DUAL MEMBRANE SYSTEMS IN SEAWATER ...watec.gr/wp-content/uploads/2017/06/RWL-DualMembrane...IDA Desalination Industry Action for Good / Santa Margherita, Portofino, Italy May 16-18,

IDA Desalination Industry Action for Good / Santa Margherita, Portofino, Italy May 16-18, 2011REF: IDA/PORT2011-056

-1-

DUAL MEMBRANE SYSTEMS IN SEAWATER DESALINATION: DRIVERSFOR SELECTION AND FIELD EXPERIENCES

Authors: Eduard Gasia-Bruch, Markus Busch, Verónica García-Molina, Udo Kolbe

Presenter: Udo Kolbe, Commercial Development Leader Ultrafiltration, Dow Water and ProcessSolutions

AbstractTraditionally, Sea Water Reverse Osmosis (SWRO) installations have been operated with conventionalpretreatments based on single or two staged media filtration occasionally preceded bycoagulation/flocculation processes. In the last decade however, microfiltration and ultrafiltrationpretreatment have increasingly gained acceptance as viable pretreatment alternatives for seawaterdesalination.

This paper describes the commonly reported drivers for adopting membrane pretreatment techniques, suchas: greater capability to cope with fluctuations and high solid loads in raw waters, smaller footprint, higherenvironmental sustainability, lower SWRO stage cost both from design as well as from the operation side.Among the drivers, this report will give special emphasis on the benefits that membrane pretreatmenttechnology offers to the desalination industry to efficiently satisfy temporary and/or emergency fresh waterdemand, e.g. installing containerized and mobile units with compact and modular membrane pretreatmentschemes.

Despite the fact that many SWRO facilities include low pressure membrane pretreatment, quantifiableperformance data reported is still relatively scarce. This publication will report field experiences ofintegrated membrane systems using ultrafiltration (UF) membranes as pretreatment for reverse osmosis (RO)membranes in the sea water desalination field. Two examples of industrial scale integrated membranesystems operating with DOWTM Ultrafiltration technology and DOW FILMTECTM reverse osmosismembranes are the focus of the study.

One of the studied systems is the Moni Desalination plant, which is a mobile system using a completemembrane integrated scheme. The plant has been fully operational since the end of 2008 and supplies 20,000m3/day of drinking water to the city of Limassol and surroundings (Cyprus). The plant reliably achieves therequired permeate flow together with the required quality (Boron concentration lower than 1 mg/l). The plantpresents optimized cleaning procedures which minimize the chemical consumptions during ChemicalEnhanced Backwash (CEB). The average transmembrane pressure (TMP) ranges at around 0.86 bar. Theother system studied is a containerized integrated membrane system installed in a refinery in Taranto (Italy).This system has been operational since mid 2009. The plant produces 5,760 m3/day of SWRO permeatewhich is sent to a second pass of RO in order to produce boiler feed water. One of the key characteristics ofthis system is the successful operation of the system at a high recovery (60%) in the SWRO section requiringonly two Cleaning In Place (CIP) actions in 20 months of operation. This, could be partially, attributed to thehigh water quality provided by the ultrafiltration membranes which positively affect the sustainable highrecovery applied in the SWRO system.

The characteristics and the performance of these integrated membrane systems together with the operationalexperiences acquired during the time of operation will be discussed in detail in this publication.

TM Trademark of The Dow Chemical Company ("Dow") or an affiliated company of Dow


-2-

I. INTRODUCTION

Traditionally, Sea Water Reverse Osmosis (SWRO) plants have been operated with conventionalpretreatments based on single or two staged media filtration occasionally preceded bycoagulation/flocculation processes. In the last decade, microfiltration (MF) and ultrafiltration (UF)pretreatment have increasingly gained acceptance as viable pretreatment options for seawaterdesalination While in the period prior to 2002, mostly pilot studies had been carried out using membranetechnology as pretreatment, in recent years several large (> 100,000 m³/d) SWRO installations haveimplemented UF or MF pretreatment. There are already more than 40 large-scale desalination plantsusing low pressure membrane pretreatment, i.e., UF or MF [1]. Figure 1 shows the cumulative capacityof SWRO plants with UF pretreatment v.s. time, which gives a clear indication of the overall fastadoption speed of membrane pretreatment technology.

Figure 1: Cumulative capacity of SWRO plants with UF pretreatment [1]

Systems combining ultrafiltration or microfiltration and Reverse Osmosis (RO) are usually called dualmembrane systems or integrated membrane systems. The selection of ultrafiltration as reverse osmosispretreatment against conventional pretreatment schemes is driven by several considerations. The keydrivers are described in this report.

Despite the fact that many SWRO installations include low pressure membrane pretreatment, there isfew quantifiable performance data reported. Because of that, this publication will provide fieldexperiences of two industrial scale systems using ultrafiltration as pretreatment for the reverse osmosismembranes in the sea water desalination application. Both reported systems are operating with DOW™Ultrafiltration technology and DOW FILMTEC™ reverse osmosis membranes.

The DOW™ Ultrafiltration modules are made from high strength, hollow fiber membranes featuring thefollowing key characteristics [2]:

- Uniform 0.03 μm nominal pore diameter for removal of bacteria, viruses, and particulates including colloids.

- Polyvinylidene fluoride (PVDF) polymeric hollow fibers for high strength and chemical resistance.The hollow fiber membrane are 1.3 mm outside diameter and 0.7 mm inside diameter

- Hydrophilic PVDF fibers for easy cleaning and wettability that help maintain long termperformance

- Outside-In flow configuration for high tolerance to feed solids and the use of air scour cleaning

10,000

100,000

1,000,000

10,000,000

1998 2000 2002 2004 2006 2008 2010 2012

Cu

mu

lati

ve

ins

talle

dS

WR

Oca

pac

ity

wit

hU

Fp

retr

eat

me

nt,

[m³/

d]


-3-

- The pressurized vertical shell-and-tube design from Unplasticized Polyvinyl Chloride (U-PVC)eliminates the need for separate pressure vessels and allows easy removal of air from cleaningand integrity testing steps.

One of the field experiences with dual membrane systems reported in this paper is the MoniDesalination plant. The Moni plant is a mobile system using a complete membrane integrated scheme.The plant has been fully operational since the end of 2008 and supplies 20,000 m3/day of drinking waterto the city of Limassol and surroundings (Cyprus). This study complements the previously publisheddata from the first 1 ½ years of operation of the system [3] and provides extended performance datafrom 10 recent months.

The other system studied in this publication is a containerized integrated membrane system installed in arefinery in Taranto (Italy). The plant produces 5,760 m3/day of SWRO permeate which is sent to asecond pass of RO in order to produce boiler feed water. A first part of the system of 3,460 m3/day wasstarted up mid 2009 and the second phase producing 2,300 m3/day of SWRO permeate was started-up inOctober 2010.

In this publication, the main characteristics of the core pretreatment and RO units will be describedtogether with the operational experiences since the start up.

II. ULTRAFILTRATION SELECTION DRIVERS

The key drivers that have been quoted frequently include the greater capability to cope with fluctuationsand high solid loads in raw waters, smaller footprint, higher environmental sustainability, lower SWROstage cost (both from capital as well as from the operating expenses). Additionally, some authors alsohighlight that membrane filtration technology could be break even or even be the lower unit operationcost compared to conventional pretreatment schemes, especially the more extensive ones. Easier designand operation have also been reported as key drivers. An extensive analysis of the market needs andultrafiltration adoption drivers has been developed by Busch et al. [1].

This chapter describes some of the common drivers for adopting membrane pretreatment techniques andamong them special emphasis is given to the benefits that the membrane pretreatment technology offersto the desalination industry to respond efficiently to temporary and/or emergency fresh water needs, e.g.installing containerized and mobile units with compact and modular membrane pretreatment schemes.

2.1 Greater capability to cope with difficult waters – Primary pretreatment stage design

Ultrafiltration has a greater capability to cope with high solid loads in raw waters while maintaining agood filtrate quality constant despite of feed water quality fluctuations. Difficult waters often haverequired a very extensive pretreatment chain, consisting of coagulation, flocculation, clarification andmedia filtration in often multiple stages, in order to provide acceptable water quality to the RO stage. Ithas been mentioned and demonstrated repeatedly that ultrafiltration could achieve acceptable waterquality without the use of extensive pretreatment [1].

An example from one of the field experiences reported in this report (see chapter 3.2.3) will describe anexperience which demonstrates this capability of the ultrafiltration technology to work with high solidloads in the raw waters while maintaining a consistent good filtrate quality.


-4-

2.2 Low environmental impact

Higher chemical doses and sludge quantities resulting from larger required coagulant doses inconventional pretreatment systems sometimes cannot be tolerated environmentally. This trend isapparent in some Australia and California projects.

Usually it is mentioned that ultrafiltration pretreatment can provide reductions in chemical consumption(mainly coagulant, acid and polymer). The key argument is that coagulation (which in conventionalsystems is using at a level of typically at least about 1 up to various mg/L of Fe) can be reduced orpossibly eliminated altogether by the use of ultrafiltration. Busch [4] reported a study where around40 % of the evaluated pilot studies eliminated the use of coagulant. In most conventional systems withcoagulation, significant acid dosing (range of 20-30 mg/L of HCl or H2SO4) is used to bring the pH tothe optimum coagulation pH of approximately 7. In some systems, also polymers are used to enhancethe flocs before media filtration – this chemical could also be eliminated in a UF scheme.

The possibility to avoid coagulation results in a direct, and significant, reduction of sludge mass created(in terms of Total Suspended Solids), by a factor in the range of 2 or larger (depending on raw waterTSS as well as coagulant dose). In certain regions, sludge cannot be injected to the brine for discharge tothe sea. Sludge needs to be concentrated, and then disposed separately [5, 6]. Both the sludge treatmentas well as the disposal can represent significant cost items in plants that are forced to use this approach[6]. The reduction of sludge would be a significant cost saving enabled by membrane filtration.

2.3 Automation

Sometimes it is reported that ultrafiltration systems can be provided with a higher degree of automationthan a media filtration system [7]. Another argument is that in the case of UF, the operation and cleaningprotocols and programming are provided from a very experienced UF technology supplier, while in thecase of media filtration, both protocol and programming needs to be developed by the system integrator.

2.4 Spatial aspect – Containerized systems

UF systems usually require a smaller footprint than media filtration systems, especially where dual-stagemedia filters are combined with sedimentation or flotation. Site-specific conditions, building costs, andother limitations may affect footprint size.

Additionally, UF skids are smaller and more compact than conventional systems. Due to their morecompact, modular characteristics and lower weight, ultrafiltration technology is the preferred choice formobile units. There are multiple examples of mobile systems built and delivered in modular components(containers), or operated on barges which are moved to areas with highest water need. The benefits ofbuilding containerized systems is the minimization of civil works and related costs, mobility of thesystems, faster system installation, etc.

The Moni sea water desalination plant is a containerized dual membrane system which was installed toprovide potable water to the people of Limassol (Cyprus) for 3 years. The Moni desalination equipmentcan be removed to an alternate location or placed onto a barge at the end of the contract, thanks to thespecially designed containers and movable plant. The Moni plant project was carried out during 2008


-5-

period when Cyprus was undergoing a severe drought. Therefore a fast action was critical to maintainthe fresh water supply to the inhabitants of Limassol and surroundings. The facility was constructed andinstalled in less than eight months. There are no precedents in the global desalination industry of asuccessful completion of such a project in such few months. The unique mobility aspect of the solutionas well as the speed of deployment is seen as one of the few options available to Governments facingimmediate problems [8].

Taranto dual membrane system, which is described in section IV is also an example of a containerizedsystem. An ultrafiltration system in Madagascar to provide direct potable water using DOW™Ultrafiltraton technology also uses the containerized approach.

Another feasible alternative to be able to supply drinking water within a very limited time are thoseSWRO installed in barges. Examples of plants using ultrafiltration as SWRO pretreatment operated onbarges can be found in United Arabic Emirates [9] and Saudi Arabia [10]. Shoaiba Barge SWRO plant islocated off the shore of Saudi Arabia and represents the world’s largest sea-based desalination plantproducing 52,000 m3/day of permeate water [10].

2.5 Lower SWRO stage cost

A vast number of benefits have been described related to the SWRO stage of integrated systems. TheSWRO stage is the heart of the process and represents the highest complexity and the highest capitalcost of the system. The most reliable UF product quality enables high operating fluxes in the SWROtrains together with lower Cleaning In Place (CIP) protocols requirements. All of these potential benefitsobviously build on the better UF water quality, which is expected to smoothen operation of the SWROstage. Many authors [e.g. 11-13] have actually assumed higher cost of UF pretreatment (as opposed tomedia filters), but the overall process results more economically viable with UF due to the savings in theReverse Osmosis section.

The benefits that ultrafiltration can provide to the SWRO skids could be divided into two:

Lower SWRO Stage Capital Cost.The potential for downstream capital costs reduction is a key consideration and it considers thatthanks to the better quality of the RO feed provided by the UF the SWRO operation can be reliablyoperated at higher flux and/or higher recovery. This would immediately result in lower number ofRO elements needed to produce the requested amount of permeate (in the case of higher flux) and/orin a reduction of the RO feed water flow needed which result in a decrease of the pretreatment sizeneeded (in the case of higher flux). The capital savings in this case would also involve the lower costassociated with the number of vessels, pipes, valves and instrumentation. SWRO costs typicallyamount to a multiple of the pretreatment costs ranging from 3 to 10 [1].In one of the field experiences reported in this paper (Taranto system, see section IV) the SWROsystem is operating reliably at a very high recovery of around 60% using two stages in the seawaterpass with an intermediate booster.

Lower Cost SWRO Stage Operation.

UF may reduce energy costs, membrane replacement costs, and SWRO cleaning costs by deliveringwater with lower fouling tendency.


-6-

Recent cost estimations stated similar capital expenses related to a UF pretreatment and to a two stageconventional pretreatment (pressurized sand filtration) and 15% lower operating expenses in theinstallation with UF pretreatment [6]. In all these calculations however, the potential positive effects ofultrafiltration pretreatment on the operation of the reverse osmosis plant, i.e., lower fouling rates, lowercleaning frequencies and longer life times were not taken into consideration and such parameters wereassumed equal for both pretreatments. This then indicates, that in reality, the operating cost of thoseinstallations with conventional pretreatments are possibly larger than 15% compared to UF pretreatment.

2.6 Pretreatment stage economical aspects

In terms of the costs, 15 years ago the use of UF as a pretreatment for SWRO resulted not economicallyviable but since approximately year 2005, the cost of the UF has become more and more competitivecompared to conventional pretreatment. It has been stated that the current cost of a UF is half of its costback in 1998 [14]. Chu et al [15] have clearly dissected pretreatment and SWRO cost and shown howtechnology changes in UF have made UF systems cost comparable to media filtration pretreatments,even in absence of cost improvements in the SWRO section. The significant reduction of cost of UFpretreatment systems to an equivalent or possible even advantaged cost position of conventionalpretreatment systems is a significant turning point for the industry which is expected to even moreaccelerate adoption of ultrafiltration technology.

IDA Desalination Industry Action for Good / Santa Margherita, Portofino, Italy May 16

III. FIELD EXPERIENCES: MONI DESALINATION PLANT

This chapter provides a description of the main characteristics of thea detailed evaluation of the UF and ROcomplements the previously published data by Gasiaof the system and provides extended performance data of 10 recent months.

3.1. The Moni desalination plant

The Moni Desalination plant is located at the Moni power station outside Limassol (Cyprus). The plantuses the electricity generated from the powerof Limassol. The plant is owned by Subsea Infrastructure Ltd. and was constrIndustries Ltd. in 8 months, which represents the fastest execution of a large scale desalination plant todate. A local company, Silnir Cyprus, was formed by Subsea Infrastructure and Nirosoft Industries Ltdto manage the operations. The contract with the Cyprus Government is for a build, operate and removeplant for three years duration. The plant is mobile and in December 2011 it can be dismantled,transported and then reassembled in a new location within a three month period.plant uses ultrafiltration technology to pretreat sea water before feeding the reverse osmosis unit. Theplant is a fully integrated membrane system that uses DOW™ Ultrafiltration as a pretreatment for DOWFILMTEC™ reverse osmosis membran(UF skids, RO racks, high pressure pumps and energy recovery devices).December 2008. Since then it supplies more than 20,000 mMoni plant is innovative for its low chemical consumption scheme.pictures of the Moni desalination plant

DOW™ Ultrafiltration skid container

Figure 3. Photographs of

The Moni desalination plant location (Satellite pictureadapted from Google™ maps).

Figure 2. Satellite picture of Moni

IDA Desalination Industry Action for Good / Santa Margherita, Portofino, Italy May 16-REF: IDA/PORT2011-056

FIELD EXPERIENCES: MONI DESALINATION PLANT

a description of the main characteristics of the Moni Desalination plantUF and RO system performances of the 2 ¼ years of operation

complements the previously published data by Gasia-Bruch et al [3] from the first 1 ½ years of operationof the system and provides extended performance data of 10 recent months.

Moni desalination plant – description

s located at the Moni power station outside Limassol (Cyprus). The plantuses the electricity generated from the power station and provides fresh drinking water to the inhabitants

. The plant is owned by Subsea Infrastructure Ltd. and was constrIndustries Ltd. in 8 months, which represents the fastest execution of a large scale desalination plant to

local company, Silnir Cyprus, was formed by Subsea Infrastructure and Nirosoft Industries Ltdhe contract with the Cyprus Government is for a build, operate and remove

lant for three years duration. The plant is mobile and in December 2011 it can be dismantled,transported and then reassembled in a new location within a three month period. Theplant uses ultrafiltration technology to pretreat sea water before feeding the reverse osmosis unit. Theplant is a fully integrated membrane system that uses DOW™ Ultrafiltration as a pretreatment for DOWFILMTEC™ reverse osmosis membranes. The main parts of the desalination system are containerized

pressure pumps and energy recovery devices). The plant was started up inDecember 2008. Since then it supplies more than 20,000 m3/day of drinking water. Furthermore,Moni plant is innovative for its low chemical consumption scheme. In Figure 2pictures of the Moni desalination plant and location are presented:

DOW™ Ultrafiltration skid container and UF rack FIMTEC™ RO rack

Photographs of UF and RO racks of the Moni desalination plant [3]

The Moni desalination plant location (Satellite pictureThe Moni Desalination plant from the control room [

Satellite picture of Moni desalination plant location and photograph of the plant

-18, 2011

-7-

Moni Desalination plant as well asyears of operation. This study

] from the first 1 ½ years of operation

s located at the Moni power station outside Limassol (Cyprus). The plantprovides fresh drinking water to the inhabitants

. The plant is owned by Subsea Infrastructure Ltd. and was constructed with NirosoftIndustries Ltd. in 8 months, which represents the fastest execution of a large scale desalination plant to

local company, Silnir Cyprus, was formed by Subsea Infrastructure and Nirosoft Industries Ltdhe contract with the Cyprus Government is for a build, operate and remove

lant for three years duration. The plant is mobile and in December 2011 it can be dismantled,The Moni desalination

plant uses ultrafiltration technology to pretreat sea water before feeding the reverse osmosis unit. Theplant is a fully integrated membrane system that uses DOW™ Ultrafiltration as a pretreatment for DOW

he main parts of the desalination system are containerizedThe plant was started up in

/day of drinking water. Furthermore, the2 and Figure 3 some

FIMTEC™ RO rack

]

Desalination plant from the control room [8]

desalination plant location and photograph of the plant


-8-

Figure 4 shows a flow diagram of the desalination process done in the Moni desalination plant.

Figure 4. Process diagram from Moni desalination plant - adapted from [3]

The raw water source is an open seawater intake in 1,100 meters distance from the beach and 22 mdepth below the sea surface. Firstly, there is a coarse screen of 10 mm size on the seawater intake. Nocoagulation is used in the feed water. Before reaching the UF system the feed water passes through a setof self-cleaning filters of 500 m screen size followed by another set of 100 m filters. The UF unitconsists of 12 racks each containing 66 x DOW™ Ultrafiltration modules, a total of 792 DOW™Ultrafiltration modules at an operating flux of 60 L/m2h. The DOW™ Ultrafiltration module used is the2860 module type [16], featuring 51 m2 of active area, 8 inches (20.3 cm) of diameter and 60 inches(152.4 cm) length. Each UF rack is housed in a 40-feet container. The UF filtrate water is collected in atank which feeds the RO system. The RO system is configured as a single pass with a single stage. Thesystem is divided into three trains; each train has one pump container and three membrane containers.The pump container houses the feed supply pump, eight isobaric energy recovery devices, the boosterpump and the high pressure feed valve. The high pressure feed pump is located outside of the pumpcontainer. Each membrane container has 31 pressure vessels with 7 membranes elements each. Twotypes of DOW FILMTEC™ membrane elements are installed within the pressure vessels in anInternally Staged Design (ISD) configuration. Within each pressure vessel, 5 x DOW FILMTEC™SW30HRLE-400 elements are installed in the first positions of the vessel and 2 x DOW FILMTEC™SW30ULE-440i elements are installed in the last positions. The total number of installed membraneelements is 1,953. The warranted RO permeate water quality is defined by a boron content below 1mg/L and a Total Dissolved Solids (TDS) value below 360 mg/l.

The table below (Table 1) includes the main parameters of the UF and RO systems as well as their mainoperating conditions from the systems studied in the present work:

Table 1. Summary of the UF and RO systems main parameters in the Moni Desalination plant

UnitTotal

Capacity(m

3/h)

Recovery (%)Operating flux

(l/m2h)

Product installedTotal

number ofmodules

Ultrafiltration 1,900 90 60 DOW™ UF 2860 792

ReverseOsmosis

833 44 11.5In each pressure vessel:

5x DOW FILMTEC™ SW30HRLE-4002x DOW FILMTEC™ SW30ULE-440i

1,953

The main details of the average feed water quality are presented in Table 2.

Self-cleaning filters:

500 µm 100 µm

UF FiltrateWater Tank

RO PermeateWater Tank

Posttreatment

UF RO


-9-

Table 2. Feed water quality for the Moni desalination plant

Parameter Unit Feed water quality

Temperature °C 15 – 28

pH 8.2

RO feed conductivity S/cm 62,000

RO feed TDS conc. mg/l 44,100*

RO feed boron conc. mg/l 5.2

UF feed water SDI15 %/min 2.4 - 6.2 (5.2 in average)

Turbidity NTU 0.4 - 10 (1.8 in average)

*calculated TDS considering a conductivity / TDS factor of 1.4

Table 2 indicates a relatively high feed water salinity. This could be attributed to a high raw water TDScombined with a partial salinity increase of the raw water due to the mixing done by the energy recoverydevices. The water feeding the UF system has an averaged SDI15 value of 5.2 (from 155 measurements).Additionally, the UF feed turbidity varies from 0.4 – 10 NTU with an average value of 1.8 NTU (from209 measurements).

3.2. Dual membrane system performance

In this chapter the performances of the ultrafiltration and reverse osmosis systems during the 2 ¼ yearsof operation are evaluated.

3.2.1. Ultrafiltration system

The DOW™ UF modules in the Moni desalination plant operate at a flux of 55-65 l/h/m2 with aBackwash (BW) frequency of 75 minutes. The ultrafiltration system shows a low environmental impactdue to the very low chemical usage. The ultrafiltation system only requires chemical addition once every7 days when a Chemical Enhanced Backwash (CEB) is performed. Most remarkably, no CIP wasrequired since the start up in December 2008.

The recovery (related to product flow) and availability of the Moni plant were calculated to be in therange of 91% and 92% respectively. This results in an overall efficiency of approximately 84% and a netflux of approximately 50.5 L/h/m².

The permeability and the transmembrane pressure (TMP) represent the key parameters showing the flowperformance and the degree of fouling of the UF membranes.

The TMP is the difference between feed pressure and filtrate pressure. To evaluate properly the UFsystem performance, the TMP values are normalized at a temperature of 25 °C. The membranepermeability is the flux (l/h/m2) divided by the TMP (bar). The permeability is also normalized bytemperature (at 25 °C) calculating it from the normalized permeate flow and the normalized TMP. Thedetails of the formulas used to normalize the UF performance were described by Gasia-Bruch et al. in aprevious publication [3].


-10-

The normalized TMP (see Figure 5) and normalized permeability (Figure 6) are plotted below. Thesefigures show the evolution of the mentioned parameters by each individual UF rack and the UF systemaverage values (thicker black colored line).

Figure 5 Normalized TMP evolution for all individual UF racks and system average

Figure 6 Normalized permeability evolution for all individual UF racks and system average

One of the key actions carried out during the first months of operation was to vary the cleaningprocedures and frequencies in order to optimize the system operation, with the target of chemicalconsumption minimization. During the first months of operation various UF cleaning protocols weretried and therefore more fluctuation of the UF performance parameters was observed. Thanks to thiscontinuous improvement focus, a low cleaning consumption scheme was achieved.

After optimization of the cleaning procedures the TMP has been well controlled (see Figure 5) and thenormalized permeability has been increased (see Figure 6). The normalized TMP is in average 0.86 bar

0.0

0.5

1.0

1.5

2.0

2.5

3.0

1-Jan-09

2-Mar-09

1-May-09

30-Jun-09

29-Aug-09

28-Oct-09

27-Dec-09

25-Feb-10

26-Apr-10

25-Jun-10

24-Aug-10

23-Oct-10

22-Dec-10

20-Feb-11

Operating Time (Date)

No

rma

lize

dT

MP

(Ba

r)

UF1 UF2 UF3 UF4 UF5 UF6 UF 7 UF 8 UF 9 UF 10 UF 11 UF 12

0

50

100

150

200

250

300

1-Jan-09

2-Mar-09

1-May-09

30-Jun-09

29-Aug-09

28-Oct-09

27-Dec-09

25-Feb-10

26-Apr-10

25-Jun-10

24-Aug-10

23-Oct-10

22-Dec-10

20-Feb-11


No

rma

lize

dP

erm

eab

ilit

yL

MH

/Ba

r

UF1 UF2 UF3 UF4 UF5 UF6 UF 7 UF 8 UF 9 UF 10 UF 11 UF 12


-11-

and the normalized permeability of the system is in average 74 lmh/bar. The current normalized TMP isvery close to the designed initial TMP 0.6-0.7 bar. This indicates that the current BW and CEBprocedures are effective in removing the fouling and restoring the UF modules performances.

At punctual moments during the operation the TMP has exceeded the maximum recommended limits(2.1 bar). Nevertheless, the fibers have kept their integrity as it has been proven by consistent goodquality of the UF filtrate.

A large pool of Silt Density Index (SDI) values obtained throughout the 2 ¼ years of operation havebeen studied. The SDI test is a means of quantifying the amount of particulate contamination in a watersource for predicting the rate of colloidal and particulate fouling of RO membranes. SDI15 analyses onUF feed and RO feed (after the chemicals additions of sodium metabisulfite and antiscalant and after thepressure exchangers) streams are regularly done by the operators of the Moni plant. The UF feed waterhas in average an SDI15 of 5.2 %/min (from 155 measurements), while the RO feed water presents anaveraged SDI15 of 1.1 %/min (from 183 measurements). More detailed look was done at the UF filtrate(RO feed) SDI15 results. Figure 7 depicts the percentage of SDI15 results distributed in various intervals.As can be seen the vast majority (84%) of the results fell in the range of 0.5 – 1.5 %/min. Additionally,it is important to mention that in the 97% of the measurements SDI15 was below 2.

Figure 7. SDI15 measurements in the SWRO feed stream of the Moni desalination plant

3.2.2. Reverse Osmosis system

In the Moni SWRO desalination plant membrane types SW30HRLE and SW30ULE are combined in anISD (Internally Staged Design) configuration in order to attain the optimum hydraulic conditions insidethe pressure vessels. The ISD design is a combination of membranes with lower permeate flowperformance in the front positions of the vessel and of membranes with higher permeate flowperformance in the rear positions of the vessel. By following this criterion, fouling is minimized, feedpressure is reduced and longer membrane life can be expected [17-19].

The RO system is operated at a recovery of 43-44% at an average system flux of 11.5 lmh. A relativelylow average system flux was selected in order minimize the energy consumption. During the evaluatedperiod (slightly more than 2 years) of operation the Moni desalination plant has reliably achieved thenominal plant capacity. The RO feed pressure ranged from 65 to 69 bar depending on the feed watertemperature and feed water salinity increase due to energy recovery system. The RO performance hasbeen evaluated using FTNORM tool. FTNORM is a Microsoft® Excel® spreadsheet that allowskeeping track of the RO plant performance at normalized conditions.

0% 20% 40% 60%

<0.5

0.5 - 1.0

1.0 - 1.5

1.5 - 2.0

2.0 - 2.5

2.5 - 3.0

>3

Percentage of occurance

SD

I15

RO

fee

d[%

/min

]

N=183 measurements=100%


-12-

In Figure 8 the relative normalized permeate flow compared to the design productivity of the plant(20,000 m3/day) is presented together with the normalized salt rejection throughout the operating time.The flow performance of the system fluctuated around the design point, and the salt rejection was betterthan designed.

Figure 8. Relative normalized permeate flow (blue diamonds) and normalized salt rejection(red triangles) of the Moni RO system

The discontinuous horizontal line in red color shown in Figure 8 is the reference salt rejection level fromthe design (99.33%). From the graph below it can be observed that the salt rejection has been higherthan designed (especially during the first year 1 ½ years of operation) or very close to the designed point.Boron concentration has been always lower than 1 mg/l as required.

The pressure drop (P) has been lower than the reference value most of the time since start-up. A P of1.46 bar is the reference value obtained from a projection at start-up conditions using ROSA (ReverseOsmosis System Analysis) design software. This value assumes 0.345 bar pressure drop in the pipingfrom the high pressure pump outlet to the first RO membrane element. The normalized pressure drop(P) has been in average 0.98 bar, which represents a lower value than projected. Lately (recent months)after 2 years of operation the normalized P temporary increased slightly to higher values (up to 1.8 barapproximately). Alkaline CIP reduced the P back to normal levels.

3.2.3. High solid load event in the raw water

Most of the desalination plants in the island of Cyprus, among them the Moni Desalination plant, facedan event of high solid load in the raw water during some days of stormy weather in Cyprus. Thisoccurred around the end of December 2010. During various days the solids load in the raw waterincreased and the turbidity values remained consistently above 10 NTU. The exact values of turbiditycould not be measured since were above the maximum limit of the installed instruments, as reported byNirosoft Industries Ltd. Several plants in the island using conventional pretreatments had to ceaseoperations or reduce capacity of their systems. However, the Moni desalination plant could keepproducing the required product flow without reducing any cubic meter of permeate water throughout thestormy period.

98.0%

98.2%

98.4%

98.6%

98.8%

99.0%

99.2%

99.4%

99.6%

99.8%

100.0%

20%

40%

60%

80%

100%

120%

140%

160%

180%

200%

2/Feb/09 3/May/09 1/Aug/09 30/Oct/09 28/Jan/10 28/Apr/10 27/Jul/10 25/Oct/10 23/Jan/11

No

rma

lize

ds

alt

reje

cti

on

.

Rela

tive

no

rmali

zed

perm

eate

flo

w


-13-

Nirosoft Industries Ltd. reported that the actions taken during the high solids load period were mainlythe increase in cleaning frequencies of the pre-filters (prior to the UF system). The UF system couldcope with the increased turbidity without any change in the backwash frequency.

This experience demonstrates that the membrane pretreatment systems offer the benefit to be very stablewhen facing high solid load peaks. Additionally, despite of high fluctuations in suspended mater contentthe filtrate quality was found to be constant at a good quality, based on the SDI15 measurements.


-14-

IV. FIELD EXPERIENCES: TARANTO DESALINATION PLANT

This section presents the field experience of the dual membrane system in Taranto site.

4.1. The Taranto desalination plant - description

The dual membrane system studied in this section is located in the refinery site of Eni SpA nearby thecity of Taranto (Italy). The seawater desalination plant was constructed and it is being operated by theItalian OEM, Bernardinello Engineering SpA. The desalination plant produces demineralized waterusing a full membrane integrated system consisting of ultrafiltration followed by a double pass systemof reverse osmosis. The product water is sent to a product water tank, which receives water from otherdemineralization processes. The product water is used as feed for boilers in the refinery. Thedesalination plant was constructed in a modular set-up. The ultrafiltration and reverse osmosis trainswere installed inside 40 feet containers. The use of ultrafiltration as pretreatment helped to minimize thefootprint of the plant and the use of containers to house the skids helped to minimize the costs related tocivil works needed to construct the plant.

Figure 9 shows a picture of the containers which house the desalination plant pretreatment and reverseosmosis system.

Figure 9. Photograph of the plant. Courtesy of Bernardinello Engineering SpA.

The integrated membrane system in Taranto has a capacity of 5,760 m3/day of SWRO permeate flow.The plant was constructed and started up in two phases. The first phase of the plant which produces3,460 m3/day of SWRO permeate flow was started-up in June 2009. A future expansion of the systemwith the same water treatment process scheme (ultrafiltration as pretreatment followed by double pass ofreverse osmosis) was carried out during 2010 and the start-up of this second phase was conducted inOctober 2010. The second phase system produces 2,300 m3/day of RO permeate.

In the next figure (Figure 10) a simplified flow diagram of the Taranto desalination system is depicted:

Figure 10. Process diagram from Taranto desalination plant

The raw water source is an open seawater intake. As in the previous example described (MoniDesalination plant, see chapter III), no coagulation is used in the feed water.

Productwater

UF FiltrateWater Tank

1st passpermeate

Water Tank

1st pass RO 2 nd pass RO

2nd passpermeate

Water Tank

Self-cleaningfilter: 200 µm

UF


-15-

A summary of the main characteristics of the feed water quality and the specified product water qualityrequired are presented in Table 3.

Table 3. Feed and product water quality summary table

Parameter Unit UF Feed water quality Required product water quality

Temperature °C 12 – 36

pH 8.0 7 - 9

Turbidity NTU 0.1 – 40

Conductivity S/cm 55,300 < 20

Total Dissolved Solids mg/l 39,600* < 10

Chloride (Cl-) mg/L 22,100 < 5.5

* The feed water TDS has been calculated based on the factor “conductivity / TDS” of 1.4

As can be observed in Table 3, the feed water temperature and turbidity values vary significantlythroughout the year. The Feed water turbidity has been as high as 40 NTU in stormy events.

The raw water is firstly pre-filtered in a set of self-cleaning filters of 200 m screen size, in order toremove big particles from the UF feed stream. The UF unit consists of 14 racks each containing 15 xDOW™ Ultrafiltration modules, a total of 210 DOW™ Ultrafiltration modules. The DOW™Ultrafiltration module used is the 2860 module type [16], featuring 51 m2 of active area. Each 40-feetcontainer can house up to 4 UF racks.

The UF filtrate water is then collected in a tank which feeds the RO system. The RO system consists oftwo passes system. The UF filtrate is firstly treated by the first pass system which is configured in a twostage system with an inter-stage turbocharger which is used to boost the pressure of the second stagefeed. The first pas system is divided in 5 trains, each one contains 10 pressure vessels in the first stageand 10 pressure vessels in the second stage. Each train is housed in a 40-feet container as can beobserved in Figure 11. Within each pressure vessel, 7 x DOW FILMTEC™ SW30HRLE-400imembrane elements are installed. One of the key aspects of the system is the high operating recovery inthe first pass which is 58-60%.

In Figure 11 two photographs of a UF rack and a photograph of a SWRO rack are shown.

Figure 11. Photographs of UF (left) and SWRO (right) racks of the Taranto desalination plant. Courtesy of BernardinelloEngineering SpA.

The permeate water of the first pass system is collected to a tank. The second pass system is also dividedin 4 trains, each one contains 6 pressure vessels in the first stage and 3 pressure vessels in the second


-16-

stage. Within each pressure vessel, 6 membrane elements are installed. The second pass system isoperated at a recovery around 85-88 %. The concentrate stream of the second pass system is recycledand sent upstream to the UF filtrate tank in order to decrease the salinity of the feed water of the firstpass RO system and to increase the overall system recovery.

The second pass system from the phase II contains 2 trains. This section uses Seawater RO membraneelements (DOW FILMTEC™ SW30HRLE-400i) in order to maximize the salt rejection and to improvethe final water product quality. The study of the 2nd pass system from phase I remained out of the scopeof this study.

The table below includes the main parameters of the UF and RO systems as well as their main operatingconditions from the systems studied in the present work:

Table 4. Summary of the UF and RO systems main parameters in the Taranto Desalination plant

UnitTotal

Capacity(m

3/h)

Operating flux(l/m

2h)

Number oftrains/stages

Product installedTotal numberof modules

Ultrafiltration(phase I and II)

400 50 N.A. DOW™ UF 2860 210

1st

pass RO(phase I and II)

240 9.2 5/2DOW FILMTEC™SW30HRLE-400i

700

2nd

pass RO(phase II)*

82 20.5 2/2DOW FILMTEC™SW30HRLE-400i

108

*the study of the 2nd pass system from phase I remained out of the scope of this study.

4.2. Dual membrane system performance

In this chapter the performances of the ultrafiltration and reverse osmosis systems from Taranto iscovered.

4.2.1. Ultrafiltration system

The ultrafiltration system in Taranto runs at an operating flux of around 50 l/h/m2 and the backwash iscarried out once every 35 minutes. The backwash operation is performed with UF filtrate without anychemical addition.

The normalized TMP and normalized permeability data of the UF system are plotted below. The datafrom phase I and phase II have been plotted in two separate graphs, as can be observed in Figure 12 andFigure 13. The figures below show both parameters for each of the individual racks of the system,except for UF 8 from phase I since the operating data was not available for the study. A horizontaldashed line indicates the average value of each parameter.


-17-

Figure 12 Normalized TMP and Permeability for all UF racks in Taranto plant Phase I.

Figure 13 Normalized TMP and Permeability for all UF racks in Taranto plant Phase II.

The UF trains have shown a robust and reliable flow performance despite of the high feed waterturbidity events faced throughout the operation.

The UF trains from phase I show a stable performance at an average normalized TMP of 0.76 bar and atan average normalized permeability of 67 lmh/bar. The UF trains from phase II show a normalizedpermeability around 100 lmh/bar and a normalized TMP around 0.46 after 3 months since start-up.

The permeability of the UF trains from phase I has decreased from the initial 100-110 lmh/bar to around70 lmh. This indicates that the UF fibers run with some fouling which is not removed with the frequentoxidant CEB actions (using NaOCl) done every day. Caustic and acid CEB cleanings are done only in a

(80)

(60)

(40)

(20)

0

20

40

60

80

100

120

140

160

0.0

0.3

0.5

0.8

1.0

1.3

1.5

1.8

2.0

2.3

2.5

2.8

3.0

No

rmal

ize

dT

MP

(Bar

)


UF1 UF2 UF3 UF4 UF5 UF6 UF 7

No

rma

lized

pe

rme

ab

ility

(lm

h/b

ar)

Norm

aliz

ed

TM

P(b

ar)

(80)

(60)

(40)

(20)

0

20

40

60

80

100

120

140

160

0.0

0.3

0.5

0.8

1.0

1.3

1.5

1.8

2.0

2.3

2.5

2.8

3.0

No

rmal

ize

dT

MP

(Bar

)


UF1 UF2 UF3 UF4 UF5 UF6

No

rma

lized

pe

rme

ab

ility

(lm

h/b

ar)

Norm

aliz

ed

TM

P(b

ar)


-18-

needed basis (approximately twice every month). No CIP has been performed in the UF modules sincestart-up.

It has been identified that iron fouling is present on the fibers. Therefore, an acid CIP with 2% (inweight) of oxalic acid or citric acid would be beneficial to restore the TMP. Additionally, an alkalineCIP action could also be beneficial to efficiently eliminate all biological or organic fouling that couldnot be removed by the programmed BW and CEB actions. CIP actions have been found very effectiveprocedure to restore the flow performances in other reported cases [20,21]. However, at present, theoperator is very satisfied with the hydraulic performance and has not taken action as of yet.

With regards to the UF filtrate water quality, since start-up the UF system has provided filtrate water ofconsistent good quality. The UF filtrate quality is monitored frequently with SDI15 measurements. Apool of 275 SDI15 measurements have been analyzed and plotted in a graph as shown in Figure 14. Inaverage, the SDI15 has been 0.92 %/min. It is worthwhile to point out that the SDI15 values have beenalways less than 2 %/min and that 90% of the measurements shown values below 1.5 %/min.

Figure 14. SDI15 measurements in the SWRO feed stream of Taranto desalination plant

4.2.2. First pass reverse osmosis system

The first pass system is operated, as indicated in previous chapters, at very high recovery ratios between58 and 60 %. This high recovery is achieved by using a double stage RO system with an inter-stageturbocharger which boosts the pressure in order to balance the flow rates between stages. The operatingflux was designed relatively low in order to minimize the pressure requirements of the system. Theaverage system flux of the first pass system is set at 9.2 l/h/m2.

The first pass system is fed with a blending stream between the UF filtrate (89.5 %) and second pass ROsystem concentrate (10.5 %). The feed water conductivity amounts in average 49,000 S/cm.

The first stage feed pressure ranges between 40 and 50 bar. A turbocharger adds 22-26 bar to the firststage concentrate pressure leading to second stage feed pressures of around 62 – 73 bar.

In Figure 15 the normalized permeate flow and the feed water temperature data are plotted. Twodifferentiate series of data have been plotted:

Series of data points shown as blue colored diamonds indicate the average normalized permeateflow of the three RO trains started up in June 2009.

0% 20% 40% 60% 80%

<0.5

0.5 - 1.0

1.0 - 1.5

1.5 - 2.0

2.0 - 2.5

2.5 - 3.0

>3

Percentage of occuranceS

DI1

5R

Ofe

ed

[%/m

in]

N=275 measurements=100%


-19-

Series of data points shown as green colored triangles indicate the average normalized permeateflow of the three RO trains started up in October 2010.

Figure 15. Normalized permeate flow of the first pas RO system. Taranto desalination plant.

The normalized permeate flow of the three RO trains started up in mid 2009 showed some variabilitybetween 22 and 52 m3/h. The averaged value throughout the reported 1 ½ years has been 32 m3/h. Thetwo trains started up in October 2010 showed in average 29% higher normalized flow compared to thosetrains started up 15 months earlier.

The pressure drop data available shows some fluctuation possibly due to the fact that P is calculated bytwo readings from two PIT instruments instead of a PIT. The P across both stages observed in thelines running since 2009 was for one year below the design value (2.1 bar as predicted by ROSA).During summer period of 2010 the P of all three lines started to increase above the predicted P values.Two complete CIP cycles (Caustic cleaning followed by an acidic cleaning) were applied and the systemrestored completely the design value of P.

The permeate conductivity of the first pass system ranges between 500 to 1,200 S/cm dependingmainly on the feed water temperature.

4.2.3. Second pass reverse osmosis (Phase II)

The second pass system from Phase II is in operation since October, 2010. The recovery ratios at whichit is operated range between 85 and 88%. The membrane elements installed in the two trains from thePhase II use SWRO membrane elements (DOW FILMTEC™ SW30HRLE-400i membrane elements) tomaximize the salt rejection in the second pass system. The membrane elements of this section are beingoperated at an average system flux of 20.5 l/h/m2.

The normalized pressure drop across the two stages of the second pass system is constant at around40 % lower value than predicted by ROSA software. In average the normalized P has been 1.2 bar

-20

-15

-10

-5

0

5

10

15

20

25

30

35

40

0

10

20

30

40

50

60

70

80

90

100

110

120

Perm

eate

Flo

w(m

3/

hr)

Operating Time

Norm

aliz

ed

perm

eate

flow

(m3/h

)

Feed

wate

rte

mpera

ture

(°C

)

Normalized permeate flow Phase I Normalized permeate flow Phase II


-20-

across both stages (considering performance data from start-up until end of January 2011) while theprojected value was 2 bar.

The normalized permeate has been constant throughout the operation period with some fluctuations of10%.

The normalized salt rejection since start-up, has been in average 99.4 % which is higher than designed(99.2%). The conductivity of the permeate water has ranged from 1.3 to 6.9 S/cm, being 2.7S/cm theaveraged value. During this studied period, the feed water temperature varied from 27 to 15 °C, and theaverage feed water temperature have been 20.3 °C.

As expected, after four months of operation CIP action is not yet needed.


-21-

V. SUMMARY AND CONCLUSIONS

In the last decade, ultrafiltration pretreatment for RO systems have increasingly gained acceptance asviable pretreatment alternatives for seawater desalination. The key drivers for adopting membranepretreatment techniques that have been quoted frequently include the greater capability to cope withfluctuations and high solid loads in raw waters, smaller footprint, higher environmental sustainability,lower SWRO stage cost. Easier design and operation have also been reported as key drivers. UF offersthe possibility to containerize the pretreatment due to the more compact, modular characteristics andlower weight which makes ultrafiltration technology the preferred choice for mobile units. The benefitsof building containerized systems are the minimization of civil works and related costs, mobility of thesystems, faster system installation, etc. These benefits that the membrane pretreatment technology offersto the desalination industry provides to the industry with the ability to respond more efficiently totemporary and/or emergency fresh water needs.

Despite the fact that many SWRO facilities include low pressure membrane pretreatment, quantifiableperformance data reported is still relatively scarce. This publication reported field experiences with twoindustrial scale integrated membrane systems in the seawater desalination field using DOW™Ultrafiltration membranes as pretreatment for DOW FILMTEC™ RO membranes. Both systemsshowed robust and reliably performance with a consistent pretreated water quality despite of thefluctuations of the raw water quality.

Since December 2008, the Moni sea water desalination plant continuously and reliably supplies 20,000m3/day of drinking water for the population of Limassol (Cyprus). The plant is a mobile system wherethe various units are containerized. This mobility aspect of the system is regarded as one of the fewoptions when an immediate solution against water scarcity is needed. The UF system operates at arecovery of 91% with an availability of 92%. The normalized TMP and normalized permeability are inaverage 0.86 bar and 74 lmh/bar, respectively. The current normalized TMP is very close to thedesigned initial TMP 0.6-0.7 bar. This indicates that the current BW and CEB procedures are effectivein removing the fouling and restoring the UF modules performances. The plant presents optimizedcleaning procedures which minimize the chemical consumptions during Chemical Enhanced Backwash(CEB). The UF filtrate quality has been consistently of good quality. The SDI15 has been in 97% of thetimes (183 measurements) below 2 %/min with an average SDI15 value of 1.1 %/min. The dualmembrane system in Moni reliably achieves the required permeate flow together with the requiredquality drinking water quality, with the boron concentration lower than 1 mg/l. The normalized pressuredrop (P) across the RO pressure vessels has been in average 0.98 bar, which represents a lower valuethan projected. Additionally, the integrated membrane system at Moni could overcome without losingany productivity a stormy weather event which increased very significantly the solids load of the rawwater. This fact could demonstrate that the membrane pretreatment systems offer the benefit to be morestable than conventional pretreatment systems when overcoming high solid load peaks. Additionally,despite of extreme fluctuations in suspended mater content the filtrate quality was found to be constantat a good quality, based on the routine SDI15 measurements.

The Taranto Sea water desalination plant reported in this publication is a containerized integratedmembrane system installed in a refinery nearby Taranto (Italy). The system is operational since mid2009 and it has currently a capacity of 5,760 m3/day of SWRO permeate flow. The system uses twopasses of RO to produce demineralized water to feed the boilers of the refinery. The UF system has beenoperating with feed waters containing high turbidities, up to levels of 40 NTU. Despite this the UF


-22-

system (operating at around 50 l/h/m2 with BW frequencies of 35 min) showed robust and reliableperformance. Average normalized permeabilities of 67 lmh/bar and 100 lmh/bar are observed in theolder UF trains (20 months in operation) and in the newer UF trains (3 months in operation),respectively. This indicates a higher degree of fouling on the fibers running since longer time. CIPactions, which have not been done yet, could be beneficial in restoring the permeability of the older UFlines. The UF filtrate SDI15 values have shown excellent numbers. Up to 90% of the 275 measurementsshowed SDI15 values below 1.5 %, being 0.92 %/min the average value. One of the key characteristicsof the SWRO system from the Taranto desalination plant is the successful operation of the system at ahigh recovery (60%). The RO section has shown a constant performance and only two Cleaning In Place(CIP) actions in more than 1 ½ years were needed. This, could be partially, attributed to the high waterquality provided by the ultrafiltration membranes which positively affect the sustainable high recoveryapplied in the SWRO system.

Both systems studied (Moni and Taranto) show that the integrated ultrafiltration and reverse osmosissystems in sea water desalination industry improve the water supply options by the following attributes:

- Modularity and Mobility: both systems are containerized. This minimizes the civil worksrequired. Additionally, the Moni desalination plant is a mobile unit and can be dismantled,transported and then reassembled in a new location in a very short period of time (3 months asreported by SubSea Infrastructure Ltd.).

- Low environmental impact: Both UF systems are operated without coagulation which reducesthe sludge mass created by this pretreatment technique and the pH adjustment in the feed water.

- High capability to cope with difficult waters: UF feed waters in both cases reached highturbidities above 10 NTU (quantified up to 40 NTU in Taranto plant). Despite this, the UFsystem maintained the hydraulic output.

- Robustness: Despite of the significant variations in UF feed water quality in terms of solids loadthe UF filtrate water has been of constant good quality throughout the operation time enablingreliable RO performance. The SDI15 values have been consistently below 2 %/min in 100% ofthe measurements (275 measurements) in the Taranto plant and in 97% of the measurements(183 measurements) in the Moni desalination plant.

- Automation: both UF systems are operated with a high degree of automation.- Reduction of system capital cost. The Taranto plant is operated at high recovery of around 60%

using two stages in the seawater pass with an intermediate booster pump. This enables areduction of the RO feed flow and thus a reduction of the pretreatment size.


-23-

V. ACKNOWLEDGEMENTS

The authors would like to acknowledge the following contributors:o David Dwek and Murray Eldridge from SubSea Infrastructure Ltd and Mino Negrin, Andrew

Urquhart, Avi Bonibay and Moshe Amir and from Nirosoft Industries Ltd. for theircollaboration in providing the access to the data reported from the Moni Desalination plant.

o Federico Nicolazzi, Federica Fizzarin and Mario De Cristofaro from BernardinelloEngineering SpA. for their collaboration in providing the access to the data reported from theTaranto Desalination plant.

o Francesco Lanari from Dow Water and Process Solutions for his support in this study.

VI. INDUSTRY DATA AND PRODUCT INFORMATION

The industry data provided in this article were collected from 2009 until the first quarter of 2011 and areincluded to illustrate the field experiences with dual membrane systems in the seawater desalinationarena. Industry data and product information are provided in good faith for informational purposes only.The authors assume no obligation or liability for such data presented herein. No warranties are given; allimplied warranties of merchantability or fitness for a particular purpose are expressly excluded.

VII. REFERENCES

1 Busch, Markus; Chu, Robert; Rosenberg, Steve, Novel Trends in Dual Membrane Systems for SeawaterDesalination: Minimum Primary Pretreatment and Low Environmental Impact Treatment Schemes,International Desalination Association (IDA) Journal, First Quarter 2010, Vol. 2, Number 1, pages 56-71,2010.

2 DOW™ Ultrafiltration Product Manual,http://www.dowwaterandprocess.com/support_training/literature_manuals/prod_manuals.htm, April 2011

3 Gasia-Bruch, Eduard; Sehn, Peter, García-Molina, Verónica; Busch, Markus; Raize, Ofer; Negrin, Mino,Field experience with a 20,000 m3/day integrated membrane sea water desalination plant in Cyprus,European Desalination Society (EDS) EUROMED Conference, Israel, Oct-2010.

4 Busch, Markus; Evaluation and Cost Modeling for Ultrafiltration Pretreatment of Seawater in ReverseOsmosis Desalination Plants, Dissertation, submitted to Wroclaw University of Technology, 2010

5 Mauguin, G.; Corsin, P.; Concentrate and other waste disposals from SWRO plants: characterization andreduction of their environmental impact, Desalination 182 (2005) 355–364 Desalination Society (EDS)EUROMED Conference, Israel, Oct-2010.

6 García-Molina, Verónica; Marzal, Miguel Ángel; Hoehn, Kai-Uwe; Designing membrane systems for thecoming future: Perth II desalination plant, International Desalination Association (IDA) World Congress –Atlantis, The Palm – Dubai, UAE November 7-12, 2009

7 Knops, Frans; 'Comparing conventional pre-treatment to ultrafiltration for SWRO pre-treatment,International Desalination Association (IDA) International Water Forum, March 5-6. 2006, Dubai, UAE.


-24-

8 European Water Partnership, Subsea Infrastructure has delivered over 2 billion litres of drinking water sinceDecember, EWP web site < http://www.ewp.eu/news/ewp-news>, May 5th 2009.

9 Galloway, Merrilee; von Gottberg, Antonia; Mahoney, James, UF for SWRO pretreatment, InternationalDesalination Association (IDA) World Congress, Bahamas, 2003

10 Gorenflo, Andreas; Saleh Abu Amirah, Rami; Smaili, Abdulkarim; Bajunaid, Abdullah A.; Shoaiba Barge:The Largest Sea Based Desalination Plant in the World, International Desalination Association (IDA) WorldCongress – Atlantis, The Palm – Dubai, UAE November 7-12, 2009

11 Henthorne, L, Evaluation of Membrane Pretreatment for Seawater RO DesalinationUF for SWROpretreatment, United States Bureau of Reclamation, Desalination and Water Purification Research andDevelopment Program, Report No. 106, 2005.

12 Knops, F., Comparing conventional pre-treatment to ultrafiltration for SWRO pre-treatment, InternationalDesalination Association (IDA) International Water Forum, Dubai, UAE, March 5-6. 2006.

13 Knops, F.; van Hoof, S.; Zark, A., Operating Experience of a New Ultrafiltration Membrane for Pre-treatment of Seawater Reverse Osmosis, International Desalination Association (IDA) World Congress,Maspalomas, Gran Canaria, Spain, October 21-26, 2007.

14 Pearce, G., SWRO pre-treatment: Cost and Sustainability. Filtration + Separtation, March/April 2010.

15 Chu, R.; Wei, J.; and Busch, M., Economic evaluation of UF + SWRO in seawater desalination. ChineseDesalination Association Conference, Qing Dao, China, 2009.

16 Product Information DOWTM Ultrafiltration Modules: Model SFP-2860 and SFD-2860,www.dowwaterandprocess.com, Form No. 795-00013-1109, March 2011

17 Mickols, W.E., Busch, M., Maeda, Y., Tonner, J., A novel design approach for seawater plants, IDA 2005World Congress on Desalination and Water Reuse, SP05-052, 2005

18 Busch, Markus, et al., Novel Internally Staged Design Approach: Reducing Seawater Desalination Costswith Higher and More Balanced Flux, Desalination & Water Reuse Vol. 18/1, pages 29-32, 2008.

19 García-Molina, Verónica; Busch, Markus; Sehn, Peter, Cost savings by novel seawater reverse osmosiselements and design concepts, EDS EUROMED Conference, Jordan, Nov.2008.

20 Riaza, A.; Shao, E.; Busch, M.; Suarez, J.; Salas, J.; and Wang, D., Performance of outside-in pressurizedUF in Qingdao pilot test, European Desalination Society (EDS) Conference, Baden-Baden, Germany, 2009.

21 Busch, M.; Chu, R.; Kolbe, U.; Meng, Q.; Li, S. J., Integrated UF and Reverse Osmosis Membrane SystemFor Seawater Desalination—1,000 Days Field Experience With Dow UF And Filmtec Technology In TheWang Tan Datang Power Plant, European Desalination Society (EDS) Conference, EUROMED series, DeadSea, Jordan, 2008.

Documents

DUAL MEMBRANE SYSTEMS IN SEAWATER ...watec.gr/wp-content/uploads/2017/06/RWL-DualMembrane...IDA Desalination Industry Action for Good / Santa Margherita, Portofino, Italy May 16-18,