DESALINATION

ELSEVIER Desalination 156 (2003) 323-332

www.elsevier.com/locate/desal

Analysis of an innovative water desalination system using

low-grade solar heat
S. Al-Kharabsheh, D. Yogi Goswami*
SolarEnergv and Energy Conversion Laboratory, Mechanical and Aerospace Engineering Department,
University of Florida, 220 MEB, PO Box 116300, Gainesville, FL 3261 l-6300, USA
Tel. +I (352) 392-0812; Fax +I (352) 392-1071; email: [email protected]

Received 7 February 2003; accepted 12 February 2003

Abstract
This paper presents a theoretical analysis and preliminary experimental results for an innovative water desalination
system using low-grade solar heat. The system utilizes natural means (gravity and atmospheric pressure) to create a
vacuum under which water can be rapidly evaporated at much lower temperatures and with less energy than conventional
techniques. The system consists of an evaporator connected to a condenser. The vapor produced in the evaporator is
driven to the condenser where it condenses and is collected as a product. The effect of various operating conditions,
namely, withdrawal rate, depth of water body in the evaporator, temperature of the heat source, and condenser
temperature, on the systemperformance were studied. Numerical simulations and preliminary experimental resultsshow
that the performance of this system is superior to a flat-basin solar still, and the output may be twice that of a flat-basin
solar still for the same input. Vacuum equivalent to 4 Wa (abs) or less can be created depending on the ambient
temperature at which condensation takes place.

Keywords: Low-grade solar heat; Solar desalination; Vacuum; Solar still; Desalination

1. Introduction small fraction, about 0.3%, ofthe available water

resources is available as fresh water [l]. A
Fresh water resources are under heavy
drinking water shortage is expected to become
pressure due to population growth rates and
one of the biggest problems facing the world. To
pollution caused by industrial wastes. A very
compensate for this, desalination of saline water
appears to be the best solution, since the only
*Correspondingauthor. inexhaustible source of water is the ocean.

Presented at the European Conference on Desalination and the Environment: Fresh Waterfor All, Malta, 4-8 May 2003.
European Desalination Society, International Water Association.

00 11-9 164/03/!§- See front matter Cp2003 Elsevier Science B.V. All rights reserved
PII: SO01 1-9164(03)00363-l
324 S. Al-Kharabsheh, D. Yogi Goswami /Desalination 156 (2003) 323-332

Desalination processes consume significant vacuum distillation without requiring additional

amounts of energy, and many countries in the energy to create the vacuum.
world, particularly those suffering from severe
water shortages, cannot afford the energy re-
quired for desalination. Fortunately, many of 2. Proposed system and its operating principle
those countries lie in areas with high insolation The atmospheric pressure is equivalent to the
rates. Therefore, solar desalination can be a hydrostatic pressure generated by a column of
suitable alternative, provided efficient tech- water of about 10 m high. So, if a column of
nologies are developed to utilize the solar energy height more than 10 m and closed from the top is
in a cost-effective way. Solar energy can be used filled with water and the water is allowed to fall
to produce fresh water directly in a solar still or under the effect of gravity, it will fall to a height
indirectly where the thermal energy from a solar of about 10 m, creating a vacuum in the part
energy system is supplied to a desalination unit. above. The desalination system analyzed in this
A number of efforts have been made to study makes use of the above concept, which was
develop and improve the performance of solar first proposed by Sharma and Goswami [ 1I]. It
desalination systems, particularly solar stills. To consists of a solar heating system and an
increase the temperature of the water inside the evaporator and a condenser at a height of about
still, some researchers [2-4] suggested coupling 10 m above ground level, connected via pipes to
the still to solar collectors. The results showed an a saline water supply, concentrated brine dis-
improvement in the still’s performance. One of charge, and a fresh water tank, all at the ground
the main reasons behind the low efficiency of level. Fig. 1 shows a schematic of the system. A
solar stills, which is about 30-40% [ 11,is the loss vacuum is created by balancing the hydrostatic
of the latent heat of condensation to the environ- and the atmospheric pressures in the supply and
ment and the sensible heat carried away by the discharge pipes.
condensate. The use of latent heat of conden- The evaporator has a provision to supply it
sation to preheat the feed water has shown good with solar or other low-grade thermal energy
improvement in the still’s performance [5,6]. The through a closed loop heat exchanger. The feed
use of latent heat of condensation of one stage to water and concentrated brine discharge pass
evaporate water in another stage, as in multi- through a tube-in-tube heat exchanger in order to
effect stills, has been studied by many re- extract the maximum possible energy from the
searchers showing very good improvement in the hot brine. The evaporator is connected to a
still’s performance [7,8]. Other researchers [9, lo] condenser, which dissipates the heat of conden-
have investigated the concept of evaporation at sation to the environment. Provisions are made
low temperatures under vacuum conditions and for periodically flushing the system and restarting
reported good improvement in the system per- it, thus removing any accumulated non-con-
formance. However, they used vacuum pumps, densable gases.
which require additional energy input to the If two chambers containing water are
system. connected together, water will distill from the
This paper presents a theoretical analysis and higher vapor pressure side to the other. The vapor
preliminary test results of a solar desalination pressure of seawater is about 1.84% less than that
system which utilizes natural forces - atmos- of fresh water over the temperature range of
pheric pressure and gravity - to create a 0-100°C. This means that if a saline water
vacuum. The new concept has the advantages of chamber (evaporator) and a fresh water chamber
 S. Al-Kharabsheh, D. Yogi Goswami /Desalination 156 (2003) 323-332 325

Saline water tank

.._
c
i+csh water tank
Con~enhated brine tank

Fig. 1. Schematic of the experimental

desalination system.

(condenser) are connected from the top while withdrawn water flows in the annulus in the
maintained at the same temperature, water will opposite direction. The heat exchanger area
distill from the fresh water side to the saline should be sized to recover a major part of the
water side. In order to maintain distillation of energy. Under vacuum conditions at the saline
water from the saline water to the fresh water water surface in the evaporator, the supply water
side, the vapor pressure of the saline water must can be injected by the effect of atmospheric
be kept above that of the fresh water by pressure; hence no pumping power is required.
maintaining it at a higher temperature. In the This makes the proposed system of a continuous
proposed process, this will be done by utilizing process type, unlike a flat-basin solar still, which
solar energy. Under vacuum conditions water can is usually a batch process.
be evaporated at a low temperature level, thus
requiring a smaller amount of thermal energy.
This energy may be provided from simple flat- 3. Theoretical analysis
plate solar collectors, which will operate at a
To simulate the system performance, mass,
higher efficiency because of lower operating
energy and salt balances are needed. In applying
temperatures.
those balances, it is assumed that no temperature
Evaporation from saline water increases its
stratification occurs in the system. Application of
salinity, which tends to decrease the evaporation
conservation of mass, solute concentration, and
rate and increase the chances of scale formation.
energy gives the following equations.
So it becomes necessary to withdraw the concen-
?? Conservation of mass:
trated brine at a certain flow rate and inject saline
water at a rate equivalent to the sum of the
withdrawal and evaporation rates. The withdrawn p,P; = p,Pw+p,Ve (1)
brine will be at a high temperature (the evapo-
rator temperature), so it is important to recover ?? Conservation of solute concentration:
energy from it. A tube-in-tube heat exchanger
may be used for this purpose, such that the
supply water flows in the inner tube and the
(2)
326 S. Al-Kharabsheh, D. Yogi Goswami / Desalination 156 (2003) 323-332

?? Conservation of energy:

where the heat transfer coefficients may be cal-

culated from the relations given by Rohsenow et
al. [ 131.Heat loss to the environment is assumed
The heat input to the system is assumed to be
to be due to natural convection only.
in the form hot water from a flat-plate solar
The operating pressure of the proposed unit
collector. Heat of evaporation is calculated as
may be taken as the sum of the pressure in the
given by Bemporad [9] after accounting for the
vapor space at the point of condensation and the
pressure drop between the evaporator and con-
pressure drop occurring in the column. This
denser. The heat of evaporation will mainly be
pressure drop can be calculated by applying
dissipated to the environment via the condenser
continuity equation and the energy equation
during the process of vapor condensation and the
between the inlet and outlet of the column:
rest will be carried away by the fresh water
produced. So the condenser will dissipate the
amount of energy given by:

Q,= rhhf; (4) P out

2
P. v!
Vout
+-+z =A.!! +m +Zin-hL (9)
out
Y 2g Y 2g
where $i is the modified latent heat of
condensation [ 121,
Because of a larger volume of vapor in
vacuum distillation, the tubing should be as short
hi = hrg + 0.68 C,/(T,-T,,) (5) and as wide as possible to avoid any significant
impedance to vapor flow in the connecting pipes.
It is a normal practice to use connecting tubes of
This amount of heat is conducted through the
diameter larger than (l/10) of the evaporator
condenser wall, and eventually transferred to the
diameter, at least in the upper part of the
environment by convection (if we neglect radia-
evaporator, where the lowest pressure prevails
tive transfer). For heat conduction through the
and flooding is most likely to take place [ 141.
condenser wall,

27czckc
(Tci-Tco) 4. Results and discussion
hhi= in(rc,lPci) (6)
For all simulations the system specifications
and dimensions were assumed to be the same as
For convective heat transfer to the ambient, the ones in the actual experimental system. The
the condenser is assumed to be a horizontal tube coil supplying heat to the saline water is a copper
with circular fins. The rate of heat transferred tube of 2.4 m length and 1.27 cm outside
from the condenser (fins and prime surface) may diameter. The evaporator is a cylinder of 0.2 m2
be calculated as cross sectional area, 0.2 m height, with a
 S. Al-Kharabsheh, D. Yogi Goswami / Desalination 156 (2003) 323-332 327

truncated cone attached on top of it. The

evaporator has a provision for feed water supply,
through a 1.27 cm copper tube, enclosed by a
2.54 cm PVC pipe that is used for withdrawing
the concentrated brine. The two pipes form a
tube-in-tube heat exchanger. The condenser is a
4” copper tube of 0.5 m length and 0.25 cm thick-
ness. On its lateral surface, 10 fins of 25.4 cm
diameter and 0.0635 cm thickness are soldered
4 cm apart, The other end of the condenser is
connected to a condensate receiver via a 1.27 cm Fig. 2. Solar radiation and useful heat gain from the
PVC pipe. For theoretical simulations, a collector.
reference state was fixed as 25°C reference
I I
temperature, 3.5% reference solute concentration
and 102 1 kg/m3 reference density. A solar
collector with an efficiency given by,

q = 0.72 - 56AT
(10)
1,
was used. A collector area equal to the evapo-
rator area of 0.2 m* was assumed. For all
calculations the ambient temperature was taken
as 25OC. The solar radiation values were taken Fig. 3. Variation of saline water and collector outlet
for a clear day in July for Gainesville, Florida temperatures with time
(latitude 29.68ON, longitude 82.27”W) for a
collector facing south, tilted at an angle equal to
the latitude. temperature reaches about 6 1.2OC,and the water
Fig. 2 shows the amount of incident solar temperature is about 45.6OC. The corresponding
radiation on the collector and the amount of use- values at a water depth of 0.1 m are 58 and
ful heat supplied by the collector to the system. 43.15 “C, respectively. The peak temperature is
This is shown for different depths of water body reached faster at a smaller depth of water.
inside the evaporator. The smaller the water The variation of output from the system with
depth, the lower the amount of useful energy time for different amounts of water inside the
gain. During the peak of incident solar radiation, evaporator is shown in Fig. 4. The highest
the collector efficiency is about 60%. amount is for the minimum amount of water. The
Variation of water temperature inside the maximum evaporation rate at a water depth of
evaporator and the collector outlet temperature 0.04 m is about 5.1 *lo-’ kg/s; as water depth
for different depths of water body are shown in increases to 0.1 m, this rate is reduced to about
Fig. 3. As is clear from the figure, the smaller the 4.2* 10e5kg/s. The accumulated output for water
depth of water body, the higher the temperature depth of 0.04 m is about 1.3 kg, and this amount
during the peak solar radiation. At a water depth decreases to about 1.l kg at a water depth of
of 0.04 m, the maximum collector outlet 0.1 m.
328 S. Al-Kharabsheh, D. Yogi Goswami / Desalination I56 (2003) 323-332

6e-5 1.4
7‘

1.2
,.,
,I

1.0
a
5

0.8 t!
b.. 0.6 p
‘.

0.4 'j
9
02

0.0

Time (s)
Fig. 4. Systemoutput at differentdepths of water with time.

5e-5

4e-5
------.
_..- .“.” “,-

s? 0.6
P

1e-5

20000

Time (s)

Fig. 5. Systemoutput at differentwithdrawalratewith time.

Fig. 5 shows the variation of the system drawal rate increases to 0.1 kg/h. The accumu-
output at different withdrawal rates with time. As lated output at this withdrawal rate is about
the withdrawal rate increases, the system output 1.l kg. As the withdrawal rate increases to
decreases because the withdrawn water carries an 2 kg/h, the evaporation rate reduces to 3.8*
amount of heat with it as it leaves the evaporator. IO-’kg/s and the accumulated output reduces to
The maximum evaporation rate at a withdrawal about 0.97 kg. At a withdrawal rate of 0.1 kg/h,
rate of 0.05 kg/h is about 4.2*10e5 kg/s; this the loss is low and the possibility of scale for-
amount remains almost constant as the with- mation is minimized.
 S. Al-Kharabsheh, D. Yogi Goswami / Desalination 156 (2003) 323-332 329

4e-5
- 35
s

;
-30 f
b
?6

;
-25 Lc

Accumulated output

- 20
0 10000 20000 30000

Time (s)

//
/ \
\
Fig. 6. Effect of fresh water temperature on the system performance.

/ \
\ --.
4800

// ___l___
I .
4600

4400

4200

4000

3800

3600

3400

3200

3.45 -I 3000
0 10000 20000

Time (s)

Fig. 7. Variation of concentration and pressure with time.

The effect of fresh water temperature on the between the increase in the output and the
system performance is shown in Fig. 6. Different increase in the condenser cost. Based on the
values for fresh water temperature are obtained above simulation for Gainesville, Florida, the
by varying the condenser heat transfer area. As daily output from a system of 1 m* evaporator
the area increases, the fresh water temperature area with a 1 m* solar collector area could reach
decreases, hence more output. For example, if the 6.5 kg, which is almost double the amount from
number of condenser fins is increased from 10 to flat basin solar still, that is about 3-4 kg/day.m*
20, the evaporation rate and the daily accumu- L.151.
lated output will increase from 4.2* 10m5kg/s and Fig. 7 shows the variation of pressure and
1.07 kg to 4.6*10W5kg/s and 1.17 kg, respect- concentration with time. The pressure reaches a
ively. However, a compromise must be reached maximum value of about 4.7 kPa absolute (the
330 S. Al-Kharabsheh, D. Yogi Goswami / Desalination 1.56 (2003) 323-332

3 1.130
0 20 40 60

Time (day)

Fig. 8. Variation of concentration and accumulated output with time.

7o
L 5oo
- 300
z
‘;
P
- 200

0 !
0 5000 10000 15000

Time (s)

Fig. 9. Experimental and theoretical temperature and heat input with time.

effect of non-condensable gases is neglected). accumulated output, which decreases as the

The concentration continues to increase till the concentration increases. This result agrees with
end of the test day. Considering the value at the the result of Keren et al. [ 163,who found that the
end of a day as the initial concentration for the evaporation rate decreases by about 1% far each
next day, a steady-state value of 7.53% will be 1% increase in the salinity.
reached after 53 days, as shown in Fig. 8. The Fig. 9 shows the simulated and experimental
steady-state condition is assumed to be achieved variation of water temperature and heat input
if the increase in the concentration for a day is with time. The experimental results were ob-
less than 0.01%. Also shown in the figure is the tained for the system operating under batch
 S. Al-Kharabsheh, D. Yogi Goswami /Desalination 156 (2003) 323-332 331

process, i.e., no withdrawal or injection, and with system can be operated as a continuous or batch
an initial water body depth of 0.08 m. The heat process type. The results show that the output
source for this test was hot water at 60°C from the proposed system can reach 6.5 kg/d.m*
supplied by an electric water heater. Simulation evaporator area, as compared to 3-4 kg/d.m*
was also carried out for the same experimental from a conventional flat-basin solar still.
conditions. The temperature of the water inside
the evaporator reaches almost the same value of 6. Symbols
44SOC by the end of the test day for both the
A - Area, m*
experimental and simulation results. The maxi-
mum difference between the two is about 2.5”C c - Solute concentration, %
- Specific heat, J/kg.“C
around the mid-point of the test period. The heat CP -
input starts from a high value where the system g Gravitational acceleration, m/s*
h - Convection heat transfer coefficient,
begins to heat up and a major part of the energy
is used to raise the temperature of the water and W/m2.K; height, m
hfg - Latent heat of vaporization, J/kg
the evaporator material, i.e., stored as sensible
I - Incident solar radiation, W/m2
heat. This amount decreases as the system heads
k - Thermal conductivity, W/m.K
towards steady-state conditions. The steady-state
1 - Length, m
energy input reaches about 158 W for the experi-
- Mass flow rate, kg/s
mental test and 150.3 W for simulation. The
p” - Pressure, Pa
experimental accumulated output for the 6-h test
was 0.598 kg as compared to a simulated value of Q - Heat transfer rate, W
T - Temperature, “C
0.675 kg. Towards the end of the test, when the
V - Volume, m3
system reaches steady-state conditions, the
P - Volumetric flow rate, m3/s
hourly output is 0.124 kg as compared to the
V - Velocity, m/s
simulated value of 0.15 kg. The difference
z - Elevation, m
between experimental and simulation results is in
part due to the fact that some of the water vapor
condenses back in the evaporator, whereas in Greek
simulation it was assumed that all vapor will P - Density, kg/m3
reach the condenser. Another reason is that in Y - Kinematic viscosity, m*/s; specific
simulation the heat loss from the system was weight, N/m3
assumed to be due only to natural convection; Q - Efficiency
however, the actual heat loss in the outdoor test
is due to a combination of convection (natural
Subscripts
and forced) and radiation. Besides an uncertainty
analysis showed that the errors in measurements a - Ambient
result in an uncertainty of about 6% for the c - Cold, condenser, collector
hourly distillate output. e - Evaporation
- Of the injection pipe, inside
; - Loss
0 - Outside, outlet
5. Conclusions -
s Of seawater chamber or the evapo-
An innovative water desalination system using rator; surface
low-grade solar heat was studied and tested. The w - Of the withdrawal pipe
332 S. Al-Kharabsheh, D. Yogi Goswami / Desalination 156 (2003) 323-332

References M. Sando, Simulationof a multi-effect solar still and

the static characteristics, J. Chem. Engineering Jpn.,
PI E. Delyannis and V. Belessiotis, Solar energy and 20 (1987) 473-478.
desalination, in: Advances in Solar Energy, An
PI G.A. Bemporad, Basic hydrodynamic aspects of a
Annual Review of Research and Development, D.Y. solar energy based desalination process, Desali-
Goswami, ed., Vol. 14, American Solar Energy nation, 54 (1995) 125-134.
Society, Boulder, Colorado, 2001, pp. 287-330.
WI B.A. Jubran, MI. Ahmed, A.F. Ismai! and Y.A.
PI S. Kumar and G.N. Tiwari, Performance evaluation Aba!car,Numerical modeling of a multi-stage solar
of an active solar distillation system, Energy, 21 still, Energy Conv. Mgmt., 41 (2002) 1107-l 121.
(1996) 805-808.
1111 S.K. Sharma and D.Y. Goswami, Low temperature
c31 S.A. Lawrence and G.N. Tiwari, Theoretical evalu- energy conversion system, unpublished document,
ation of solar distillation under natural circulation University of Florida, 1994.
with heat exchanger, Energy Conv. Mgmt., 30 (1990)
r121 F.P. Incropera and D.P. Dewitt, Fundamentals of
205-213. Heat and Mass Transfer, 4th ed., Wiley, New York,
r41 Y.P. Yadav and L.K. Jha, A double-basin solar still 1996, p. 559.
coupled to collector and operating in the thermo-
u31 W.M. Rohsenow, J.P. Hartuett and Y.I. Cho,
siphon mode, Solar Energy, 14 (1989) 653-659. Handbook of Heat Transfer, 3rd ed., McGraw-Hill,
[51 H. Kunze, A new approach to solar desalination for New York, 1998, pp. 6.31-6.41.
small-and medium- size use in remote areas,
[I41 E. Krell, Handbook of Laboratory Distillation,
Desalination, 139 (2001) 35-41. Elsevier Scientific, Amsterdam, 1982, p. 159.
161 K. Schwarzer, M.E. Vieira, C. Farber and C. Muller,
P51 S. Kalogirou, Survey of solar desa!inaEionsystems
Solar thermal desalinationsystemwith heat recovery, and system selection, Energy, 22( 1) (1997) 69-8 1.
Desalination, 137 (2001) 23-29.
WI Y. Keren, H. Rubin, J. Atkinson, M. Priven and
171 J.L. Femandez and N. Chargoy, Multi-stage, A. Bemporad,Theoretical and experimentalcompari-
indirectly heated solar still, Solar Energy, 44 (1990) son of conventional and advanced soiar pand perfor-
215-223. mance, Solar Energy, 5 l(4) (1993) 255-270.
PI S. Toyama, T. Araga!ci,H.M. Salah, K. Murase and