Desalination 254 (2010) 170–174

Contents lists available at ScienceDirect

Desalination

j ourna l homepage: www.e lsev ie r.com/ locate /desa l

Comparative study on stand-alone and parallel operating schemes of energy recoverydevice for SWRO system

Xiaopeng Wang, Yue Wang ⁎, Jianping Wang, Shichang Xu, Yuxin Wang, Shichang WangChemical Engineering Research Center, School of Chemical Engineering and Technology, Tianjin University, State Key Laboratory of Chemical Engineering,Tianjin State Key Lab of Membrane Science and Desalination technology, Tianjin 300072, China

⁎ Corresponding author. Tel./fax: +86 22 27404347.E-mail address: tdwy75@tju.edu.cn (Y. Wang).

0011-9164/$ – see front matter © 2009 Elsevier B.V. Adoi:10.1016/j.desal.2009.11.038

a b s t r a c t

a r t i c l e i n f o

Article history:Received 12 September 2009Received in revised form 23 November 2009Accepted 27 November 2009Available online 6 January 2010

Keywords:Energy recovery deviceSeawater reverse osmosisParallel operationRotary fluid switcher

As known, positive displacement (PD) energy recovery device (ERD) transfers hydraulic energy from theejected brine of the RO modules directly to the seawater feed and its energy recovery efficiency achieved canbe as high as 95% or more. The PD type ERD has been concerned widely in the market place and globallyadopted into seawater reverse osmosis (SWRO) desalination plant by designers and operators. Usually, tosatisfy the needs of large to super-large SWRO plant capacity, parallel operations of ERD facilities are themost convenient and effective way that can be chosen. In this article, a PD type ERD, named FS-ERD wasintroduced and a parallel FS-ERD setup was designed and built. Operating schemes for stand-alone operationand parallel operation of the setup were developed and experimentally compared for different test capacity.

ll rights reserved.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

Seawater reverse osmosis (SWRO) desalination systems arewidely used to produce potable water from the sea nowadays.However, the excessive power consumption is still an emergentproblem to be solved for the system. As is known, the SWRO is apressure-drivenmembrane process and themain power consumptionarises from the high pressure (HP) pump, which accounts for at least35% of operating costs in total system [1]. Since more than half of thehydraulic energy supplied by the HP pump are incorporated in theejected brine of the RO modules with a traditional conversion rate of35–45%, so it is important to recover the thus lost hydraulic energy ofthe brine through the energy recovery approach.

Currently in the market place, there are principally two categoriesof energy recovery device (ERD), the positive displacement (PD) typeand the centrifugal type. The PD type ERD transfers hydraulic energyfrom the ejected brine of RO modules directly to the seawater feedand due to its direct transfer manner, the energy recovery efficiencyachieved is higher than that of the centrifugal type, and can be as highas 95% or more [2]. Currently the PD type ERD has become one of themost efficient ERDs in themarket place and has been globally adoptedinto seawater reverse osmosis (SWRO) desalination plant bydesigners and operators [3].

As the development and maturation of RO technology, large oreven super-large SWRO plants are established around the world, and

the per train size has reached to 10,000–15,000 m3/d. To satisfy theincreased needs of SWRO plant capacity, many efforts have beenput into practice, including enlargement of the ERD per unit size, andalso parallel operations of ERD facilities [4,5]. Attentively, the paralleloperation is considered the most common and effective way to bechosen, which is because the parallel operatingmode not only extendsthe device's capacity, but also has the potential to significantly reduceor even eliminate fluctuations of the ERD working streams [6].

2. Experimental setup

2.1. Working principle of the FS-ERD

In this work, a newly developed PD type ERD named FS-ERD wasinvestigated. The FS-ERD was mainly composed of three portions, arotary fluid switcher, two pressure cylinders and a check valve nest. Thecore component of the FS-ERD is the rotary fluid switcher, which isfeatured with four joint ports and twoworking phases similar to a two-position four-way valve. Fig. 1 gives theworking principle of the FS-ERDin phaseI. Under this condition, the HP brine stream is imported intocylinder 1 and therein the pre-filled low pressure (LP) seawater feed ispressurized and pumped out, which is called the pressurizing stroke.Simultaneously, HP brine stream in cylinder 2 would be depressurizedand was drained out by the incoming LP seawater feed, and the processis called depressurizing stroke. Thereafter, when the FS-ERD accom-plishes its pressurizing stroke (and also the depressurizing stroke), theswitcher would rotate to working phaseIIat a low speed of 7.5 rpmdriven by motor, which denotes that the stroke modes in cylinders arealternated to each other.

Fig. 1. Working principle of FS-ERD at phase I.

Fig. 3. Schematic diagram of the experiment.

171X. Wang et al. / Desalination 254 (2010) 170–174

A piston is installed in each cylinder to isolate seawater from brineand ensure minimum mixing between them during the operation.Position of the piston is detected to judge whether the pressurizing ordepressurizing stroke has been accomplished, and when the fluidswitcher should be changed to the next working phase.

Further, an experimental parallel ERD unit, which comprises twoidentical FS-ERDs and can be operated independently, was set up asshown in Fig. 2. Here, the cylinders of the FS-ERD are made ofpolymethyl methacrylate for convenient observing of the experimen-tal process. Fig. 3 gives the schematic diagram of the experiments,which is similar to the actual SWRO system, except that the ROmembranemodules are not incorporated, and the operating pressuresadopted are comparatively lower (below 0.6 MPa). Tap water is usedas the working medium instead of actual seawater and brine in SWROdesalination plants.

Data acquisition system is designed and incorporated in theexperimental system. Flow rates of LP seawater (Qsi) and HP brine(Qbi) feeding to the ERDs are measured by the flow transmittersrespectively with a precision of ±0.5%. Pressures of the streams to andfrom the ERDs are also tested by the pressure transducers withmeasuring precision of ±0.5%.

2.2. Opening change mechanism of rotary fluid switcher

The rotary fluid switcher is the core subassembly of the FS-ERDunit and the switcher accomplishes its phase change by rotating itsmulti-channel rotor around the switcher's shell. Here a working phaseis defined at the location where the specific channel in the rotorsuperposes accurately with the window in the shell. Departure fromthe working phase position, the channel in the rotor would not be

Fig. 2. Experimental setup of the parallel FS-ERDs.

sufficiently opened and thus the stream to and from the rotor channelwill be influenced or even interrupted. Fig. 4 gives the opening changemechanism of the rotary fluid switcher. It can be seen that workingphaseIand phaseIIcorresponding to the full opening line alternatesequentially, and the duration time for working phaseIorIIsigned asT1. The switch time of the switcher is defined as T2, which includesthe switcher's response time and the full closure remaining time T3.Here the time T3 is determined by the channel size of the rotor, whilethe time difference between T2 and T3 decides from the rotatingspeed of the rotor.

In this paper, operating schemes of stand-alone operation andparallel operation are developed and uploaded in the PLC controlsystem. Dynamic performances of the experimental setup under thetwo operating schemes were tested and comparatively analyzed.

3. Stand-alone operation experiments

In stand-aloneoperation, the fluid dynamics performances of FS-ERDdevice were studied at test capacity of 1.0 m3/h and 2.0 m3/hrespectively.

3.1. Flow rate fluctuations of streams in stand-alone operation

Fig. 4 illustrates the flow rate fluctuations of LP seawater (Qsi) andHP brine (Qbi) at test capacity of 1.0 m3/h. It can be seen that both Qsiand Qbi have a downward fluctuation periodically and also theamplitude of Qbi is significantly much larger than that of Qsi.Referring the opening change mechanism described in Fig. 4, it isbelieved that the periodical downward fluctuations of Qsi and Qbistreams derive from the cyclic switch of the fluid switcher and furtherthe sudden interruption of streams during the switch process. Thecause of the larger amplitude of Qbi is that the brine stream is directly

Fig. 4. Openings of the fluid switcher vs. time for stand-alone operation.

Fig. 5. Fluctuations of Qbi and Qsi with time when test capacity is 1.0 m3/h in stand-alone operation.

Fig. 8. Fluctuations of Pbi and Pbo with time when test capacity is 1.0 m3/h in stand-alone operation.

172 X. Wang et al. / Desalination 254 (2010) 170–174

controlled by the switcher and its responding to the interruptionappears muchmore tempestuously. While for the LP seawater stream,it flows into cylinder through check valve nest, and when facing thesudden interruption, the cylinder depresses the amplitude of Qsi like abuffer pocket.

When the test capacity adds to 2.0 m3/h as shown in Fig. 6, thedownward fluctuation trends of Qsi and Qbi conform to Fig. 5, buttheir fluctuations amplitude and frequency become much more

Fig. 6. Fluctuations of Qbi and Qsi with time when test capacity is 2.0 m3/h in stand-alone operation.

Fig. 7. Fluctuations of Psi and Pso with time when test capacity is 1.0 m3/h in stand-alone operation.

tempestuous (Fig. 6). Compared to Fig. 5, the horizontal steady linefor the flow rate curves shrinks to nearly a point, which reveals thatthe operating capacity of the given ERD unit has exceeded its handlinglimit, also, the cyclic period has approached to the switch time of theswitcher.

3.2. Pressure fluctuations of streams in stand-alone operation

Figs. 7 and 8 illustrate the pressure fluctuations of seawaterstreams and brine streams. It can be seen that the pressure curves ofLP seawater (Psi) and HP brine (Pbi) all present periodic upwardfluctuations at the operating capacity of 1.0 m3/h, and the cyclicfrequency is uniform to that of flow rate in Fig. 5. So when consideringthe flow rate fluctuations in Fig. 5, conclusions can be made that theupward fluctuations of Psi and Pbi all origin from the downwardfluctuations of flows. Here the upward fluctuations of the pressure canbe characterized as a phenomenon of momentary accumulations ofstream pressure in the relevant pipelines. For actual plant conditions,the pressure accumulations may evolve into an actual “waterhammer” due to the much longer length of connecting pipelines tothe ERD devices and the higher flow velocity adopted, and furtherdestroys the stability of the devices and the application system.

However, the pressure curves of HP seawater (Pso) and LP brine(Pbo) fluctuate downwardly and periodically. That is because the HPseawater flow is pressurized and driven by the HP brine flow, andwhen HP brine flow shrinks or even cuts off, the HP seawater flowwill

Fig. 9. Fluctuations of Psi and Pso with time when test capacity is 2.0 m3/h in stand-alone operation.

Fig. 10. Fluctuations of Pbi and Pbo with time when test capacity is 2.0 m3/h in stand-alone operation.

Fig. 12. Fluctuations of Qbi and Qsi with time when test capacity is 2.0 m3/h in paralleloperation.

173X. Wang et al. / Desalination 254 (2010) 170–174

lose its driving power and the Pso gets lower due to the momentarysiphon for HP seawater pipeline. The same reason for fluctuation ofPbo is also applicable.

When the test capacity of the ERD unit increases from 1.0 m3/h to2.0 m3/h, pressure fluctuations shown in Figs. 9 and 10, reflect thesimilar trend as in Figs. 7 and 8. Comparatively, fluctuations in Figs.8 and 9 become more tempestuously and frequently, the reasons arethe same as that of Fig. 5.

To sum up, the fluctuations of pressure and flow rate for piston-typeERD are inevitable in stand-alone operation, which are determined bythe inherent characteristics of rotor structure. To solve the problems, apotential way is to operate the ERDs in parallel which will be discussedmainly in the following sections.

4. Parallel operation experiments

The fluid dynamic performances of the parallel FS-ERDs are alsoexperimentally studied at the test capacity of 1.0 m3/h and 2.0 m3/hrespectively. Here, the synchrony of the parallel devices is ensured byfixed timing of the PLC controller system.

4.1. Flow rate fluctuations of streams in parallel operation

Fig. 11 illustrates the fluctuations of Qsi and Qbi at test capacity of1.0 m3/h. It can be seen that curves of Qsi and Qbi nearly keep constantand the previous fluctuations for stand-alone operation in Fig. 4 has

Fig. 11. Fluctuations of Qbi and Qsi with time when test capacity is 1.0 m3/h in paralleloperation.

been smoothedor even eliminated. The significant improvement shouldattribute to the parallel operating scheme of two FS-ERDs, whichpromises at least one set of FS-ERD keeping at the status of steadypressurizing (or depressurizing) stroke and ensures the continuities ofstreams to and from the ERD unit.

Fig. 12 shows the fluctuation curves of Qsi and Qbi when the testcapacity adds to 2.0 m3/h. Compared to Fig. 10, since the test capacityof the system has been enlarged to twice, the flow curves could stillmaintain its stabilization. That means the parallel ERDs can not onlystabilize the working streams, but also have a good adaptability forincreased capacity of the system.

4.2. Pressure fluctuations of streams in parallel operation

Figs. 13 and 14 describe the pressure fluctuations of seawater (Psi,Pso) and brine (Pbi, Pbo) streams synchronizing to the flow ratecurves in Fig. 11. It can be seen the pressure curves keep almostconstant during the experiments. The cause can be explained that thepressure fluctuations are in most cases influenced and decided by theflow fluctuations of the relative streams. And since the fluctuationproblems of Qsi and Qbi have been successfully smoothed throughparallel operatingmode, the pressure fluctuationswill no longer occurfor the same operating conditions. Pressure fluctuation curves at testcapacity of 2.0 m3/h reflect the same fluctuation trend and are shownin Figs. 15 and 16.

Fig. 13. Fluctuations of Psi and Pso with time when test capacity is 1.0 m3/h in paralleloperation.

Fig. 15. Fluctuations of Psi and Pso with time when test capacity is 2.0 m3/h in paralleloperation.

Fig. 16. Fluctuations of Pbi and Pbo with time when test capacity is 2.0 m3/h in paralleloperation.Fig. 14. Fluctuations of Pbi and Pbo with time when test capacity is 1.0 m3/h in parallel

operation.

174 X. Wang et al. / Desalination 254 (2010) 170–174

5. Conclusions

A newly development PD type ERD named FS-ERD was introducedand experimentally studied. Performances of the FS-ERD in single-alone operation and parallel operationwere tested and compared. The

results showed that the parallel operation of two sets of ERDs can notonly extend the capacity of the system but also remarkably improvethe stability and continuity of the working streams to and from theERDs. The demonstration of the parallel FS-ERDs under a real SWROoperating pressure is in progress at the FS-ERD test stand with acapacity of 2×500 m3/d in our laboratory.

Acknowledgements

This research is supported by the National Key Technologies R&DProgram (No. 2006 BAB03A021).

References

[1] E. Oklejas, W.F. Pergande, Integration of advanced high-pressure pumps and energyrecovery equipment yields reduced capital and operating costs of seawater ROsystems, Desalination 127 (1999) 181–188.

[2] Y. Wang, S. Wang, S. Xu, Experimental studies on dynamic process of energyrecovery device for RO desalination plants, Desalination 160 (2004) 187–193.

[3] W.T. Andrews, D.S. Laker, A twelve-year history of large scale application of work-exchanger energy recovery technology, Desalination 138 (2001) 201–206.

[4] G. Kronenberg, The Ashkelon 100 MCM/year BOT project, Desalination 152 (2002)103–112.

[5] B. Schneider, Selection, operation and control of a work exchanger energy recoverysystem based on the Singapore project, Desalination 184 (2005) 197–210.

[6] J. Sun, Y. Wang, S. xu, S. wang, Energy recovery device with a fluid switcher forseawater reverse osmosis system, Chinese Journal of Chemical Engineering 16 (2)(2008) 329–332.