Use of parabolic trough solar energy collectors for sea-water

Use of parabolic trough solar energy collectorsfor sea-water desalination

Soteris Kalogirou *Department of Mechanical Engineering, Higher Technical Institute, PO Box 423, Nicosia, Cyprus

Abstract

The various desalination methods are analysed with respect to their primary energy con-sumption, sea-water treatment requirement and equipment cost. From this analysis, the mul-tiple-e�ect boiling evaporator is concluded to be the most suitable method for stimulation by

solar energy. The parabolic-trough solar-collector is selected mainly due to its ability tofunction at high temperatures with high e�ciency. The design of the ¯ash vessel and thedesalination system circuit are presented. System modelling is used to predict the rate of fresh

water produced by four sizes of systems, varying from small 10m2 to large 2160m2 collector-area applications. The economic analysis performed, showed that prices of about 0.89 C£/m2

can be achieved with the larger applications. Nevertheless, it is not cost e�ective to operate the

system solely on solar energy due to the relatively high cost of the equipment and the highpercentage of inactive time. # 1998 Elsevier Science Ltd. All rights reserved.

Nomenclature

CR Concentration ratiod Market discount-rate (%)FR Heat-removal factorI Beam solar radiation (W/m2)In Test intercept of collector performance graphN Number of yearsn Thermal e�ciencyno Optical e�ciencyS Test slope of collector's performance graph

Applied Energy 60 (1998) 65±88

0306-2619/98/$Ðsee front matter # 1998 Elsevier Science Ltd. All rights reserved.

PII: S0306-2619(98)00018-X

* Tel.: +357-2-306199; Fax: +357-2-494953; e-mail: [email protected]

Ta Ambient temperature (K)Ti Collector's inlet-temperature (K)Tr Mean receiver temperature (K)UL Steady-state heat-loss coe�cient (W/m2K)�T Temperature di�erence, (Ti ÿ Ta) (K)

Abbreviations

C£ Cyprus pound (August 1998: 1 C£=1.17 £ sterling)CPC Compound parabolic-collectorED ElectrodialysisER±RO Energy recovery±reverse osmosisFPC Flat-plate collectorLCS Life-cycle savingsMEB Multiple-e�ect boiling (evaporator)MES Multiple-e�ect stack (evaporator)MSF Multi-e�ect ¯ash (evaporator)ppm Parts per millionPR Performance ratioPTC Parabolic-trough collectorPW Present worthPWF Present worth factorRO Reverse osmosisTDS Total dissolve solidsVC Vapour compression (evaporator)

1. Introduction

Water is one of the most abundant resources on Earth, covering approximatelythree-quarters of the planet's surface. About 97% of the Earth's water is salt waterin the oceans: 3% of all fresh water is in ground water, lakes and rivers, which sup-ply most of that needed by humans and animals.Water is essential to life. The importance of supplying potable water can hardly be

overstressed. Man has been dependent on rivers, lakes and underground water-reservoirs for fresh-water requirements in domestic life, agriculture and industry.However, rapid industrial-growth and the population explosion world-wide haveresulted in a large escalation of the demand for fresh water. Added to this is theproblem of pollution of rivers and lakes by industrial wastes and the large amountsof sewage discharged. On a global scale, man-made pollution of natural sources ofwater is becoming the single largest cause for fresh-water shortages [1]. Besides theonly inexhaustible sources of water are the oceans. Their main drawback, however,is the high salinity of such water. It would be attractive to tackle the water-shortageproblem with desalination of this water, which may be mixed with brackish water to

66 S. Kalogirou/Applied Energy 60 (1998) 65±88

increase the amount of fresh water and reduce the concentration of salts to around500 ppm [1].Solar distillation has been practised for many generations. According to Malik et

al. [1], the earliest documented work is that of an Arab alchemist in the 15th century,as reported by Mouchot in 1869. Mouchot stated that the Arab alchemist had usedpolished Damascus mirrors for solar distillation. The great French chemist Lavoisier(1862) used large glass lenses, mounted on elaborate supporting-structures, to con-centrate solar energy on the contents of distillation ¯asks [1]. The use of silver- oraluminum-coated glass re¯ectors to concentrate solar energy for distillation was alsobeen described by Mouchot.Solar stills were the ®rst to be used in large-scale, distilled-water production. The

®rst water-distillation plant constructed was a system built at Las Salinas, Chile, in1874 [1,2]. The still covered 4700m2 and produced up to 23 000 litres of fresh waterper day (4.9 litre/m2) in clear sky conditions. The still was operated for 40 years andonly abandoned after a fresh-water pipe was installed to supply water to the areafrom the mountain region.The use of solar concentrators in solar distillation was reported by Pasteur (1928) [1],

who used a concentrator to focus solar rays onto a copper boiler containing water. Thesteam generated from the boiler was piped to a conventional water-cooled condenserin which distilled water was accumulated. Renewal of interest in solar distillationoccurred soon after the First World War, during which several new devices had beendeveloped, such as the roof-type, tilted-wick, inclined-tray and in¯ated stills.Desalination can be achieved by using several techniques. These may be classi®ed

into the following categories: (i) phase-change or thermal processes and (ii) mem-brane or single-phase processesÐsee Table 1 . In the phase-change or thermal pro-cesses, the distillation of sea water is achieved by utilising a heat source. The thermalenergy may be obtained from a conventional fossil-fuel source, nuclear energy orfrom a non-conventional solar-energy source. In the membrane processes, electricityis used either for driving high pressure pumps or for ionisation of salts contained inthe sea water.Desalination processes require signi®cant quantities of energy to achieve separa-

tion. This is highly signi®cant as it is a recurrent cost which few of the water-short

Table 1

Desalination processes

Phase-change processes Membrane processes

1. Multi-e�ect ¯ash (MSF) 1. Reverse osmosis (RO)

2. Multiple e�ect boiling (MEB) - RO without energy recovery

3. Vapour compression (VC) - RO with energy recovery (ER±RO)

4. Freezing 2. Electrodialysis (ED)

5. Solar stills

- conventional stills

- special stills

- wick-type stills

- multiple-wick-type stills

S. Kalogirou/Applied Energy 60 (1998) 65±88 67

areas of the world can a�ord. Many countries in the Middle East, because of oilincome, have enough money to invest and run desalination equipment. People inmany other areas of the world have neither the cash nor the oil resources to allowthem to develop in a similar manner. According to Marinos et al. [3] and Morris andHanbury [4], the installed capacity of desalinated water systems in 1990 reached 13million m3/day, which, by the year 2000, is expected to double. The dramaticincrease in desalinated water supply will create a series of problems, the most sig-ni®cant of which are those related to energy consumption. It has been estimated thata production of 13 million m3 of portable water per day requires 130 million tons ofoil per year. Even if oil were much more widely available, could we a�ord to burn iton the scale needed to provide everyone with fresh water? Given the current under-standing of the greenhouse e�ect and the importance of CO2 levels, this use of oil isdebatable. Thus, apart from satisfying the additional energy-demand, environmentalpollution would be a major concern. If desalination is accomplished by conventionaltechnology, then it will require the burning of substantial quantities of fossil fuels.Given that conventional sources of energy are polluting, sources of energy that arenot polluting will have to be used. Fortunately, there are many parts of the worldthat are short of water but have exploitable renewable-energy sources that could beused to drive desalination processes.Solar desalination is used in nature to produce rain, which is the main source of

fresh-water supply. Solar radiation falling on the surface of the sea is absorbed asheat and causes evaporation of the water. The vapour rises above the surface and ismoved by winds. When this vapour cools down to its dew point, condensationoccurs and fresh water precipitates as rain. All available man-made distillation sys-tems are small scale duplications of this natural process.Solar energy can be used for sea-water desalination either by producing the ther-

mal energy required to drive the phase-change processes or by generating the elec-tricity required to drive the membrane processes. Solar-desalination systems are thusclassi®ed into two categories, i.e. direct and indirect collection-systems. As theirname implies, direct-collection systems use solar-energy to produce distillate directlyin the solar collector, whereas in indirect collection systems, two sub-systems areemployed (one for solar-energy collection and one for desalination). Conventionaldesalination systems are similar to solar systems because the same type of equipmentis applied. The prime di�erence is that in the former, either a conventional boiler isused to provide the required heat or mains electricity is used to provide the requiredelectric power, whereas in the latter, solar energy is applied.

2. Desalination-process selection

During the design e�ect, there is a need to select a process suitable for a particularapplication. The factors to be considered during such a selection are:

1. Suitability of the process for solar-energy application.2. The e�ectiveness of the process with respect to energy consumption.


3. The amount of fresh water required in a particular application in combinationwith the range of applicability of the various desalination-processes.

4. The sea-water treatment requirements.5. The capital cost of the equipment.6. The land area required, or could be made available, for the installation of the

equipment.

Solar energy can generally be converted into useful energy either as heat, withsolar collectors and solar ponds, or as electricity, via photovoltaic cells. Both meth-ods have been used to power desalination systems. The direct collection systems canonly utilise solar energy whenever it is available, and their collection is ine�cient.Alternatively, in the indirect collection systems, solar energy can be collected, bymore-e�cient solar collectors, and be in the form of hot water or steam. It should benoted, however, that solar energy is only available for almost half of the day. Thisimplies that the process operates for only half the time available unless some storagedevice is used. The latter, which is usually expensive, can be replaced by a back-upboiler or electricity from the grid in order to operate the system during low-insola-tion periods or during the night. When such a system operates without thermalbu�ering, the desalination sub-system must be able to follow a variable energy sup-ply, without upset.The energy required for various desalination-processes, as obtained from a survey

of manufacturers' data, is shown in Table 2. It can be seen, from Table 2, that theprocess with the smallest energy requirement is RO with energy recovery. But this isonly viable for very large systems due to the high cost of the energy-recovery tur-bine. The next lowest is the RO without energy recovery and the MEB system. Acomparison of the desalination equipment cost and the sea-water treatmentrequirement as obtained from a survey of manufacturers' data, is shown in Table 3.The cheapest of all the systems considered is the solar still. This is a direct-collectionsystem, which is very simple to construct and operate. The disadvantage of thisprocess is the very low yield, which implies that large areas of ¯at ground are required.It is questionable whether such a process can be viable unless cheap desert-like land is

Table 2

Energy consumptions of desalination systems

Process Heat input

(kJ/kg of product)

Mechanical power-input

(kWh/m3 of product)

Prime-energy consumption

(kJ/kg of product)

MSF 294 3.7 338.4

MEB 123 2.2 149.4

VC ± 16 192

RO ± 12 144

ER±RO ± 7.9 94.8

ED ± 12 144

Solar still 2330 0.3 2333.6

Note: Assumed conversion e�ciency for electricity generation=30%.


available near the sea. The MEB system is the cheapest of all the indirect collec-tion systems and also requires the simplest sea-water treatment. RO, althoughrequiring a smaller amount of energy, is expensive and requires a complex sea-water treatment.One alternative, which is usually considered for solar-powered desalination, is to

use an RO system powered with photovoltaic cells. This is more suitable for inter-mittent operation than the conventional distillation processes and has higher yieldsper unit of energy collected. According to Zarza et al. [5], who compared the ROpowered by photovoltaic-generated electricity with MEB plant coupled to parabolictrough collectors:

1. The total cost of fresh water produced by an MEB plant coupled to parabolictrough collectors is less than that of the RO plant with photovoltaic cells, dueto the high cost of the photovoltaic-generated electricity.

2. The highly reliable MEB plant operation makes its installation possible inthose countries with high insolation levels, but lacking in experienced person-nel. A serious mistake during the operation of a RO plant can ruin its mem-branes: these plants must be operated by skilled manpower.

Also, as renewable energy is expensive to collect and store, an energy-recoveryturbine is normally ®tted to recover the energy from the rejected brine stream: thisincreases the RO plant cost considerably. Additionally, in polluted areas, distillationprocesses are preferred for desalination because the water is boiled, which ensuresthat the resulting distilled water is unlikely to contain any harmful micro-organisms.It is believed that solar energy is more e�ectively and cheaply harnessed with heat-

collection systems. Therefore, the two processes that should be considered are theMSF and the MEB. Both systems have been used in various applications coupledwith solar-energy collectors. According to Tables 2 and 3, the MEB requires lessspeci®c energy, is cheaper and requires only a very simple sea-water treatment whencompared with the MSF. In addition, MEB exhibits various advantages whencompared with other distillation processes. According to Porteous [6], these are asfollows:

Table 3

Comparison of desalination plants

Item MSF MEB VC RO Solar still

Scale of

application

Medium±large Small±medium Small Small±large Small

Sea-water

treatment

Scale inhibitor

anti foam

chemical

Scale

inhibitor

Scale inhibitor Sterilizer Coagulant

Acid Deoxidiser

±

Equipment price

(C£/m3) (1993 prices)

1200±2000 1250±1900 1800±2900 2000±2550

Membrane

replacement every

3±4 years

900±1000


1. Energy economy as the brine is not heated to above its boiling-point as in theMSF process. This leads to inherently less irreversibilities in the MEB processas the vapour is used at the temperature at which it is generated.

2. The feed is at its lowest concentration at the highest plant temperature, so thatscale formation risks are minimised.

3. The feed ¯ows through the plant in series and, as the maximum concentrationonly occurs at the last e�ect, the worst boiling-point elevation is con®ned tothis e�ect.

4. The other processes have high electrical demands, because of the recirculationpump in the MSF or the vapour compressor in the VC systems.

5. MSF is prone to equilibrium problems, which re¯ect themselves in a reductionin PR. In MEB plants, the vapour generated in one e�ect, is used in the nextand PR is not subject to equilibrium problems.

6. Plant simplicity is promoted by the MEB process as less e�ects are required togive a certain PR.

Therefore, the MEB process appears to be the most suitable to be used with solarenergy. The temperature required for the heating medium is between 70 and 100�Cand can be achieved with low-pressure steam.

3. Solar collector design

From the many types of solar collectors developed, three types merit further con-sideration for steam generation: the parabolic-trough collector (PTC), the com-pound parabolic collector (CPC) and the ¯at-plate collector (FPC). The ®rst one is atracking collector, whereas the last two are stationary. PTCs are generally of med-ium concentration ratio (15±40) whereas CPCs are generally of low concentrationratios (1.5±5). The low concentration-ratios of the latter allow them to work withouta need for tracking of the Sun.

3.1. Collector type selection

In general, concentrating collectors exhibit certain advantages as compared withthe conventional ¯at-plate type. The main ones are:

1. The working ¯uid can achieve higher temperatures in a concentrator systemwhen compared with a ¯at-plate system of the same solar-energy collect-ing surface. This means that a higher thermodynamic e�ciency can beachieved.

2. It is possible with a concentrator system, to achieve a thermodynamic matchbetween temperature level and task. The task may be to operate thermionic,thermodynamic, or other higher-temperature devices.

3. The thermal e�ciency is greater because of the small heat-loss area relative tothe receiver area.


4. Re¯ecting surfaces require less material and are structurally simpler than ¯at-plate collectors. For a concentrating collector, the cost per unit area of thesolar collecting surface is therefore less than that of a ¯at-plate collector.

5. Owing to the relatively small area of receiver per unit of collected solar energy,selective surface treatment and vacuum insulation to reduce heat losses andimprove the collector e�ciency are often economically viable.

Their disadvantages are:

1. Concentrator systems collect little di�use radiation, the rate depending on theconcentration ratio.

2. Some form of tracking system is required, so as to enable the collector to fol-low the Sun.

3. Solar-re¯ecting surfaces may lose their re¯ectance with time and may requireperiodic cleaning and refurbishing.

Perhaps their most important advantage is the enhanced thermal-e�ciency andtherefore this is further analysed. The thermal e�ciency of a concentrating collectoris de®ned as the ratio of the useful energy delivered to the energy incident at theconcentrator aperture. This may be calculated from an energy balance on the recei-ver [7,8] which is given by:

n � no ÿ UL Tr ÿ Ta� �ICR

� ��1�

or in terms of the heat-removal factor:

n � FR no ÿUL Ti ÿ Ta� �ICR

� ��2�

From both equations, it can be concluded that the e�ciency of a concentrating col-lector depends on the optical e�ciency (no) which is determined by the opticalproperties of the various materials used in the construction of the collector and themagnitude of the heat losses, as indicated by the second term in Eq. (1). Theadvantage of concentrating collectors is that the heat losses are inversely propor-tional to the concentration ratio (CR). The standard collector-performance can beindicated by the corresponding straight line, whose slope and intercept are thenindications of performanceÐsee Eq. (3):

n � In ÿ S�T

I

� ��3�

where In � Intercept � FRno and S � Slope � FRUL� �=CR.The same relations apply to a ¯at plate-collector, in which case CR � 1. The small

heat-loss term in Eqs. (1) and (2) for the parabolic trough collector leads to a small slopeof the typical collector±performance curve, Fig. 1: this does not apply for ¯at-plate


collectors. This means that the e�ciency in the PTCs remains high at high inlet-water temperatures. Therefore, at a temperature of 100�C, which occurs at a �T=Ivalue of about 0.1, PTCs work at an e�ciency of about 62%, CPCs at about 32%and the FPC at about 10%. This clearly suggests that the PTC is the best type ofcollector for this application.

3.2. Parabolic-trough collector design

Parabolic trough collectors are employed in a variety of applications,includingindustrial steam production [9] and hot-water production [10]. These are preferredfor solar steam-generation because, as was seen above, high temperatures can beobtained without any serious degradation of the collector e�ciency. In this paper,PTCs are used for puri®ed-water production by producing the steam used to powera MEB evaporator. The design of the parabolic trough collector system is detailed inKalogirou et al. [11] and Kalogirou [12]. Four sizes of applications are analysedhere, with aperture area, varying from 10 to 2160m2. The speci®cations of thecollector are shown in Table 4. The same collector characteristics are applicable toall the collector sizes employed.

4. Design of the steam-generation method

Three methods have been employed to generate steam using parabolic-troughcollectors: [9]

1. The direct or in-situ concept in which two-phase ¯ow is allowed in the collectorreceiver, so that steam is generated directly.

Fig. 1. Typical collector±performance curves.


2. The steam-¯ash concept, in which pressurised water is heated in the collectorand then ¯ashed to steam in a separate vessel.

3. The un®red-boiler concept, in which a heat-transfer ¯uid (e.g. therminol 55) iscirculated through the collector and steam is generated via heat-exchange in anun®red boiler.

These steam-generation methods are analysed here with respect to the system'ssimplicity, capital-cost and stability.

4.1. Selection of the steam-generation method

A diagram of a steam-¯ash system is shown in Fig. 2. Water, pressurised to pre-vent boiling, is circulated through the collector and then ¯ashed across a throttlingvalve into a ¯ash vessel. Treated-feedwater input maintains the level in the ¯ashvessel and the subcooled liquid is recirculated through the collector. The direct or in-situ boiling concept, shown in Fig. 3, uses a similar system con®guration without a¯ash valve. Subcooled water is heated to boiling and steam forms directly in thereceiver tube. Capital costs associated with a direct-steam and a ¯ash-steam systemwould be approximately identical [13].Although the steam-¯ash system uses water, a superior heat-transport ¯uid, the

in-situ boiling system is more advantageous. The ¯ash system uses a sensible heatchange in the working ¯uid, which makes the temperature di�erential across thecollector relatively high. The rapid increase in water-vapour pressure with tempera-ture requires a corresponding increase in the system's operating pressure to preventboiling. Increased operating temperatures reduce the thermal e�ciency of the solarcollector. Increased pressures within the system require more robust collector com-ponents, such as receivers and piping. The di�erential pressure over the deliveredsteam pressure, required to prevent boiling, is supplied by the circulation pump and

Table 4

Parabolic-trough collector speci®cations

Item Value or type

Collector's aperture-area 10±2160m2

Collector' aperture 1.46m

Aperture-to-length ratio 0.64

Rim angle 90�

Glass-to-receiver ratio 2.17

Receiver diameter 22mm

Concentration ratio 21.2

Collector's intercept factor 0.95

Collector's test intercept 0.638

Collector's test slope 0.387W/m2K

Tracking-mechanism controller type Electronic

Mode of tracking E±W horizontal

Mass ¯ow rate 0.012 kg/s m2


is irreversibly dissipated across the ¯ash valve. When boiling occurs in the collectors,as in an in-situ boiler, the system pressure-drop and consequently, the electrical-power consumption, are greatly reduced. In addition, the latent heat-transfer pro-cess minimizes the temperature rise across the solar collector. Disadvantages of in-situ boiling are the possibility of stability problems [14] and the fact that, even with avery good feedwater treatment system, scaling in the receiver is unavoidable.In multiple-row collector arrays, the occurrence of ¯ow instabilities could result inloss of ¯ow in the a�ected row. This in turn could result in tube dryout with con-sequent damage of the receiver's selective-coating. No signi®cant instabilities werereported by Hurtado and Kast [13] when experimentally testing a single row 120 ftsystem.

Fig. 2. The steam-¯ash steam-generation concept.

Fig. 3. The direct steam-generation concept.


In the un®red-boiler system, shown in Fig. 4, the heat-transfer ¯uid should benon-freezing and non-corrosive: system pressures are low and control is straight-forward. These factors largely overcome the disadvantages of water systems, and arethe main reasons for the predominant use of heat-transfer oil systems in currentindustrial steam-generating solar systems.The major disadvantage of the system results from the characteristics of the heat-

transfer ¯uid. These ¯uids are hard to contain, and most heat-transfer ¯uids are¯ammable. Decomposition, when the ¯uids are exposed to air, can greatly reduceignition-point temperatures, and leaks into certain types of insulation can causecombustion at temperatures that are considerably lower than measured self-ignitiontemperatures. Heat-transfer ¯uids, are also relatively expensive and present apotential pollution-problem that makes them unsuitable for food-industry applica-tions [9]. Heat-transfer ¯uids have much poorer heat-transfer characteristics thanwater. They are more viscous at ambient temperatures, are less dense and have lowerspeci®c-heats and thermal-conductivities than water. These characteristics mean thathigher ¯ow-rates, higher collector di�erential-temperatures, and greater-pumpingpower are required to obtain the equivalent quantity of energy transport whencompared with a system using water. In addition, heat-transfer coe�cients arelower, so there is a larger temperature-di�erential between the receiver tube and thecollector ¯uid. Higher temperatures are also necessary to achieve cost-e�ective heat-exchange. These e�ects result in reduced collector e�ciencies.From the above discussion, it can be said that water-based systems are simpler

and safer for desalination. With proper selection of ¯ow rate and the desalination-system's steam-supply pressure, the pump power can be kept to a minimum. Thisreduces the main disadvantage of the steam-¯ash system against the in-situ system:as their costs are similar, the steam-¯ash system is selected. For a maximum value ofsolar radiation of 1000W/m2, the outlet temperature of the water, for a 100�C inlettemperature (i.e. the pressure in the separator being equal to atmospheric and the¯ow-rate equal to 0.012 kg/s m2) would be 120�C. This is considered a reasonablevalue, not causing the collector to work at excessively high temperatures, and onlyrequires a pressure of 2 bar to avoid boiling.

Fig. 4. The un®red-boiler steam-generation concept.


4.2. Flash-vessel design

In order to separate steam at a lower pressure, a ¯ash vessel is used. This is avertical vessel as shown in Fig. 5, with the inlet for the water located about one thirdof the way up its side. The standard design of ¯ash vessels requires that the diameterof the vessel is chosen so that the steam ¯ows towards the top outlet connection at nomore than about 3m/s. This should ensure that any water droplets can fall throughthe steam (i.e. in contra-¯ow), to the bottom of the vessel. Adequate height above theinlet is necessary to ensure separation. The separation is also facilitated by having theinlet projecting downwards into the vessel. The water-outlet connection is sized tominimise the pressure drop from the vessel to the pump inlet to avoid cavitation. The¯ash valve connected to the vessel inlet is spring loaded for adjustment purposes.In order to maximise the system's steam production, the heat-up energy require-

ments should be kept to a minimum. This is because energy invested in the pre-heating of the ¯ash vessel is inevitably lost due to the nature of the diurnal cycle.The losses during the long overnight shut-down return the vessel to near ambientconditions each morning. This could be readily achieved by optimising the ¯ash-vessel's water inventory and dimensions in order to lower the system's thermalcapacity and losses. The following constraints on the optimisation should be noted,however:

1. The mass of the circulating water contained in the pipes cannot be changed.2. The water inventory in the ¯ash vessel should not be reduced below a certain

level, because the system's performance will deteriorate. This is due to theaddition of make-up water, which is continuously supplied to keep the water

Fig. 5. Flash-vessel design details.


level in the ¯ash vessel constant, would then ``dilute'' the system temperatureand possibly result in instabilities.

The height of the ¯ash vessel should also be kept to a minimum, which in combi-nation with the right steam velocity would avoid the possibility of ``contamination''of the steam with water droplets (i.e. carry-over). Furthermore, a reduced vesselheight and hence a consequent reduction in the system's thermal capacity, will leadto a faster response of the system. In an earlier report [15], the problem of systemoptimisation through variation of the ¯ash-vessel's design was studied in detail. Theoptimal ¯ash vessel design parameters for a system with a collector area of 10m2 areshown in Fig. 5.

5. Design of the desalination system

5.1. System circuit arrangement

The circuit must be able to carry the sea-water from the sea to the MEB evap-orator and return the rejected brine back to the sea. These two streams must beremote from each other to avoid potential mixing problems. The circuit diagram,shown in Fig. 6, gives details of only the intake stream. Whenever possible, theintake from a well next to the coast line is preferred because as the water passesthrough the sand it is ®ltered. The water, after passing through a ®lter is directed tothe MEB evaporator's last e�ect, to cool the steam produced in the previous e�ect.Part of this water is then returned to the sea as warm brine and part as feedwaterdirected to the evaporator's top e�ect after a scale inhibitor is ejected (see Fig. 7). InFig. 6, the solar collectors and the steam-generation system-piping layout is alsoshown. A back-up boiler is also shown in Fig. 6. This is necessary for the operationof the evaporator during days of low insolation and/or during the night. As can beseen from Fig. 6, no complicated controllers are required as reported by Meaburnand Hughes [16]. This is because the steam delivery temperature is constant (i.e.dependent on the evaporator pressure) and the operation of the boiler can be con-trolled by a simple thermostat located at the pipe before the ¯ash vessel. The sameprinciple applies for the operation of the boiler during day-time (back-up of thesolar system) and night-time.

5.2. Evaporator design

Of the various types of MEB evaporators, the Multiple E�ect Stack (MES) type isthe most appropriate for solar-energy application. This features several advantages,the most important of which is the stable operation between virtually zero and100% output, even when sudden changes are made, as well as its ability to follow avarying steam supply without upset. In Fig. 7, a four-e�ect MES evaporator isshown. Sea-water is sprayed into the top of the evaporator and descends as a thin®lm over the horizontally-arranged tube-bundle in each e�ect. In the top (hottest)


Fig.6.Desalinationsystem

arrangem

ent.


e�ect, steam from the solar-collector system condenses inside the tubes. Because ofthe low pressure created in the plant by the vent ejector system, the thin sea-water®lm boils on the outside of the tubes, so creating new vapour at a lower temperaturethan the condensing steam.

Fig. 7. Schematic of the MES evaporator.


The sea-water falling to the ¯oor of the ®rst e�ect is cooled by ¯ashing throughnozzles into the second e�ect, which is at a lower pressure. The vapour made in the®rst e�ect is ducted into the inside of the tubes in the second e�ect, where it con-denses to form part of the product. Again, the condensing warm vapour causes thecooler external sea-water ®lm to boil at the reduced pressure.The evaporation±condensation process is repeated from e�ect-to-e�ect down the

plant, creating an almost equal amount of product inside the tubes of each e�ect.The vapour made in the last e�ect is condensed on the outside of a tube bundlecooled by raw sea-water. Most of the warmer sea-water is then returned to the sea,and a small part is used as feedwater to the plant. After being treated with acid todestroy scale-forming compounds, the feedwater passes up the stack through a seriesof pre-heaters that use a little of the vapour from each e�ect to gradually raise itstemperature, before it is sprayed into the top of the plant. The water produced fromeach e�ect is ¯ashed in cascade down the plant so that it can be withdrawn in a coolcondition at the bottom of the stack. The concentrated brine is also withdrawn atthe bottom of the stack.The MES process is completely stable in operation and automatically adjusts to

changing steam conditions, even if they are suddenly applied, so it is suitable forload-following applications. It is a once-through process that minimises the risk ofscale formation without incurring a large chemical-scale dosing cost. The typicalproduct purity is less than 5 ppm TDS and does not deteriorate as the plant ages.Therefore, the MEB process and in particular the MES-type evaporator appears tobe the most suitable to be used with solar energy.

6. System modelling

The modelling program is used to predict the quantity of the steam produced bythe collector and the ¯ash vessel, and subsequently the amount of desalinated waterproduced by the various systems. The principle of operation of the program is that itemploys the values of the solar radiation and ambient-air temperature from areference year developed previously [17]. The values of the solar radiation are cor-rected hourly for the collector's inclination.In the analysis, a representative day for each month is taken as shown in Table 5.

These are chosen because the value of extraterrestrial solar-radiation is closest to themonth's average for that day [7].

Table 5

Average day of each month

Month Day Month Day Month Day Month Day

January 17 April 15 July 17 October 15

February 16 May 15 August 16 November 14

March 16 June 11 September 15 December 10


In the program, the actual measured collector performance parameters of testslope and intercept are required. These were obtained by testing the collectoraccording to the procedures outlined in ASHRAE Standard 93 [18]. The programtakes into account, in addition to the sensible heat and the thermal capacity of allthe system components, all the heat losses from the system i.e. the ¯ash-vessel body,pipes and pump body. After all these losses are estimated, the ¯ash-vessel's inletwater temperature is determined. From the di�erence in enthalpy of this hot waterfrom that of the water contained in the ¯ash vessel, i.e. the steam production is cal-culated. The accuracy of the simulation depends to a great extent on the validity ofthe reference year. This was investigated when modelling the performance for hotwater production from PTCs [19]. Although the variation reported was 7%, thiscannot be generalised as an expected variation. Details about the structure of theprogram and the validation of the model are given in Kalogirou et al. [20]. The rateof fresh water produced by the desalination system is evaluated by using the eva-porator performance ratio ®gure. Several systems were considered in this study withaperture areas varying from small 10m2 to large 2160m2. The smallest system issuitable for supplying water in a block of 3±4 houses and the largest for a village ofabout 400 persons. The modelled performances of the systems are shown in Table 6.By studying Table 6, it can be seen that the system's performance is in phase with the

weather, i.e. during periods of dry weather (summer) the system's production is at itsgreatest. This is considered to be the most important advantage of solar desalination.

7. System economics

Economic viability studies of the various systems investigated were performed, byusing a life-cycle analysis method. All the parameters required for the economicanalysis, together with the values used, where applicable, are shown in Table 7. Inaddition to these parameters, the amount of water produced in di�erent months of the

Table 6

Modelled performance of PTC desalination systems

Month System production (litres/month)

Area=10m2 Area=60m2 Area=540m2 Area=2160m2

January 31 153 2488 11 197

February 56 341 4672 20 010

March 176 1237 11 452 59 331

April 250 1788 20 503 82 944

May 327 2335 26 510 107 205

June 481 3514 39 113 157 671

July 517 3795 42 139 169 672

August 456 3370 37 493 151 165

September 355 2635 29 497 119 051

October 194 1429 16 453 66 770

November 83 566 7096 29 238

December 39 233 3376 15 541


year is required for which the performance values given in Section 6 are used. The,analysis is performed annually and costs for the following are evaluated: Water cost,mortgage payment, maintenance cost, pumping cost, fuel cost, (if any), tax savings,system annual cost.In equation form, the annual cost for either solar or non-solar systems to meet an

energy need can be expressed as:

System annual cost � mortgage payment�maintenance cost� pumping cost� fuel cost

ÿwater costÿ fuel savingsÿ tax savings

�4�Finally, the present worth (PW) of the system annual cost is expressed as:

PW � System Annual Cost

1� d� �N �5�

where d � market discount rate, N � number of years.The price of water per cubic metre is a variable in the calculation. In the analysis,

this unit price is varied until the LCS of the system gives a value close to zero. This canbe considered as the price of water at which there is no loss or gain from the system.It can also be considered as the water price that the owner of the system couldcharge in order to sell the water without losing money, i.e. the market cost of water.The economic scenario used in the analysis is to pay 30% of the cost of the sys-

tems in advance and the remaining 70% in equal instalments over the life of thesystem. It is also estimated that the system is sold at the end of its life at 30% of theinitial cost (i.e. its resale value).The mortgage payment is the annual sum of money required to cover the funds

borrowed at the beginning to install the system. This includes interest and principalpayments. The estimation of the annual mortgage payment can be found by dividingthe amount borrowed by the Present Worth Factor (PWF). The PWF is estimated

Table 7

Parameters, a�ecting the chosen design

Item Value Units

Collector area Depending on the case m2

Area-dependent cost Taken from Table 8 C£/m2

Area-independent cost Taken from Tables 8 and 9 C£

Period for economic analysis 20 years

Market discount rate 7.84 %

Maintenance in year one 2 %

Annual increase in maintenance 2 %

Total pumping-power Taken from Table 9 kW

Price of electricity 0.04 C£/kWh

Annual increase in electricity price 2.9 %

First-year fuel cost Depending on the case C£

First-year fuel savings Depending on the case C£

Fuel cost annual-increase 0.6 %

Resale value relative to original value 30 %


by using the in¯ation rate equal to zero (i.e. equal payments) and with the marketdiscount rate equal to the mortgage interest rate taken as (9%).In Cyprus, the law allows a 20% investment allowance as an incentive to owners

to build up new businesses and an annual 10% wear-and-tear allowance for tenyears. Both allowances are calculated with respect to the total investment cost.

7.1. Cost parameters

The cost parameters of the various systems investigated are divided into twocategories, one for the collector and one for the desalination system. The costs of thecollector systems are tabulated in Table 8, together with the costs of the circulationpump and the ¯ash vessel, as well as the installation labour cost. The costs of thedesalination equipment associated with each application are tabulated in Table 9together with the performance ratio (PR) of each evaporator, the costs of otherauxiliary equipment, and labour costs to install the plant and the piping from the seato-and-from the evaporator. The desalination plant capacity is determined by themaximum steam capacity of the solar system.

7.2. Economic analysis

The economic analysis is performed for three types of operation mode. The ®rstone is for the system operated solely with conventional fuels, the second is for thesystem operated solely with solar energy and the third is a combination of the two,

Table 8

Parabolic-trough collector system costs

Cost (C£)

Item (10m2) (60m2) (540m2) (2160m2)

Parabola 160 960 Nine Four times

Re¯ective material 140 840 times the the costs

Receiver 28 168 costs of of 540m2

Framework 50 266 60m2

Labour cost 92 606

Sub-total 470 2840 25 560 102 240

Design supervision, overheads and pro®t 141 852 6390 20 448

Tracking mechanism 150 350 3150 12 600

Piping, ®ttings and insulation 20 300 2440 11 100

Electrical installation 10 50 400 1500

Total 791 4392 37 940 147 888

Area dependent cost (C£/m2) 79.1 73.2 70.26 68.47

Pump 200 250 350 745

Flash vessel 30 150 270 555

Installation labour-cost 100 500 3000 10 000

Grand total 1121 5292 41 560 159 188


i.e. system operation with solar energy during day-time and conventional fuel duringnight-time. Diesel fuel in Cyprus is subsidised by the Government. The normal priceof such fuel is double the today's price: therefore, in addition to the normal waterprice (calculated when using the subsidised fuel price), the water price for the non-subsidised fuel cost is calculated in the present analysis. The results of the economicanalysis are shown in Table 10, from which it can be seen that the operation of thesystem solely with solar energy is not cost e�ective mainly due to the high cost of therequired equipment and the high percentage of the inactive time. The fuel only sys-tem gives somewhat better results for the small-area applications, whereas the sameis correct for the combined system and the large-area applications. It should benoted though that a signi®cant di�erence occurs in the cases where non-subsidisedfuel cost is considered in the analyses.

Table 9

Desalination system-cost parameters (from manufacturers' data)

Item Collector area (m2)

10 60 540 2160

Desalination equipment data

Maximum daily-water production (m3) 1.3 10 110 430

Performance ratio 8 10 12 12

Electrical power (kW) 0.2 0.9 8.7 36

Desalination system costs (in C£)

MEB Evaporator 2280 16 200 152 000 585 000

Piping 50 100 300 400

Pumps 100 250 400 800

Electrical installation 100 150 200 300

Labour cost to install plant and pipes 400 600 1000 1500

Boiler cost with auxiliaries 600 600 2280 5620

Total 3530 17 900 156 180 593 620

Table 10

Water prices for the various applications considered

Operation mode Area=10m2 Area=60m2 Area=540m2 Area=2160m2

Sub. No sub. Sub. No sub. Sub. No sub. Sub. No sub.

Fuel only 1.97 2.99 1.09 1.68 0.89 1.39 0.88 1.37

Solar only 6.70 6.70 3.32 3.32 2.43 2.43 2.28 2.28

Combined 2.20 3.00 1.13 1.60 0.89 1.29 0.87 1.28

Water price in C£/m3.

Sub.=water price for subsidised fuel cost.

No sub.=water price for non-subsidised fuel cost.


7.3. Sensitivity analysis

An analysis was carried out to investigate the sensitivity of the economic modeland the e�ects of the variations of the various parameters on the water price. Thebase ®gure is considered to be the full fuel back-up system for the 540m2 case, whichfrom Table 10 gives a value of C£0.89/m3.The e�ects of the area dependent and independent costs, i.e. solar and desalina-

tion systems, are shown in Table 11. The di�erence in water price for 20% variationis about 8%, which is a reasonable ®gure.The doubling of the rate of increase of electricity and fuel price gives 3.4 and 2.2%

increases in the unit water price, respectively. The reduction of the initial payment tozero imposes a reduction of 1.1% on the water price, whereas the elimination of anyresale value leads to an increase of 3.4%. Finally, the modi®cation of using theoriginal analysis with only a 10-year mortgage recovery gives an increase of 1.9%.From the above discussion, it is clear that the water prices given in this section

re¯ect the true costs and their variation, due to possible change in the direct costsand method of payment are relatively insigni®cant.Lately, in Cyprus there is a trend to liberate the bank interest rates. The Govern-

ment declares that this is imposed by the European Union countries in view ofCyprus wishing to join the Union. In such a case, the market discount rate would bereduced to 6%. Such a change will reduce even further the water costs by 1.1%.Further tax bene®ts are provided in sectors of the economy that the Government

wants to promote, like the purchase of automation systems. This bene®t is the returnof 100% wear-and-tear allowance at the end of the ®rst year as if the equipment hasa life of only 1 year. This is an extra incentive for enterprise owners to upgrade theirbusinesses. As Cyprus is potentially facing a water-shortage problem, the authorbelieves that such an incentive should be given to solar desalination as well. This willresult in a reduction of the unit water price by 4.3%.

8. Conclusions

Solar desalination can be viable for the two bigger installations considered. Theunit water cost is insensitive to changes in the method of payment or to variations in

Table 11

E�ects of the desalination and solar cost on water price

Item Percentage di�erence

+20 +10 ÿ10 ÿ20Area dependent cost Price (C£/m3) 0.91 0.90 0.88 0.87

Di�erence +2.2% +1.1% ÿ1.1% ÿ2.2%Area independent cost Price (C£/m3) 0.96 0.93 0.86 0.83

Di�erence +7.9% +4.5% ÿ3.4% ÿ6.7%


direct costs. However, it is not usually worth operating the desalination systemsolely on solar energy due to the high cost of the desalination system and the highpercentage of inactive time.The author believes that even in cases where the fuel only systems result in lower

or equal water prices compared with a solar-plus-fuel system, the solar alternativeshould not be abandoned because as it was proven a possible increase in fuel priceturns the system viability in favour of the solar system. The issues of global warmingand climate change resulting from the increase in greenhouse gases due to theburning of fuels should not be underestimated. According to a world-wide accep-table scenario of human activities, an increase in global average temperature by0.3K per decade will occur [21].Although water prices above the ones charged by the various water boards in

Cyprus (about 0.50C£/m3) have been obtained from the presented analysis, theauthor believes that the present system o�ers some bene®ts which should not beunderestimated. In particular, security of supply is very important to the hotelindustry with Cyprus endeavouring to upgrade its tourist facilities. Most of thehotels in Cyprus have stand-by generators installed for security of electricity supply.Based on the same principle, the author believes that water supply should be treatedthe same way.

References

[1] Malik MAS, Tiwari GN, Kumar A, Sodha MS. Solar distillation. Oxford: Pergamon Press, 1985.

[2] Meinel AB, Meinel MP. Applied solar energyÐan introduction. Reading, MA: Addison±Wesley,

1976.

[3] Marinos D, Assimacopoulos D, Provatas F. The experience of local authorities with water desali-

nation in the Islands of Southern Europe. Proceedings of the New Technologies for the Use of

Renewable Energy Sources in Water Desalination, Athens, Greece, Section V, 1991:48±56.

[4] Morris RM, Hanbury WT. Renewable energy and desalinationÐa review. Proceedings of the New

Technologies for the Use of Renewable Energy Sources in Water Desalination, Athens, Greece,

Section I, 1991:30±50.

[5] Zarza E, Ajona, JI, Leon J, Genthner K, Gregorzewski A. Solar thermal desalination project at the

Plataforma Solar De Almeria. Proceedings of the Biennial Congress of the International Solar

Energy Society, Denver, CO, Vol. 1, Part II, 1991:2270±2275.

[6] Porteous A. Saline water distillation processes. Barking, UK: Longman, 1975.

[7] Du�e JA, Beckman WA. Solar engineering of thermal processes. 2nd ed. New York: John Wiley &

Sons, 1991.

[8] Kreith F, Kreider JF. Principles of solar engineering. New York: McGraw±Hill, 1985.

[9] Murphy LM, May KE. Steam generation in line-focus solar collectors: a comparative assessment of

thermal performance, operating stability, and cost issues. SERI/TR-1311, 1982.

[10] Kalogirou S, Lloyd S. Use of solar parabolic trough collectors for hot water production in CyprusÐ

a feasibility study. Renewable Energy Journal 1992;2(2):117±124.

[11] Kalogirou S, Eleftheriou P, Lloyd S, Ward J. Design and performance characteristics of a parabolic-

trough solar-collector system. Applied Energy 1994;47:341±354.

[12] Kalogirou S. Parabolic-trough collector system for low temperature steam generation-design and

performance characteristics. Applied Energy, 1996;55:1±19.

[13] Hurtado P, Kast M. Experimental study of direct in-situ generation line focus solar collector. SERI,

1984.


[14] Peterson PJ, Keneth E. Flow instability during direct steam generation in a line-focus solar collector

system. SERI/TR-1354.

[15] Kalogirou S, Eleftheriou P, Lloyd S, Ward J. Optimisation of the initial response of a solar steam

generation plant. Proceedings of the 1995 ASME/JSME/JSES International Solar Energy Con-

ference, Maui, HI, 1995:513±520.

[16] Meaburn A, Hughes FM. A simple predictive controller for use on large scale arrays of parabolic

trough collectors. Solar Energy 1996;56(6):583±595.

[17] Kalogirou S. Solar energy utilisation using parabolic-trough collectors in Cyprus. MPhil thesis, The

Polytechnic of Wales, 1991.

[18] ASHRAE Standard 93-1986. Methods of testing to determine the thermal performance of solar

collectors. ANSI/ASHRAE 93, 1986.

[19] Kalogirou S, Lloyd S, Ward J. Modelling of a parabolic trough collector system for hot-water pro-

duction. Proceedings of the ISES Solar World Congress, Budapest, Hungary 1993;5:145±150.

[20] Kalogirou S, Lloyd S, Ward J. Modelling optimisation and performance characteristics of a para-

bolic trough solar collector steam generation system. Solar Energy Journal, 1997;60:49±59.

[21] Houghton J. Climate change: the current state of scienti®c knowledge. Proceedings of the Second

World Renewable Energy Congress, Reading, UK, 1993;1:1±23.


Documents

Use of parabolic trough solar energy collectors for sea-water