7 Mid"). ‘Three reasons have been sited by Churchouse and Wildgoose for this expansion:- 1. Design fluxes have increased. 2. Membrane life has increased from 3 to 8 years 3. Increased levels of manufacture have reduced fabrication costs. ‘The result is that MBR technology is now able to compete with traditional effluent ‘treatment technologies, and the MBR is now available to solve various effluent problems. ANOXIC AERATION TANK TANK MEMBRANE BIO-REACTOR gece | ag Boa Figure 1 Effluent Treatment System Schematle 1.2 RO Developments The RO industry has been steadily expanding and there are now a number of major players offering various types of RO membrane elements for applications ranging from desalination to wastewater reuse. Use of RO in effluent treatment has been restricted in the United Kingdon but a substantial amount of work has been undertaken in the United States and the Far East utilising RO technology in conjuction with microfiltration as an effective pre-treatment. Recent developments leading to higher surface area elements has reduced the capital cost of installations and improving manufacturing techniques have lead to a reduction in the feed pressures required resulting in lower power consumption and the consequentIndustrial Applications 2 reduction in operating costs. In parallel to these positive developments has been the improvement in permeate (product) quality achievable due to increased rejection of dissolved ions. Work with RO in the wastewater field has shown this membrane technology to be a consistently reliable technique for separating an effluent stream into a purified permeate for re-use and a concentrated membrane reject for disposal. 1.3 Combination of Technology This project has involved the combination of the MBR and RO technology in « manner that has capitalised on the recent developments of the membrane separation processes described above. It has only been recently that both technologies have been economically attractive enough tobe included ina capital plant projet, and this application isthe first commercial facility to be installed in the United Kingdom. 2, PROCESS DESCRIPTION The process installed in depicted in Figure I on the previous page. It comprises activated sludge with an anoxic zone and an MBR with product from this further processed by the RO system, 2.1 ‘Traditional Activated Sludge Enhanced by MBR Technology Supernatant from the dissolved air flotation (DAF) process is pumped to the aeration tank of the activated sludge system via an anoxic tank for nitrification. The typical analysis of| the feed to the plant supplied by ACWa Services Limited is shown in Table 1 below. The chemical oxygen demand (COD) entering the process following wool scouring and DAF reaches 5000 mgl" with a total dissolved solids (TDS) of up to 8000 mel’. Retum activated sludge is mixed with the DAF feed in this anoxic tank and the facility exists to add chemical nutrients as required to promote biological growth in the activated sludge system. Following oxidation ofthe biomass the mixed liquor gravitates to the MBR. ‘This application of the ZeeWeed"™ MBR process enables the separate anoxic zone to be accommodated outside of MBR reactor vessel. Also, the separate bio-eactor enables the development of an active mixed liquor with levels of suspended solids up to 12,000 mg as opposed to a typical 4,000 mel" found in traditional activated sludge systems. The separate MBR unit also enables easier cleaning-in-place (CIP) of the membranes. Husain & Coté (1999) have also described the MBR system without the additional bio-reactor capacity. This configuration is mainly used for straightforward municipal wastewater applications. ‘The micro-filtration membrane system is rated to remove particles nominally above 0.1 tum in size and the filtered product is drawn up under vacuum through 24 membrane elements packed with polypropylene based hollow fibres. These elements ean be seen suspended in the reactor vessel in Figure 2. The separated biomass is pumped back to the anoxic tank as retum activated sludge and periodically waste activated sludge is removed from the system to maintain the correct sludge age.m2 Membrane Technology in Wer and Wastewater Treaoment Table 1 Influent Design Basis Parameter Value / Unit Volume 312 m‘day" Flowrate 3m'h" ‘Chemical Oxygen Demand (COD) Concentration 3,500-5,000 mgt" Biochemical Oxygen Demand (BOD) Concentration ‘Suspended Solids Settled Solids Total Dissolved Solids Conductivity Ammonia Total Nitrogen Total Phosphate <0.2-04 mgt" Iron <80 mgl" pH 10-15 ‘Temperature 3010 35°C. Figure 2 ZeeWeed™ Microfiltration Unit 2.2 RO Two Stage Concentration Filtered water from the MBR enters the frst stage ofthe RO plant. The fed is pressurised {040 bar g by an inverter controlled multi-stage centrifugal pump to maintain a fixed feed flowrate before entering a 2:1 array of pressure tubes each containing six high rejectionIndustral Applications 2 thin film composite membranes supplied by Koch ~ Fluid Systems. The pressure is boosted between the two arrays by a fixed speed centrifugal pump. The RO process works by pressurising the feed side of the membrane with a pressure ‘eater than the natural osmotic pressure of the solution. Where the natural process of ‘osmosis results in the movement of water from a high asmotic pressure to a lower osmotic pressure through the semi-permeable membrane, the application of a net driving pressure in excess of the natural osmotic pressure reverses the process to produce a small volume of concentrated solution. ‘The amount of permeate produced divided by the volume treated is termed the recovery of the RO system. A two stream RO plant provided for treatment ‘of'a borehole supply is depicted in Figure 3 on page 225. “The first stage recovers 70% of the MBR filtrate as permeate for re-use and utilisation in the wool scouring process. The typical permeate quality required is shown in Table 2 below. An Allen-Bradley programmable logic controller automatically controls the recovery rate of the RO plant along with all other control interlocks for safe and efficient ‘operation of the system. Operator monitoring and contro initiation is provided via a user interface sereen ‘The concentrate stream from the first stage RO is futher processed by the second stage RO plant. A fixed feed flowrate is maintained using a positive displacement plunger type pump capable of generating pressures upto 80 bar g. The feed is thus pressurised onto a single pressure tube comprising five membranes suitable forthe saline concentrate from the first stage of the RO system. 50% of this concentrate is recovered from the second stage to add to the first stage permeate. ‘The overall recovery of both RO systems in this two stage concentrator is thus 85%. RO permeate recycled to the wool scouring process is recovered at a continuous rate of 11.05 mh with a TDS of less than 300 mg/l and a COD of less than 10 mg. The level of dissolved solids is less in the recycled water than in the mains supply. Table2 RO Permeate Typical Specification Value / Unit 13h! 11.05 mt <25 mgt! <05-3.0 mel" <0.1 mgt" Manganese <0.02 mgt" Alkalinity <50 mgt Total Dissolved Solids <300 mgt! Colour <20 Hazen, 3. OPERATING REQUIREMENTS. At the time of writing this document the plant has not been operating for a sufficiently long period to be able to provide detailed operating costs for the process. However, it is hoped to provide details at the Conference in March.24 ‘Membrane Technology in Water and Wastewater Treatment 3.1 Nutrient Dosing ‘The provision for nutrient dosing into the feed from the DAF plant has been provided but this has not been uilised to date. It would be possible to add chemicals providing a combination of N P and K to enhance the biological activity in the activated sludge plant and the MBR. 32 MBR Cleaning ‘The MBR control logic allows for backflushing of the hollow fibre membranes on a periodic basis generally set at one hour intervals. Hypochlorite can be introduced into the backflush water to provide a chemically enhanced backwash and this is generally undertaken on a weekly basis. ‘The trans-membrane pressure (TMP) of the system is typically expected to increase from 0.4 to 1.2 bar g over a period of two months following which dilute backflushing with acid and / or caustic sotutions will be required. 3.3 RO Dosing and Cleaning ‘An anti-sealant is continually dosed at 25 mg" concentration in the feed to the RO plant in order to minimise scaling on the concentrate side of the membrane and also to minimise biological fouling. In addition to the anti-scalant, every week for 30 minutes # non-oxidising biocide is dosed at 400 mel" concentration. This shock dose prevents the formation of biofilms on the RO membrane surface and has the effect of increasing the period required between cleaning-in-place (CIP). CIP chemicals will be required on a monthly basis and the facility exists to use either high or low pH proprietary solutions for periodic recirculation through and soaking of the RO membranes. solution is made up manually in the CIP tank and heated up to 40 °C. A small centrifugal pump i then used to force the CIP liquid through the plant. Effective CAP procedures will reduce the net driving pressure required down close to initial requirements thus prolonging the membrane life 4. CONCLUSIONS ‘The integration of the MBR and RO membrane technologies alongside the tradit vated sludge process has resulted in an efficient system allowing the recovery of water for re-use on-site. This has provided the advantages of reduced mains water consumption, 8 6-fold reduction in waste disposal costs in order to comply with the zero discharge ‘consent and the recycling of RO permeate with 2 lower dissolved solids content than the ‘mains water supplyIndustral Applications ns Figure 3 Two Stream Reverse Osmosis System built by ACWa Services Limited References 1, H, Husain & P. C8té, The Zenon Experience with Membrane Bio-Reactors for Municipal Wastewater’ Treatment, Published in MBR-2 Conference Proceedings, Cranfield University, 1999. 2. S.Churchouse & D. Wildgoose, Membrane Bio-Reactors Hit the Big Time:- from Lab 10 Full Scale Application, Published in MBR-2 Conference Proceedings, Cranfield University, 1999.PERFORMANCE ON A REAL INDUSTRIAL EFFLUENT USING A ZenoGem?MBR, D. Mallon and F. Steen K. Brindle, Anglian Water Services Ltd, ‘School of Water Sciences, ‘Thorpe Wood House, Cranfield University, Peterborough, Cranfield, PE3 OWT Beds, MK43 OAL. Abstract ‘A Zenon ZenoGen pilot MBR (membrane bioreactor) was studied using submerged ‘ZecWeed® MF (microfiltrtion) membranes in activated sludge, treating effluent from a chicken processing factory. The performance of the MBR was monitored with regard to ‘the final permeate quality pre-treatment requirements and potential re-use opportunites. The paper assesses the performance and compares the advantages and disadvantages of the process together with likely costs. The influent had a COD between 2,000 - 4,500 mp/l, BOD average 1853 mg/l, with high SS (suspended solids) and high ammonia content. The COD, BOD, ammonia and phosphate removal efficiencies were higher than 92%, 99%, 95% and 98% respectively. Phosphate removal/assimilation was greater than expected; this was thought to be due to the acclimatisation of the biomass to high phosphate levels, around 50 mgil, in the chicken blood in the influent. The effluent was ‘ofa high enough quality to discharge directly to river making this process competitive in price wit other conventional processes. The effluent was unsuitable for reuse within the factory. Keywords ‘Membrane bioreactor; wastewater treatment; activated sludge, chicken processing effluent. 1 INTRODUCTION In conventional aerobic activated sludge systems, biomass separation from the treated liquor relies on sedimentation of aggregated mixed microbial flocs. If biomass separation ‘rom the treated effluent is facilitated by physical retention within the bioreactor the need for flocculation is removed. The necessary operating conditions can be achieved with cross-flow membrane bioreactors where the membrane acts as a filter to provide betterIndustral Applications mm ‘than clarified effluent. Progress in the engineering of membrane bioreactors and the «development of immersed membrane configurations, has produced compact, robust, cost- effective systems. Immersed membrane systems use either hollow fibre membranes or flat sheet membranes (Bussion etal. 1998)' and consist of membrane modules submerged directly into the activated sludge compartment, Typically they operate in an outside- inside filtration mode at low suction pressures, (I-5 bar). Operation at low suction pressure means a reduction in the transmembrane pressure. This means reduced fouling of the membrane and slower formation of a cake layer, resulting in less frequent backwashing and lower operating costs, as smaller pumps are required. However, ‘because the force driving the liquid through the membranes is relatively low, a large ‘membrane surface is required to compensate for the low flux rates achieved. Sludge ‘wastage is minimised by maintaining a low F/M ratio whl the footprint of the plant is reduced by operating at high biomass concentrations, typically 15 - 20 g/l (Cote and Pound 1997}. In addition, biomass retention allows build-up of a waste-specific microbial population, of particular use when dealing with industrial effluents, thereby providing the most effective biological treatment. The main improvements arising from the coupling of membrane technology with biological wastewater treatment are summarised in Table 1. ‘Table 1. Conventional and membrane biological treatment comparisons. Problems and limitations associated with conventional biological treatment Tmprovement from the inclusion of| membranes in the process ‘Variations inthe pollutant loading “Acceptance of variation in concentration of activated sludge including high concentrations Very slow rate of Kinetic reaction Natural selection and tial retention of the bacterial population Settlement of bacteria in clarifier can rate limit process Removal of bacteria from effluent not dependent upon settling characteristics of bacteria Posi-ireaiment removal required prior to reuse of viruses Significant removal of viruses from effluent make direct reuse possible Tnsufficient contact me — between macromolecules and micro-organisms Terease in contact time allowing effective treatment of low biodegradable products Rate of sludge production creates a problem for disposal Sludge volume reduced with minimised cost of post-treatment Trregular quality of Geated water and absence of effective barrier to bacteria Produetion of high quality effluent, free from bacteria, offering the possibility of| reeyeling Bally Foot-print reduced28 Membrane Tecnology tn Water and Wastewater Treaoment LA Materials and Methods Ceaving in Docc Pane) Sen on mee net sein Coane ale stn Figure 1. Schematic diagram of the pilot membrane bioreactor. All chemical analysis was carried out on site with a Dr Lange colorimeter using standard methods. These results and all microbiological analysis were verified at 2 NAMAS accredited laboratory using standard methodologies’. Bacteriological samples ‘were monitored for E.Coli and Coliforms and viral samples were monitored for Human Enteric Viruses ‘A ZenoGem® pilot plant was fed with effluent from a chicken processing factory that ‘was high in COD, BOD, ammonia, suspended solids and fat. I also contained blood and feathers from the culling process. The effluent was screened through a 2. mm drum and then through two crude fat removal tanks, (Figure 3). These consisted of Im? tanks ‘operated with a constant overflow, the feed being taken from the bottom of the tank using ‘small submersible pump. The pilot plant used inthis study had total working volume of 4 m’, (Figures 1 and 2). It consisted of two reactors in series (of 3 m’ and 0.8 m’, respectively), followed by a membrane compartment, and CIP (cleaning in place) tank, ‘operated under aerobic conditions. Fine bubble aeration was supplied atthe bottom of the first bioreactor and the second was mechanically stirred, Two hollow fibre membrane ‘modules, (surface area 13.8m? each, pore size 0.1 microns supplied by Zenon Gmbli), were submersed in the third compartment. The membranes are chlorine resistant made from plastic, woven to obtain the required pore size. Hollow fibre membranes provide a higher surface area to volume ratio than flat late membranes, thus oceupying less reactor ‘volume. Coarse bubble aerators were positioned at the bottom of the membranes supplying air to increase cross flow velocity and increase mass transfer in the vicinity of the membranes. The membranes were backwashed with permeate for 40 seconds after every 300 seconds of operation. This is to prevent surface fouling and dislodge the cake layer formed at the membrane surface. Provision was made inthe pilot plant for chemical cleaning with sodium hypochlorite, however a chemical clean was not required during this stady. A biomass recirculation pump was in constant operation transfering biomassIndustral Applications 29 from the membrane compartment back into the first reactor. The pH was maintained between 6.75 and 7.25 by acid and alkali dosing, Figure 2. Picture ofthe pilot plant Figure 3. Picture of first grease trap 1.2 Feed Quality ‘The characteristics ofthe screened feed are given in Table 2. ‘Table 2. Feed characteristics “Average Minimum | Maxim Toal COD GagOyD) 3I8i 1280 3068 Total BOD (mgO,) 1652 156 2950 ‘Ammonia (mgN/) 20.78 0.58 848 “Total Phosphate (mgP/I) 19.20, 187 629 2 RESULTS AND DISCUSSION 2.1 Plant Operation The bioreactor was seeded from a local sewage works; the biomass concentration slowly increased over the first month. Initially foaming inhibited the development of a significant microbial population. This was overcome by drip-feed application of ant- foam oil and a reduction in aeration rate. However, overnight foaming caused the MBR. to shutdown when influent detergent levels rose due to factory cleaning. This was reduced by using a timer switch on the firs fat removal tank pump to ensure feeding only during production hours. Foaming stopped when the MLSS reached approx. 10 g/l and sludge levels increased rapidly, around 1g/V/4, to a maximum of 28 g/l. The suction pressure delivered by the pump dropped from 1 bar to 0.68 bar at MLSS of approx. 15-18 #/, indicating a rise in transmembrane pressure (Figure 4). The increase in MLSS was accompanied by a decrease in DO fiom over 2.5 mg/l to below 04 mg/l and a deterioration of permeate quality with the COD rising from 100 - 200 mg/l to aver 350 mg/l. Maintaining the DO is critical to maintaining permeate quality; unfortunately, increasing the air flowrate through the blowers on the pilot plant would have produced230 “Membrane Technology ln Water and Wastewater Treament large rather than fine bubble aeration and a subsequent loss in oxygen transfer. This is not expected to be a problem on a full size plant, which would incorporate DO control. 095, Sludge MLSS. 8 Suction Pressure 08 os Figure 4. MLSS concentration and Suction Pressures. Permeate quality was of a consistently high enough standard to discharge directly to river, BOD levels typically less than 10 mg/l, with an average 92% and 99% removal of COD and BOD, respectively, but had a slight colour from the blood in the influent, (iguee 5). The lower influent COD values around the 22" October correspond to the timer on the feed stream being removed and samples being taken around midday, when ‘the plant was producing permeate from a less concentrate overnight feed. The lower COD values on the 9" and 10* of November correspond to low influent COD concentrations. Flow balancing should ensure a more consistent and better quality permeate, Samples of ‘the permeate showed bacteria and virus concentrations were below the level of detection. ‘Ammonia and phosphate removal efficiencies were excellent, with an average of over 95% and 98% removal respectively, (Figure 6). Average ammonia in the permeate was 0.86 mg/ and average phosphate was 0.23 mg/l. The phosphate removal was particularly good and showed a marked decrease from the start of October onwards, which was accompanied by a slight rise in ammonia concentration inthe permeate. The cause is not fully understood but may be due to one or more ofthe following; 2.2Results o x0 x0 B= ‘@Permeate COD a {ePermeate BOD 10 a 0 Date Figure 5. Permeate COD and BOD concentrations.Industral Applications a 1. Variation of the feed concentrations of ammonia and phosphate, Again reinforcing the ‘need for buffering upstream of a full size plant. 2. Biological phosphate assimilation. 3. Simultaneous nitrification and denitrification caused by an anacrobic nucleus forming in the MLSS, which at higher concentrations became more viscous. 4. Concentrated blood containing up to 50 mg/l phosphate being added atthe start of the study, conditioning the MLSS microbial population enhancing phosphate removal ‘when this high concentration was later removed. ¢Permeate Phospate as P ‘wPermoate NHS a3 7 oh Fano! 7 Figure 6, Permeate ammonia and phosphate concentrations. Loading rate on the bioreactor was 6.2 kg COD/m'/d, 3.2 kg BOD/m'Vd and 0.9 kg CODim'/d, 0.5 kg BOD/m/d on the membrane. Samples taken from the permeate and ‘mains water supply to the site showed a large increase in the concentration of dissolved salts in the permeate. Chlorides rose from 68 to 172 mg/l Cl and sodium rose from 47.6 to 127 mg/l Na. Reuse of the permeate from the MBR in the factory, without further treatment, would create an environment in the bioreactor that would be toxic to the ‘microbial population. However, the permeate was treated using a Reverse Osmosis (RO) plant and caused no fouling to the RO membranes. 23 Cost Benefit Analysis The processing plant where the MBR was trialed curently has DAF pre-treatment before discharge to sewer. Trade effluent bills for a flow of approx. 900 md are crea £300 K per annum and volume discharged is expected to double. Effluent treatment plant onsite requires a small footprint, as space is at a premium, and low odours, as the factory is opposite a residential area. Estimated costs of a ZenoGem® membrane bioreactor to treat 1800 mld are: Capital Expenditure - £1,400,000 Operating Expenditure -18.3p/m’ permeate ‘The permeate from the pilot MBR was of a high enough quality to discharge direct to river and savings in trade effluent costs give an estimated payback period of approx. 3 years.a2 ‘Membrane Technology in Water and Wastewater Treatment 3 CONCLUSION ‘This study monitored the performance of a Zenon ZenoGem* with a real industrial cffluent. Problems occurred with foaming of the biomass that stopped when then the MLSS reached 10 g/. Suction pressures dropped at operation over 15 g/l. The permeate was ofa high quality, COD and BOD 242 mg/l and 6 mg/l, respectively, with no visible suspended solids. Ammonia removal efficiency was 95.8% with an average concentration in the permeate of 0.86 mg/l. Phosphate removal efficiency was excellent and greater than expected at 95.9%, with an average concentration in the permeate of 0.23 mg/l The permeate was unsuitable for reuse in the factory without further treatment. However, preliminary costings show that a direct discharge to river would make an MBR plant a viable economic option for treating this case study waste 4 ACKNOWLEDGEMENT The authors are grateful to Frank Duty and Marilena Demetriadi at Anglian Water Services Ltd. for their help in taking samples and assistance with maintaining and running the pilot plant. References 1. H. Buisson, P, Cote, M, Praderie and H. Palla. The Use of Immersed Membranes for Upgrading Wastewater Treatment Plants Wat. Sci. Tech. Vol. 37, No. 9, pp. 89-95, 1998. 2. P. Cote and C. Pound Results of Pilot Studies for the Reuse of Municipal Wastewater resented at the Workshop on Membranes in Drinking Water Production - Technical Innovations and Health Aspects, L’ Aquila, Italy, June 1-4, 1997. 3. HMSO 1994 Report 71: Microbiological Analysis of Drinking Water. 4, HMSO Methods for the Isolation of Human Enteric Viruses from Waters and Associated Materials. 5. HIMSO pablications Methods for the Examination of Waters and Associated Materials (Blue Books).MEMBRANE TECHNOLOGY IN WOOD, PULP AND PAPER INDUSTRIES Jorgen Wagner (Osmonies Desal DK-2820 Gentofte, Denmark 1 INTRODUCTION This document will describe three membrane applications used in the wood, pulp and paper industries: ‘© Reverse Osmosis (RO) of Medium Density Fibreboard (MDF) effluent using spiral ‘wound membrane elements. ‘© Ultrafiltration (UF) and RO of Spent Sulphite Liquor (SSL) using a combination of a PCT tubular UF system and a Desal™ high temperature spiral wound membrane clement RO system. ‘© UF and Nanofitration (NF) of white water using combination of the plate and frame CR filter from Valmet Flootek and Desal spiral wound NF elements designed for high temperatures. ‘The pulp and paper industry, much like other, is characterised by extremely lange process flows and high volumes of fresh water consumption. Membrane filtration offers the possibility to reduce water consumption, reduce solids discharge or simply enhance ‘existing equipment such as an evaporator. 11 RO of MDF Effluent ‘This process is one of the success stories of recent years. What started as a dream was realised in the cretion of a zero discharge facility ffom an otherwise polluting plant. MDF is a wood based panel, similar to particleboard but much more versatile since it can be machined (planed, turned and routed). It is generally manufactured from softwoods, such as pine. MDF ean be produced according to a wet and to a dry method According to the wet method wood is fist chipped and washed to remove any extraneous debris such as sand. The chips are then steamed to soften the fibres and then fed into a refiner. Excess liquid (MDF effluent is separated from the fibres, which are then mixed with glue and flash dried. The dried mixture is then distributed onto a forming line to create a uniform mat and then finally a heated continuous press presses it to around {Y40th. The final product can then be used with or without @ coating such as laminate. ‘Applications include fumiture, kitchens, laminate flooring and house infemals such as skirting boards and picture rails.24 Membrane Technology ln Water and Wastewater Treament The MDF effluent is traditionally treated in a biological treatment plant. However, a very high Biological Oxygen Demand (BOD) load and formaldehyde from the glue kitchen makes it dffieult to treat and results in poor treated effluent quality. A biological plant occupies much space and consumes a lot of energy in the aeration system. Consequently, there are several incentives to investigate alternative processes that operate ‘more economically. RO membrane technology was identified as such a process. Esmil Process Systems LTD, a long-time Osmonies partner, started fest work in 1993. ‘The objective was to develop a process that recovers all suspended and dissolved solids from the MDF effluent, and allows reuse of the solids recovered. Both goals were achieved. Pre-treatment was, and continues to be, a challenge. It is very difficult for a spiral ‘wound membrane element to handle the fibres that inevitably are present in the MDF cffluent. It was also realised that there were various naturally occurring resins in the ‘water which could disturb the membrane process, and which are difficult or impossible to remove by conventional cleaning methods. Esmil chose flocculation for pre-treatment that proved extremely challenging. Extensive test work was required to identify a flocculent which achieved the pre- treatment goals without damaging the membranes which are typically susceptible to the cationic nature of effective flocculent. Esmil was successful in identifying a satisfactory flocculent although precise dosing became a necessity. If dosing is too low excess suspended solids will be fed into the membrane system. If dosing is too high polymer ‘may blind the membranes and membrane flux will suffer. ‘After flocculation, the liquid undergoes several filtration steps. First, itis sent through a large filter press, the filter cake from which is incinerated to produce energy for steam production. The filtrate from the filter press is fed through a sand filter followed by cartridge filters which act as protective, safety filters more than anything else since physical blocking of the spiral wound elements could easily occur if suspended solids were to accidentally by-pass. Although this extra pre-treatment process may sound expensive, itis, inthe author’s opinion, absolutely essential to ensure thatthe RO plant is not damaged by debris that may enter as a result of upstream failures. Although they are rare, failures eventually occur, and it is better to be safe than sorry. The sand filter is bback-flushed regularly, but the flush water is not diseahrged to sewer as usual - it is sent ‘back to the MDF raw effluent tank for floceulation and RO treated again. ‘The filtered MDF liquid is now treated by an RO system. All dissolved solids are concentrated. The concentrate is simply sprayed on the fibres before pressing. Two advantages are achieved by doing this. Firs, itis an excellent disposal route (there is mo effluent to treat). Second, and more surprising, is that it actually improves board quality and saves a little glue since the sugars polymerise during heating and create a glue-like material. The RO plant is cleaned regularly. The spent Cleaning In Place (CIP) liquid is sent back to the MDF collection tank, which is against all the rules of membrane filtration, but inthis case it works! The RO permeate isthe last stream to discuss. It typically contains <100 mg/L of COD, ‘meaning it can be used in a number of factory applications thereby reducing the consumption of fresh water. ‘The operating parameters are not extreme considering pressures do not exceed 40 bar and temperatures average 30°C. In summary, the MDF effluent and all the chemicals and water sed in the plant may be divided into the following three components:Indusrial Applications 235 Table 1 Typical MDF dara ‘eed Termeaie Concentrate joume 6 Tmt OTT P56 Th T jpround 30” T 0,000 ma FE foo 0,000 me 150 mel 00,000 melt Filter cake, which is incinerated 2. RO concentrate, which is used as glue 3. RO permeate, which is reused in the factory. LLL. No water or solids are wasted. The plant design employs the traditional ‘multistage recirculation design (Figure 1). Although it has been argued that the single- pass design used for water treatment is an alternative; itis, in the author’s opinion, not a Viable altemative since flux must be stable over time in order for single-pass designs to succeed. The flux inthis application is quite variable, which would make the operation of a single-pass system extremely difficult Pumps are, as usual, also a challenge. The feed pump(s) is the easiest to specify since it simply supplies a specified flow at a given pressure. Grundfos pumps were chosen and hhave operated fine. Although the mechanical seals have been problematic, they have actually performed better than expected considering the pump was manufactured to pump cold, clean water only. Figure 1 Membrane Skid Figure 2 Plant Overview ‘The recirculation pumps, on the other hand, have been a problem. Its tempting 10 use borehole pumps, suchas the BM type from Grundfos, snes they are inexpensive. But in the long nun the problems exceed the savings and they eventually were abandoned Problems were as follows: First, the system operating temperature was on the high side for the pump, although within specifications. Since the liquid it pumps cools the pump, i is likely thatthe warm water caused the motor to overheat. Second, the seal between the236 ‘Membrane Technology In Water and Wastewater Treatment pump motor and the product leaked occasionally, allowing MDF liquid to enter the motor which is obviously undesirable. Lastly, the cables were not sufficiently insulated, resulting in short circuits and motor burnout. More conventional pumps, albeit with quite expensive mechanical seals, have proven to operate beter. ‘An indication of this application's sucoess is the fact that Esmil has installed a total of cight plants in Europe since the first plant was installed in Chirk see Figure 1 and 2), and ‘the number is still growing 1.2 UF and RO of SSL Pulp production by cooking with highly acidic Co-sulphite isthe first and oldest method used industrially. Today, there are very few of these mills because the sulphite process has largely been replaced by alkaline cooking, otherwise called Kraft pulping. Sulphite pulping survives because cellulose made from the sulphite process produces a specific type of cellulose, which is dificult or impossible to produce by other processes. For every ton of pulp produced, one ton of solids are dissolved. This means that a factory producing 300.000 TPY of pulp also generates 300,000 TPY of waste water containing approximately 55% lignosulphonate (LS), 25% various sugars and 20% ash by products. The sheer tonnage can flood most markets since there are few buyers that can absorb so much product. The problem of selling huge amounts of a byproduct and the difficulty of finding a market for materials such as lignosulphonate has resulted in the limited use of membrane technology. Borregaard (Norway) is the exception because they employ membrane filtration plants world wide and dominate the lignosulphonate market. ‘There are other markets, but they have for non-technical reasons not been exploited yet. Nevertheless, there are a few RO systems treating Spent Sulphite Liquor that can be reported upon. RO of SSL can potentially replace evaporators, and can certainly enhance the performance of existing evaporators. For many reasons itis the latter which has been ‘most successful, although few systems have been installed. One particular application concentrates permeate from a UF plant treating SSL. An RO ‘membrane system may either treat SSL ot UF permeate, even though the composition of ‘SSL varies considerably. Table 2 shows typical values for SSL and for UF permeate, ‘The UF plant treating SSL. is a tubular system from PCI. The system incorporates12 ‘mm tubes, and the main advantage of the tubular system is its ability to handle suspended solids, especially fibres, and to do so with a minimum of pre-treatment. It is common to use a 20K MWCO PSO membrane in such an application. The operating temperature is around 70°C, and the majority of the LS are rejected by the UF membrane - only the very low MW LS may pass through the membrane. However, LS is not of great concer since LS is easy to wash away and does not influence flux very much. Table 2. Typical data for the RO plant Feedio RO [ RO concentrate | RO permeate Toral somds 6% TH% Ie Conducting 7500 nS T1500 nS 00S ‘Osmotic pressure z 16 ‘The RO system concentrates the UF permeate ftom 7% TS to 14% solids prior to evaporation. The TS in the feed can vary significantly, and the high operatingInductral Applications 237 temperature of 70°C is a challenge for the membrane elements. However, the use of Desal™ Duratherm® Excel membrane elements, and the specification of a 3,8" OD, has secured good mechanical stability for the elements. It should be noted that 3,8" is a ‘commonly used diameter in the dairy industry, but itis certainly unusual in this kind of industry. This element was chosen for its excellent mechanical strength: itis probably the sturdiest element on the market. However, itis likely that a new plant built today would use 8° elements since the designs, and mechanical strength, have improved considerably ‘over the last few years. ‘The plant is built as a two-stage recirculation plant. The plant employs 400 m? of membrane surface area, and the water removal capacity is a nominal 10,000 L/h that was in accordance to requested capacity. It tumed out that the UF plant operated at considerably higher capacity than expected, which means the RO system was too small. In order to treat all the UF permeate as fast as it is produced, a significant enlargement is needed. For many months the system performed as expected and did not show any tendency to lose flux of foul. Lately, however, the RO plant has experienced a peculiar problem of Tow capacity. Since UF permeate isthe feed, the pre-treatment is almost perfect, and no harmful chemicals from the SSL should be reaching the RO membranes. if a UF ‘membrane failed, raw SSL could penetrate, but itis rare to incur a failure so massive that itseriously affects the membranes. ‘The present problem can probably be traced to the water supply since the plant uses river water that contains humic acid. Humic acid is colloidal in nature, and there may be clusters housing microorganisms, which secrete fatty or sticky products as part of their ‘metabolism. When such water is heated, such as for steam production, the fatty partis separated from the humic acid, and the result is water, which contains a material behaving much like mineral oil. Some of these undefined oily substances have been observed, and it has had a rather devastating effect on polyamide membranes. Since pH during cleaning is limited to 11,5, itis difficult or impossible to remove the fouling layer. We are still investigating how or if the membranes can be regenerated, and how to prevent such problems in the future. 1.3 UF and NF of White Water in the Paper Industry White water is a term used to describe the wastewater produced during the formation of paper in the wet end ofa paper machine. The water is rather white due to the high content ‘of suspended solids. Paper machines have grown to almost incomprehensible size and the following are a ‘ew figures that are literally breathtaking. New machines form paper measuring up to 10 meters in width. The speed of the paper is approaching 2000 meters per minute, which is equivalent to 120 km/h, of 75 mph. The largest machines produce 40 TPH paper. Now take into account that the pulp slurry contains less than 2% cellulose fibres, meaning that the volume of white water is around 1800 m’/h. The price for a machine is well over 200 milion US dollars. ‘These numbers reveal two important aspects conceming this application. First, the paper industry’s wastewater problem is serious and very large. Second, an operator of such a machine will not install any equipment that endangers the operation of production: downtime is extremely expensive, In order to achieve the proper paper quality, it is necessary to add several chemicals to the pulp slurry, and some of those in quantity. For instance, itis common to add CaCO,238 Membrane Technology in Water and Wastewater Treatment as filler to obtain the glossy paper used in magazines. The pulp slurry is poured onto fast moving mesh where the formation of paper takes place. One of the major challenges in making paper is keeping the paperforming portion of the machine in good working order. Water is drained first by gravity, and then by vacuum, leaving the pulp. After a few seconds of dewatering, the paper is lifted off and continues into the paper machine. The wire is then subjected to intense cleaning by nozzles, and is then ready again for paper formation, Cellulose fragments some of the filler and some of the chemicals can become wedged in the wire mesh and must be removed. Otherwise, the wire will lose is ability to drain ‘water from the paper, and machine capacity will be limited. The quantity of water needed for spraying is staggering, and itis a challenge to clean the white water by conventional ‘methods to such an extent that it can be used in the nozzles. For years, UF has been used to do just this, and has performed well. However, low flux and low availability, combined with’ too Title paper machine knowledge, has prevented the widespread adoption of UF. In order to demonstrate that UF isa viable, practical process forthe filtration of white ‘water, full-scale plant must be built since a small-scale pilot will prove nothing. Valmet Flootek (former ABB Flootek and Raisio Flootek) has, for more than 5 years, been ‘operating full-scale UF systems on white water. It has been proven conclusively that UF ‘meets all the criteria for use in the harsh environment. It should be noted that Valmet is ‘one of the world’s major producers of paper machines. Therefore, they can combine their ‘paper know-how with the membrane know-how of Flootek to effectively sell a complete ‘and proven system to the end user. This combination of knowledge is unique and necessary forthe success ofthis application. The filter used by Flootek is called the CR filter. It is a plate and frame type, with approximately 84 m per filter. The CR filter gains sufficient turbulence by mechanical means, and does not rely on pumps. The filter employs something similar to pump impellers between each plate, The advantages include: ‘© The volume to be treated and the linear velocity over the membrane are independent variables + The investment is well under US $2 per m? water treated * The volumetric concentration ratio can be varied, and it can be very large without affecting the operation ‘+ The module is not sensitive to suspended solids (within reason) High, stable flux over extended periods of time is possible. The disadvantages include complex module construction The volumetric concentration ratio is approximately 25 meaning that 100 litres of feed are divided into 96 litres of permeate and 4 litres of concentrate. It is known that the permeate recovery can be increased to >99%, however, the extra volume of water recovered is quite small. It is most probably not worth while to attempt higher levels of recovery since the solids content in the concentrate increases enough that it impairs the function of the CR filte. ‘The author has stated in the “Membrane Filtration Handbook” that flux is never >100 [Lm for any length of time. The CR filter has broken that rule by maintaining flux of 200 ~ 300 Lah for weeks on end. This is, of course, not valid for all products, but is for most. ‘A major reason for this occurrence isthe high shear rate on the membrane surface. ‘There are several full-scale systems in operation. It has been proven conclusively that the UF permeate is very well suited for nozzle water and for reuse in other parts of theIndustral Applications 239 ‘paper process. The CR filter has achieved an important goal in papermaking: reduction of water consumption. 1.3.1 NF of white water UF permeate ‘The more water tat is recycled in pulp ot ‘paper mill, the more one has to watch out for salt build-up. One major step forward in the pulp industry is replacement of Cl, bleaching by oxygen bleaching because this replacement removed a major source of chloride. Chloride is feared in all pulp and paper ‘operations because chloride can so easily cause corrosion, that at best results in poor paper quality and at worst, can stop a complete system. ‘Many newer mills have recycled their water to the point that they need a method to remove salt from the system. Nanofiltration and Reverse Osmosis can achieve this ‘objective. For several years, Flootek has been piloting NF of UF permeate and the results have always been good. However, the first priority was to gain acceptance of UF. Now that that has taken place, the next step isto gain acceptance of NF and/or RO. A this paper is being written in 1999, a full-scale system is being commissioned. The data presented below reflects the results from the pilot test. Data from the full-scale, on- line system will be available during the conference in March of 2000. Tables 3 and 4 show data obtained from the full-scale UF systems and the pilot-scale NF system, ‘Table 3 Typical values. Kirkaiemi. Papermachine ! and 2 pH [nS [COD, [TS [Salis [mh [Bar UF | Feed SH [42 | 3xaD THT | aa0 [Tae |= Concentrate 30 [aD | C47 TO9TE | O74 | 180 Permeaie 32 [ar [ia [asides Te NF | Feed 30 [a2 [TIO PATS aR as Concentrate 49 [are [4309 | 1449s] TOTS ‘Permeate 30 [202 [305 | 1246345 UF | Reduction factor ofan [ie [1se NF | Reduction factor 2a [397 [347 [817 Table 4 Typical Concentration of Solutes Ta_[Me [Al [Fe [Cl 80, ]Si0, [meq [ms Feed Ts [89 [2033 1046 [oo [018 [730 Concentrate. | 700 [38 [50 [13 3658 [106 [O55 Permete [7 [01 [0 [Oi 03 [360 Reduction 99% | 99% | 100% | To0% | 7% | 100% ‘Some comments regarding the system are as follow: © The NF flux varies considerably. It appears that an average flux of 30 Lmh is reasonable at 50°C and 12 bar. These operating conditions are gentle, and the prob- ability of mechanical membrane damage is quite small. It is most likely that flux stability can be improved by increasing the cross flow. This is one of the parameters ‘that will be investigated on the full-scale system. ‘© Cleaning every 3 to 4 days is recommended in order to prevent the flux from240 ‘Membrane Technology in Water and Wastewater Treatment decreasing too much. ‘© The concentration ratio is close to 4, meaning that 75% of the feed becomes permeate while the remaining 25% are discharged as concentrate. ‘The NF shall enable the factory to recycle much more water and therefore consume less fresh water. The bigger the NF system, the more water ean be recycled. The full- scale plant will be used to investigate the relationship between water recycling and the impact on the water composition inside the paper machine, UF will make it possible to reduce fresh water consumption from 10m’ per (on pulp to Sm. NF will make it possible to reduce fresh water consumption to 2m* per ton pulp possible that is equal to zero discharge. Both are major achievements in water conservation. 2 FINAL REMARKS, Membrane technology is finally being used in the pulp and paper industry almost 30 years after the first very optimistic market studies predicted that membranes would be used on many different pulping liquids. UF of SSL and UF of bleaching effluent (E1) were the only processes that utilised membrane technology for two decades. It is interesting that 20 years passed before Esmil and Flootek realised new processes, while the known processes largely lay dormant. The author is confident that there are many more possiblities for membrane technology in this industry. However, it appears to require an unusual combination of people, membrane technology, pulp and paper companies, and, most importantly, lots of hard persistent work to achieve success, References 1 Finnemore, S. “Achieving Zero Output” and Private Conversations, Esmil Process Systems Lid, 1998. 2 Tepler, M., etal. P Raisio Flootek OY, 1998, 3 Tepler, Milan Private Conversations, Valmet Footek, 1997, 1998, 1999. 4 Wagner, J. Membrane Filtration Handbook. Wagner Publishing, Copenhagen, 1998,‘CASE STUDIES OF WASTEWATER RE-USE FOR THE PETROCHEMICAL, POWER AND PAPER INDUSTRY Bruce Durham USF Memcor Vivendi Water Wirksworth DE4 4BG, UK, L INTRODUCTION {A revolution is taking place in the water industry. The environmental issues combined with ever more stringent health expectations and the Ofwat directives are driving these changes as our industry moves towards common carriage and a better deal for the customer. How closely will we follow the example of the Gas and Electricity industries into an open and highly competitive market? Water companies are already benchmarking ‘their utility and asset management business against the Gas and Electricity suppliers. Reuse of wastewater is actively being encouraged by this water revolution. But we need innovators to champion these beneficial solutions, agreement on who owns the ‘water, logical financial incentives to interest the budget holders and reuse quality guidelines to protect all concerned. The UK has been slow to take advantage of new ‘opportunity due to the priority set following privatisation and lack of funding in a highly regulated water industry. tis easy to understand why California, Australia, Spain, South Africa and Israel are ahead of the UK. It is not so easy to understand why the Netherlands, Belgium, France, Scandinavia and Germany are also more advanced in reuse projects than the UK. This is especially difficult to comprehend when East Anglia is an arid region and we even have water shortage problems in high rainfall areas in the UK due to limited catchment reservoirs and no opportunity to build more reservoirs. Best practice studies need to consider international track record not just the limited amount of work that has been completed in the UK. ‘This paper is intended to introduce and demonstrate the viability of large reuse projects in the UK based on long term international operation experience on reuse projects for petrochemical, power and the paper industry. 2 REUSE TECHNOLOGY Reverse osmosis is the key process used in the reuse of wastewater to remove the dissolved salts. RO membranes are designed to operate under steady-state conditions.20, Membrane Technology in Water and Wastewater Treatment Sources, such as wastewaters are susceptible to rapid changes in suspended or dissolved solids loadings and are complex waters to treat ‘An RO element is not designed to handle suspended solids, as the RO membrane retains the solids. This results in rapidly declining flow and shortens the membrane life. Continuous microfiltration (CMF) systems are being used in over 40 insiallations to protect RO systems. There are two key elements to the CMF process - the 0.2um pore size hollow fibre membranes (out-to-in filtration) and patented air backwash system, both of Which offer a number of benefits: ‘+ Very high feed solids can be accommodated with no effect on the system performance ‘other than an automatic change in the frequency of the backwash. A consistent filtrate flowrate and quality is continuously sustained, with an automatic process, irespective of the quality of the feed water (Silt Density Index (SDI)<3, ‘Tusbidity<0.1 NTU) + Energy consumption is very low - typically 0.15-0.3 kWh/m? of treated water. Even ‘with several hundred mg/l of feed suspended solids, eross-flow isnot required. ‘+ The compressed air backwashing is much more effective than the conventional liquid backwashing at removing contaminants due to high energy and good distribution, It also offers the added benefits of discouraging grow-through and avoids the need for prechlorination of the backwash liquor, which is normally a prerequisite of conventional liquid backwashing membrane systems. Lime & Polymer Chlorine i Multimedia Clarification filtration Rapid mix flocculation t J Reverse ‘Carbon Backwash ene Dioxide to secondary presteatment Figure 1 “Conventional” RO Pretreatment Secondary fluent Backwash to secondary re-treatment a Microfiration Reverse Figure 2 Continuous Microfiliration RO Pretreatment ‘The following sections review three case studies.Industral Applications 28 3 WEST BASIN WATER DISTRICT, CALIFORNIA. The confidence in CMF/RO technology developed through the work done at Water Factory 2, 1] led tothe two West Basin wastewater reuse projec. ‘Over-abstraction has depleted the aquifer, causing saline ingress. In addition, industrial activity has caused further pollution making the aquifer unusable. There is complete dependence on expensive imported water. Secondary sewage from the Hyperion works is now used in a variety of ways to alleviate this. In one application, commissioned in 1997, Memcor processes are used in 15 Mlitre/day CMF/RO system to supply the nearby Mobil and Chevron refineries with feed for thei boiler water plants. In an earlier project Memcor supplied an 11.5 Mlitreiday CMF plant as pretreatment for RO to provide water used for injection in’ barrier scheme to keep out further sea water. Figure3. CMF & RO system at West Basin ‘CMF has simplified the wastewater reclamation process by eliminating the lime handling, addition and recovery systems and the flocculation, clarification, recarbonation and filtration processes, resulting in lower chemical requirements and eliminating the production and disposal of solid waste sludge. 3.1 Membrane Life The frequency and cost of replacing the membranes is of key importance to the viability of project, especially when membrane cost is typically 15% to 25% of capital cost. The CMF system at West Basin, which contains $ units, automatically checks the integrity of 9 million membrane fibres as part of the operating procedure. Two sub- ‘modules have been repaired inthe lat two years of operation. ‘The following capital and operating costs comparison forthe conventional and ‘microfiltration pretreatment to RO were presented in March 1999 at the AWWA membrane Conference. 2} ‘The operation costs for the two pretreatment systems were calculated from July 1997 to October 1998,24 Membrane Technology in Water and Wastewater Treatment Table1 Cost Summary ——Beseripton ‘Conventional Mierofiion Pretreatment © ® © © __ £ eM £ ML. [Fixed Coase Capital Coss ssoasi| 34 276720} 78 O&M Labour tras | 28 302] 4 Replacement Parts & Supplies zues} 3 9260/ d Suboel Fixed Costs $06,172) us 37s3si| 106 Variable Costs | Chemical Costs asossi| 54 ee] 7 ‘Sludge Production & Handing s1ose7| 34 7960] 2 Power isis} sagas} 1 Subtotal Variable Costs 923964] 101 nose} 32 ‘Total Fixed & Variable Costs 216 | ss Table2 Quality Summary Feed Turbidity (NTU) Fite Turbidity Comparison on spr Table3 Process Design Scope ‘Conventional ‘iierotitraton CAPITAL, ‘Clarifier ‘Strainers “Timea Filtration Micrfiation Cartridge Filters Cartridge Fitters RO RO VARIABLE ime Sodium Hypochlorite | Ferre Chloride ‘Sulphuric Acid Sodium Hypochlorite | Scale Inibition ‘Sulphur Acid Seale InhibitionIndusrial Applications 245 4 ERARING POWER STATION, AUSTRALIA This is a municipaindustrial re-use project in which secondary treated effluent from a local Sewage Treatment Plant undergoes tertiary treatment at Eraring Power Station in ‘New South Wales. The plant design combines Memcor continuous microfiltration (CMF) and reverse osmosis (cellulose acetate membranes) to produce high purity water for all power station uses except for drinking and showering. The power station i located onthe shore of Lake Macquari,a tidal saline lake near the coast and approximately one hour from Sydney. The project consisted of connecting the outlet from the Dora Creek sewage treatment works to a new membrane treatment system incorporating microfitration and reverse osmosis. The treated sewage is then used instead ‘of potable water for boiler feed, cooling water, dust suppression and fly ash handing. 4.1 Process Design = This 2-stage membrane system has been in operation since March 1995 producing 63m with a planned expansion to 168m°/hr. Effluent from Dora Creek ‘Sewage Treatment plant flows under gravity to centrifugal pumps, which deliver the feedwater via a single in-line motorised 500 micron self cleaning strainer to two Memcor CMF units Figure4 CMF System ‘Virtually all suspended solids, faecal coliforms and giardia cysts are removed. The SDI of the CMF filtrate water is consistently reduced to 1.5 allowing RO membranes to be ‘operated at approximately 40% higher flux. than is possible with a pretreatment based on ‘the conventional lime coagulation/sedimentation filtration process. Filtrate from the CMF units is dosed with sodium hypochlorite, for downstream control of biological growth. Sulphuric acid is also dosed to reduce pH and minimise hydrolysis of the cellulose acetate RO membranes. Microfiltered water is drawn fom the storage tank, dosed with antiscalant and passed through a 5 micron disposable cartridge guard filter for feed to the RO plant. ‘The RO system comprises two trains (2 x 50%), each with two stages ~ the first stage comprising 6 RO pressure vessels and the second stage comprising 3 (6:3 array). Each pressure ‘vessel houses seven RO membrane elements. The membrane elements themselves are 8.5-i diameter x 40 inches long cellulose acetate ‘membranes rated at 98% salt rejection. The RO ‘membrane, allowing only water to pass through, rejects salts and organics. Permeate (treated ‘water) which s virtually fee ofallatsand Figure RO System imiero-oganisms is piped to a degaser tower ‘to increase pH by flashing off CO; and is then stored in a 60m” treated water tank.246 Membrane Technology in Water and Wastewater Treatment The system provides the following benefits: 1) The treated water (permeate) from the RO plant has a TDS of < 40 mg/l when treating secondary sewage of 500-1500 mg/l TDS . The existing deioniser capacity has been increased by more than 3 fold as a result of the membrane pretreatment resulting in over AS150,000 savings per annum in chemical costs. 2) The membrane system removes bacteria and viruses for health risk minimisation and to protect the downstream equipment from biofouling. 3) RO permeate can be used directly for washdown, dust suppression and gland seal, with no health risks. 4) The compact size ofthe installation enabled an existing building to be used. 5) The total system is fully automatic, operated through the SCADA system and has self diagnostic data logging. 6) Membrane integrity tests automaticaly take place daily. Clean in Place (CIP) frequency is determined automatically by the control system. All cleaning chemicals are recovered and reused automatically on the microfiltration system. 7) Alleflluents from the system are recycled or used for dust suppression. 5 ZERO DISCHARGE PAI MILL, NEW MEXICO, USA. ‘One of the best known “Zero Effluent” Paper mills is MeKinley Paper Company's mill in New Mexico, USA, producing test liner from OCC (old corrugated containes). USF ied the water reclamation plant project with Continuous Microfitration (CMF), RO and with Crystallization (USF HPD S.A.). ‘The mill has operated since 1994 and it was originally designed for production of 135 00 ton/year of 100 % recycled fiber based board: Today the production is approximately 165 000 ton/year.The raw water consumption is approximately 1.5 m’fon of product. ‘This raw water is reverse osmosis treated deep well water from the Plains Electric Generation and Transmission company next door. There is no foul sewer near the mill in order to discharge the effluent and the raw water availabilty is restricted. These matters ‘were the original driving forces for the “Zero Effluent” concept. ‘The effluent from board production and OCC handling is treated biologically in an SBR (Sequenced Batch Reactor) type of activated sludge treatment plant. The biologically treated effluent is aerated and clarified in the same basin, in sequences. During clarification the aeration is simply turned off and the sludge is settled to the bottom of the basin. Excess sludge is removed from the bottom for dewatering and the clear effluent is removed from the surface of the basin. The water flow to the biological treatment is approximately 1500 m3/d. Typical COD content of the water is 3000-5500 mg/l in the inlet, TSS approx. 250 mg/l. COD in the outlet is 450-500 mg/l and TSS 40-50 mel. The effluent is cooled to suitable level (< 37 °C) before the SBR treatment in order to optimize the biological activity.Industral Applications ar ‘5.1 Membranes After Biological Treatment ‘The biologically treated effluent contains approximately 40-50 mg/l suspended solids of ‘which is removed by the CMF microfiltration system. ‘The CMF treated water has TSS <1 mg/l and SDI <3, which is very suitable for reverse osmosis treatment, in order to remove salts. Reverse osmosis treatment in McKinley is for approx. 660 m3/d, with 75% recovery, resulting in approx. 1000 m3/d ‘water recycled after microfiltration. The CMF treated water is low in BOD and COD, low in TSS (< Img/, low in bacterial (10° CFU/ml). ‘An HPD Crystallizer evaporates the concentrate from the RO plant in order to turn the “last liquid form of effluent” to solid waste. All TDS left over from biological, CMF and RO treatment is concentrated to RO brine. The amount is approx. 168 m3/d with TDS concentration as high as 8000 mg/.The Crystllizer is a tube type of falling film evaporator where the driving force for evaporation is generated by compressing the vapor pressure higher and simultaneously heightening the vapor temperature with mechanical single stage centrifugal vapor compressor. The condensate from the crystalliser is the cleanest treated water fraction from the whole water reclamation plant and it is re- circulated back to the proces ‘The final solid wast, practically 80 % NaySOy salt crystals, is disposed with the rest of the solid wastes by ‘in the multi fuel boiler in the coal mine next door. 6 CONCLUSIONS 1. Continuous Microfiltration allows Reverse Osmosis technology to treat previously impractical source waters, as microfiltration allows RO feedwater quality to be controlled and consistent 2. CMF produces a constant high quality effluent that is independent of feed water variations. 3. Microfiltration allows the membrane inventory of an RO plant to be reduced by up t0 30 t0.40%. 4. Microfiltration membranes have simplified the RO pretreatment process resulting in lowering the operating costs of both the pretreatment and RO systems. 5. CMF is proven to be more economical on whole life costs and is helping to drive down the cost of ownership of RO installations. 6. CMF allows TFC RO membranes to operate in a stable condition on municipal ‘wastewater. References [1] GL Leslie etal. , 1997. Pilot testing of microfiltration and ultrafiltration upstream of reverse osmosis during reclamation of municipal wastewater. [2] W.Won et al, 1999 Comparative life cycle costs for operation of full scale conventional pretreatment/RO and MF/RO system, AWWA Membrane Conference, Long Beach, March 1999, B] G. Craig, Overview of the long term application of wastewater reuse at a major powerstation. April 1999 [4] T Pobjolainen, 1999 Case study of ‘Closed’ board mill and application studies in Finland.PostersPRACTICAL EXPERIENCE WITH A MEMBRANE BIOREACTOR FOR WASTEWATER TREATMENT-SEMI-CROSS-FLOW ULTRAFILTRATION S. Geifler, K. Vossenkaul, Th Melin _P. Ohle, E, Brands, M. Dohmann Institut fir Verfahrenstechnik, Institut fir Siedlungswasserwirtschaf, der RWTH Aachen, der RWTH Aachen, ‘Turmstr, 46, Mies-van-der-Rohe St, 52056 Aachen 52056 Aachen Germany Germany 1 INTRODUCTION Currently, the use of ultra and microiltration for retention of activated sludge is receiving increasing attention. Using ultr- and microfilation has the advantage of ‘optimising the discharge hygiene. Further advantages are the low sludge generation and decreased tank volume, The challenges of using ultra- and microfiltration in the municipal wastewater treatment technology are to reduce investment and operating expenses and to cope with inflow variations. The development of new methods of ‘operation is necessary to fulfil these requirements. 2 PILOT PLANT The pilot plant concept consists of biological stage with nitrification and denitrification as well as a membrane plant with two module units. The bigger consists of five cushion modules with approximately 40 m? in total. This module unit ’s supposed to guarantee the operation of the biological stage. One single module is running in parallel to the first unit to optimise the operating parameters of the membrane plant and membrane materials. Using external modules it is possible to run the plant more flexibly than with submerged modules. The vertical installation of the membrane modules allows an air sparging from the bottom into the modules. This air is used to induce shear forces, hereby reducing cake formation to achieve high filtrate fluxes. To generate the necessary driving permeate sided partial vacuum is applied. 3 RESULTS During the pilot tests the operational procedure of the plant is gradually adjusted to realistic conditions of local sewage plants, i, e. variations in inflow conditions, Excellent quality effluent with potential discharge to a water course = Ability to handle difficult wastes = Enclosed plant, free from odours, => Minimal excess sludge production. = Water recycle. = Small foot print. The paper will give a general overview of different MBR techniques currently being used in Europe and also explore how the advantages of MBR can be retrofitted onto existing conventional activated sludge plants, The MBR process can be categorised into two main ‘+ The submerged technique where membranes are inserted into the BioReactor vessel. Generally used for high flow rates and resulting in low flux ratesMembrane Technology in Water and Wastewater Treatment ‘© Conventional cross flow UF technique with high flow velocities and flux rates. Very much suited to high COD strength waste streams, Extensive UK trials in the Food industry have been undertaken using a dedicated industrial base pilot plant and results will also be presented in the paper. ‘The MBR process in Europe has been developed by Wehrle Werk AG, Emmendingen, Germany and they have 10 years of MBR experience, covering 40 installations. Stork Protech UK and Wehrle have combined forces and operate in the UK as a joint venture.OPERATION OF A ZERO DISCHARGE WOOD PULP EFFLUENT TREATMENT PLANT GBateman Esmil Process Systems Ltd Westfields London Road High Wycombe HPII VHA Conventional processes for the treatment of effluent generated during the manufacture of medium density fibreboard (MDF) cannot comply with increasingly stringent environmental legislation. Advances in polymer chemistry combined with membrane ‘technology have enabled the recovery of both water and soluble organic compounds. This has resulted in the realisation of a ZERO DISCHARGE process Benefits of the membrane system include: Contaminated effluent to pure water in one process stage, soluble product recovery ~ dissolved metals, organics and inorganics, robust process - not susceptible to thermal and toxic shocks Following the success of the world’s first commercialised system at an MDF mill in Wales, a further seven plants have been installed across Europe, making this process something of an industry standard. The operational experience gained from these plants hhas placed Esmil in a strong position to offer expertise in the field of water/product recovery and reuse Process Philosophy ~ Wood pulping effluent from the refining process is flocculated, then pumped to a filter press, which produces @ filter cake with a dry solids content in excess of 50%. Filtrate passes through # multi-media filter. The filtered effluent is then {fed into the RO membrane and finally an optional carbon filter. Concentrate from the RO ‘membrane is stored in a dedicated holding tank, prior to recycling. All dirty backwash / cleaning waters from the filter press, multi-media filter and RO membrane system are recovered and retumed to the head of the plant. All solid and liquid phase outputs are recoverable, thereby resulting in a zero discharge plant. The Esmil Plant is substantially ‘automated and an operator can be fully trained within a matter of days. ‘The plant is largely mechanical with few process stages, reducing the capital investment requirement compared to other systems such as biological treatment. This process is currently being considered as a case study under the DTI's Environmental Technology Best Practice Programme and has already this year won Esmil Process Systems the Queen's Award for Environmental Achievement,INTEGRATION OF MAINTENANCE AND OPERATION INTO THE DESIGN OF REVERSE OSMOSIS MEMBRANE NETWORKS H, J. See, V, S. Vassiliadis and D. 1. Wilson ‘Department of Chemical Engineering, University of Cambridge, Pembroke Street, ‘New Museums Site, ‘Cambridge, CB2 3RA, United Kingdom, The selection of the configuration of a reverse osmosis (RO) membrane network is an important part of the design process, as the configuration dictates the flexibility of the system to variation in feed parameters and the loss of efficiency caused by fouling. Design methods have usually paid litle attention to maintenance and control issues presented by membrane fouling and deterioration owing to the difficulty in determining the actual operating cost of system subject to serious fouling and consequent cleaning. ‘A common practice to account for fouling consideration in the design process is to assume cleaning-in-place (CIP) operations based on pre-determined regeneration ‘maintenance cycles. However, these cycles are empirically developed without exact understanding of effects of fouling, nor consider the implications of downtime and imperfect cleaning. ‘This may result in higher capital costs if the fouling effect is overestimated, and higher operating costs due to frequent regeneration ifthe fouling effect is underestimated. The cleaning costs are strongly affected by the cleaning schedule uscd, so that complete optimisation of the design requires consideration of the impact of the network design on the cleaning schedule. Furthermore, replacement scheduled in advance has been acknowledged as the most effective way to control costs and maintain plant capacity consistency within design specs. This paper builds on a recently developed method for determining the optimal cleaning schedule in a given RO network subject to fouling. The scheduling problem is formulated as a mixed-integer nonlinear programming (MINLP) problem that can include control ‘action and membrane degradation. The problem is solved based on algebraic equations and time discretisation methodology under the general algebraic modelling system (GAMS) environment using the Outer-Approximation/Extended-Relaxation algorithm. Based on the algorithm, the objective function (which account for factors related to membrane regeneration and operating constrains) is minimised sequentially as a nonlinear programming (NLP) subproblem using MINOS and as a mixed-integer linear ‘programming (MILP) master problem using CPLEX. This flexible scheduling solver has been integrated into the larger design optimisation problem and the integrated problem solved using MINLP techniques. The size of the problem requires the use of approximate solutions to the design equations in the initial stage; these approximations are replaced by more exact forms in searching for the ‘optimum configuration. The paper presents some initial results from this combinedPosters 259 approach, using a previously published case study featuring a less flexible method. ‘The sensitivities of the discrete time intervals and the fixed cost incurred on each cleaning period regardless of module quantities to be cleaned are also discussed.MICROFILTRATION AND REVERS! & OSMOSIS OF KNOSTROP FINAL EFFLUENT M. Barton Yorkshire Water Services, Western Way, Halifax Road, Bradford, BD62LZ, Following the water shortages of recent summers, in particular 1995, Yorkshire Water wished © investigate potential new sources of water. Treating sewage effluent could potentially be lower cost than constructing a new raw water storage reservoir. A. second reason for carrying out this work was to maintain competitive advantage and to offer Yorkshire Water's industrial customers a choice. An on-site treatment plant which ean treat industrial effluent to a sufficiently good quality to enable it to be re-used as process water isan altemative to the provision of potable water and sewerage services. Treated sewage effluent could also potentially be used as on-site grey water, for polymer make-up or screenings washing. ‘The process chosen for this trial was continuous microfltration (CMF) followed by reverse osmosis (RO). “The aim of the trial was to treat Knostrop STW effluent to potable water quality, establishing capital and operating costs and evaluating operational problems and risks associated with the process. “The programme of work involved maximising the flow through the pilot plant optimising chemical usage, and challenging the plant with eryptosporidium and a selection of organic chemicals. The CMF unit achieved a maximum throughpot of 3.5 m’/hr and the RO plant 2.5 mf. Several operational problems were experienced during the trial, including biofouling, iron fouling and problems with hypochlorite dosing. The quality ofthe effluent produced was excellent and has been described as “better than drinking water" and “poor quality disilled water”. The eryptosporidium challenge achieved removal rates of greater than 4 Jog. ‘The trial of CMF and RO has proven that this technology is appropriate for the production of potable quality water from sewage effluent.MODELLING TEMPERATURE AND CONCENTRATION POLARISATION IN ULTRAFILTRATION OF NON-NEWTONIAN FLUID UNDER NON-ISOTHERMAL CONDITIONS Sergei P. AGASHICHEV Department of Chemical Engineering , D, Mendeleev University of Chemical Technology, Muisskaja Sa. 9, ‘Moscow 125047, Russia Development of new generation of membrane processes, in particular, in biotechnology and oil processing, requires development of advanced and more complex methods of calculation where phenomena of temperature and concentration polarization have to be described, Proposed method is based on hybridization of submodels for baro- and thermo- ‘membrane operations. Submited paper covers the following aspects: (A) temperature field; (B) concentration field; (C) non-Newtonian behaviour of fluid. The model applicable for channels with plat-and-trame configuration. The proposed solution is ‘based on the following assumptions: (1) flow is to be incompressible, laminar, continuous with uniform density field under the steady-state conditions; (2)fluid manifests non- Newtonian behaviour, viz.: apparent viscosity-1(z) is a power function of shear rate The ‘manuscript covers pseudoplastic and dilatant fluids; (3) fluid properties are independent of applied pressure and time under shear; (4) velocity profile being derived through integrating the momentum balance equation for non-Newtonian fluid has the following mathematical formulation 142) Cart -(-nyr"}; (S)Temperature profile was approximated by parabolic function (((¢)—hu)/(—tu)=1-22, where dimensionless coordinate with scale unit being equal to temperature layer thickness; The ratio of temperature (6;) to hydrodynamic (8y) boundary layer was estimated s8,/8y =Pr!; (6) Concentration profile was approximated by parabolic function (€0)-G)/(Gu-G) =07, where 0-dimensionless coordinate with scale unt being equal {0 concentration layer thickness. The ration of concentration to hydrodynamic boundary layer was estimated as, /8y = Sc"; (7) No slip at the wall; Equations for temperature and concentration boundary layers are the core of the model Integration ofthese equations over control volume has been carried out. Proposed model permits the analysis ofthe influence of mass flow, bulk concentration, tangential velocity, transmembrane flux , membrane rejection and geometry parameter on configuration of ‘concentration and temperature profile. The proposed solution can be built into complex algorithm of membrane process. The solution does not contain digital integration. The Paper includes the developed algorithm, an array of calculated data and graphical visualization of solution,NOVEL METHODS OF HOLLOW FIBRE MEMBRANE INTEGRITY MONITORING Dr. S. Williams Dr. AJ. Merry and C.V. Meadaweroft ‘Thames Water Research & Technology, _Leopold-PCI Membranes, Spencer House, Manor Farm Road, Reading, RG2 JW, Hampshire, RG27 8NR, UK UK ABSTRACT The recent widespread use of similar membranes for water treatment in Europe and North America has been largely driven by a need for improved disinfection, in particular the requirement to deal with Cryptosporidium. UltaBar is a novel hollow fibre ‘membrane system, of which its main applications are in the single stage treatment of ground and surface waters to potable standards. As the system performs as a physical barrier, the most important factor is to ensure the integrity of the system. Its essential to confirm that the barrier provided by the membranes and components within the system is intact. Water treatment other than through an intact membrane system will allow the passage of potentially harmful organisms into supply. In the case of Cryptosporidium, only a very small number of organisms may be sufficient to cause infection. Such integrity breaches may take the form of membrane fibre breakages, pin hole imperfections, potting defects or system O-ring and seal defects. In developing an integrated integrity monitoring system for the UltraBar process, a umber of methods are currently being trialed at pilot scale. These include physical ‘methods such as a water displacement test, a vacuum test and an air pressure test and a continuous online method using particle counting for an indication of permeate quality. ‘The water displacement test is based upon the theory that below a given applied pressure, the rate of air passage from the feed side to the permeate side of an intact ‘membrane will be a very small and due mainly to diffusion. However, in the ease of an integrity breach, the rate of air passage across the membrane is much higher. For the test, the feed side of the membrane elements are drained and the permeate side is left fully flooded with water. Air pressure to approximately 0.6 bar is applied to the feed side of the membrane, and the passage of air across the membrane results inthe displacement of ‘water from the permeate side. By measuring the rate at which this water is displaced, the integrity of the membrane is monitored. This method of integrity testing is curently in its developmental stages and is undergoing onsite pilot testing.Posters 268 ‘The intial results obtained on this test were generated using a single 40” XFlow clement, The results were then applied to a full scale membrane stack consisting of 84 x 40” membrane elements. in order for the test to be sensitive enough to detect a single broken fibre within an entire membrane stack of 56 element, it was calculated that the rate of water displacement due to air flow through a broken fibre on an compromised element would need to be approximately 100 times greater than the rate of water displacement duc to air diffusion in each ofthe 83 intact elements. The results obtained ‘on a single 40” clement indicated that at an applied pressure of 0.6 bar, the measured R values ranged from 300 to 980. These results indicate that the sensitivity of the water displacement test is high enough to enable the detection ofa single broken hollow fibre in over 800,000 intact fibres. Testwork carried out at a Thames Water site using a pilot plant containing 16 x 60” XFlow elements has produced comparable results. ‘The vacuum test is carried out only on elements and complete pressure vessel assemblies which have been removed ffom production and fully drained. The test is ‘based upon similar theory to the water displacement test in as far that when a vacuum of 0.2 bar absolute is applied to the permeate side of a fully drained and intact membrane clement, the rate of decay in the vacuum will be very slow and due only to the passage of air actoss the membrane from diffusion, Where an integrity breach is present however, ‘any vacuum applied to the permeate side will rapidly decay due to the influx of air through a broken fibre from the open feed side of the membrane. The mean rate of vacuum decay measured ranged from 4.0 mBar per minute on an intact membraive

You might also like

• ZwitterCo - Introduction Year 2023

  

  No ratings yet

  ZwitterCo - Introduction Year 2023

  20 pages
• Microbes and Enzymes For Water Treatment and Remediation - Sanet.st

  

  No ratings yet

  Microbes and Enzymes For Water Treatment and Remediation - Sanet.st

  227 pages
• Membrane and Separation Technology News

  

  100% (2)

  Membrane and Separation Technology News

  16 pages
• Water Management Concept

  

  No ratings yet

  Water Management Concept

  23 pages
• A Quantitative Comparison Between Chemical Dosing and Electrocoagulation

  

  No ratings yet

  A Quantitative Comparison Between Chemical Dosing and Electrocoagulation

  16 pages
• Vacumm Desalinization

  

  No ratings yet

  Vacumm Desalinization

  5 pages
• W&i Catalogue 2023 en - Low Res

  

  No ratings yet

  W&i Catalogue 2023 en - Low Res

  184 pages
• Pulsion MBR System: Low-Energy Submerged Hollow Fiber Ultrafiltration

  

  No ratings yet

  Pulsion MBR System: Low-Energy Submerged Hollow Fiber Ultrafiltration

  4 pages
• IC1172

  

  No ratings yet

  IC1172

  1 page
• Lecture # 2 - Membrane Classification

  

  No ratings yet

  Lecture # 2 - Membrane Classification

  21 pages
• Wastewater Treatment Using Membrane Technology

  

  No ratings yet

  Wastewater Treatment Using Membrane Technology

  13 pages
• Chapter 2. Membrane Technologies For Industrial Wastewaters Treatment

  

  No ratings yet

  Chapter 2. Membrane Technologies For Industrial Wastewaters Treatment

  40 pages
• D399D

  

  100% (1)

  D399D

  5 pages
• Qatar Desalination Research

  

  No ratings yet

  Qatar Desalination Research

  21 pages
• Product Safety Data Sheet Huwa San TR - 50 3

  

  No ratings yet

  Product Safety Data Sheet Huwa San TR - 50 3

  9 pages
• Progress in Materials Science: S. Ramakumar, C. Deviannapoorani, L. Dhivya, Lakshmi S. Shankar, Ramaswamy Murugan

  

  100% (1)

  Progress in Materials Science: S. Ramakumar, C. Deviannapoorani, L. Dhivya, Lakshmi S. Shankar, Ramaswamy Murugan

  87 pages
• Haake Viscotester Iq Instruction Manual - Reometro

  

  No ratings yet

  Haake Viscotester Iq Instruction Manual - Reometro

  130 pages
• 11 Chapter 11 Updated

  

  No ratings yet

  11 Chapter 11 Updated

  33 pages
• Water Treatment Through Membrane

  

  No ratings yet

  Water Treatment Through Membrane

  7 pages
• Full-Scale Wastewater Treatment Plant

  

  No ratings yet

  Full-Scale Wastewater Treatment Plant

  48 pages
• Avista Membrane Chemicals: All Avista Membrane Chemicals Are Fully Supported by Avista Services and Programs, Including

  

  No ratings yet

  Avista Membrane Chemicals: All Avista Membrane Chemicals Are Fully Supported by Avista Services and Programs, Including

  8 pages
• A Review On Industrial Wastewater Treatment Via Electrocoagulation Processes

  

  No ratings yet

  A Review On Industrial Wastewater Treatment Via Electrocoagulation Processes

  32 pages
• NOTES Unit 1 Water Technology 23022021

  

  No ratings yet

  NOTES Unit 1 Water Technology 23022021

  32 pages
• WTP Moduls 2 - Agst2018

  

  No ratings yet

  WTP Moduls 2 - Agst2018

  305 pages
• USTR3369 Fluorodyne II DBL Capsules Cartridges VG EN

  

  No ratings yet

  USTR3369 Fluorodyne II DBL Capsules Cartridges VG EN

  21 pages
• Pentair Structural Pressure Vessels L

  

  No ratings yet

  Pentair Structural Pressure Vessels L

  12 pages
• Nanofiltration and Reverse Osmosis

  

  No ratings yet

  Nanofiltration and Reverse Osmosis

  2 pages
• MB Regeneration

  

  No ratings yet

  MB Regeneration

  3 pages
• DOWEX Ion Exchange Resins: Tech Facts

  

  No ratings yet

  DOWEX Ion Exchange Resins: Tech Facts

  2 pages
• UnitOps 2019 4 Dewatering

  

  No ratings yet

  UnitOps 2019 4 Dewatering

  54 pages
• Presentation HuwaSan WT Legionella - Covid

  

  No ratings yet

  Presentation HuwaSan WT Legionella - Covid

  22 pages
• Gengard Gn7004: Description and Use

  

  No ratings yet

  Gengard Gn7004: Description and Use

  2 pages
• Unit 2 Advanced Biological Processes For Wastewater Treatment

  

  No ratings yet

  Unit 2 Advanced Biological Processes For Wastewater Treatment

  75 pages
• V 07 Hater Kurita

  

  100% (1)

  V 07 Hater Kurita

  25 pages
• Evaporative Condenser Passivation: Cameron Klein Strand Associates, Inc

  

  No ratings yet

  Evaporative Condenser Passivation: Cameron Klein Strand Associates, Inc

  27 pages
• State-Of-The-Art Reverse Osmosis Desalination PDF

  

  No ratings yet

  State-Of-The-Art Reverse Osmosis Desalination PDF

  77 pages
• Cross-Flow Microfiltration

  

  No ratings yet

  Cross-Flow Microfiltration

  8 pages
• Recommended Resins (ResinTech)

  

  No ratings yet

  Recommended Resins (ResinTech)

  1 page
• Annamox

  

  No ratings yet

  Annamox

  13 pages
• Electrochemical Technologies in Wastewater Treatment PDF

  

  No ratings yet

  Electrochemical Technologies in Wastewater Treatment PDF

  31 pages
• 450 Bittern Separation

  

  No ratings yet

  450 Bittern Separation

  5 pages
• Design and Manufacturing of Portable Lifesaving Water Filter

  

  No ratings yet

  Design and Manufacturing of Portable Lifesaving Water Filter

  20 pages
• Membrane Technology Lecture

  

  100% (1)

  Membrane Technology Lecture

  25 pages
• Spirac Grit Classifier SANDSEP - A4 - v1

  

  No ratings yet

  Spirac Grit Classifier SANDSEP - A4 - v1

  4 pages
• Wastewater Treatment: Secondary or Biological Processes

  

  No ratings yet

  Wastewater Treatment: Secondary or Biological Processes

  102 pages
• Membrane Technology in Water Treatment Applications

  

  No ratings yet

  Membrane Technology in Water Treatment Applications

  11 pages
• Screening Methods For Selection of Surfactants Formulations For IOR From Fractured Carbonate Reservoirs

  

  No ratings yet

  Screening Methods For Selection of Surfactants Formulations For IOR From Fractured Carbonate Reservoirs

  41 pages
• Water

  

  No ratings yet

  Water

  40 pages
• Water Treatment Water Treatment: Lecture 6: Softening

  

  No ratings yet

  Water Treatment Water Treatment: Lecture 6: Softening

  29 pages
• Kla Qa Optimize Aeration MBR Plants

  

  No ratings yet

  Kla Qa Optimize Aeration MBR Plants

  2 pages
• Chapter 3 Primary Treatment

  

  No ratings yet

  Chapter 3 Primary Treatment

  29 pages
• Recent Advances in Desalination Technologies Final 1

  

  No ratings yet

  Recent Advances in Desalination Technologies Final 1

  190 pages
• W3T556068 - BA - Protegra CS PRO RO EDI 125-250 - EN - DE - 01 - 04.2023

  

  No ratings yet

  W3T556068 - BA - Protegra CS PRO RO EDI 125-250 - EN - DE - 01 - 04.2023

  104 pages
• What Is The Difference Between Demineralized Water and Distilled Water

  

  No ratings yet

  What Is The Difference Between Demineralized Water and Distilled Water

  2 pages
• 820r11003 (DW Residuals Management Report)

  

  No ratings yet

  820r11003 (DW Residuals Management Report)

  377 pages
• Water Treatment Plant Process

  

  No ratings yet

  Water Treatment Plant Process

  13 pages
• Water

  

  No ratings yet

  Water

  37 pages
• Water Treatment - Lectures

  

  No ratings yet

  Water Treatment - Lectures

  80 pages
• Salient Features of Nalgonda Technique

  

  100% (1)

  Salient Features of Nalgonda Technique

  10 pages
• 4th Semester Internship Report

  

  No ratings yet

  4th Semester Internship Report

  21 pages