DESALINATION

ELSEVIER
Desalination 120 (1998) 273-280

Solar desalination with a humidification--dehumidification cycle: performance of the unit

Said A1-Hallaj a*, Mohammed Mehdi Farid b, Abdul Rahman Tamimi c
aDepartment of Chemical and Environmental Engineering, lllinois Institute of Technology, Chicago, IL 60616, USA Tel. +1 (312) 567-5118; Fax +1 (312) 567-6914; email: said@selmanl.che.iit.edu bChemical and Material Engineering Department, The University of AucMand, Auckland, New Zealand CChemical Engineering Department, Jordan University of Science and Technology, Irbid, Jordan
Received 30 June 1998 ; accepted 10 August 1998

Abstract The closed air cycle humidification-dehumidification process was used for water desalination using solar energy. The circulated air by natural or forced convection was heated and humidified by the hot water obtained either from a flat plate solar collector or from an electrical heater. The latent heat of condensation was recovered in the condenser to preheat the saline feed water. Two units of different sizes were constructed from different materials. The productivity of these units was found to be much higher than those of the single-basin stills. Moreover, these units were capable of producing a large quantity of saline warm water for domestic uses other than drinking. No significant improvement in the performance of the desalination units was achieved using forced air circulation at high temperatures. While at lower temperatures, a larger effect was noticed. This can be related to the low heat and mass transfer coefficients at low temperatures and to the non-linear increase in the water vapor pressure with temperature.

Keywords: Solar desalination; Humidification-dehumidification cycle

1. I n t r o d u c t i o n Potable water shortage is expected to be one o f the major worldwide challenges o f the near future. A small-scale, cost-effective conventional solution such as reverse osmosis (RO) or electrodialysis
* C o r r e s p o n d i n g author.

(ED) is not presently available to meet this need. A considerable increase in fresh water demand, coupled with a lack o f fossil energy sources is expected, especially for developing countries experiencing significant population growth like India and China, and also for developed countries which depend on imported oil such as the United

0011-9164/98/$ - see front matter 1998 Elsevier Science B.V. All fights reserved Pll S 0 0 1 1 - 9 1 6 4 ( 9 8 ) 0 0 2 2 4 - 0

274

S. AI-Hallaj et al. / Desalination 120 (1998) 273-280

States. This necessitates designing desalination processes that use renewable energy such as solar energy in the least capital-intensive manner. Saline water desalination plants based on RO, multi-stage flash (MSF) and ED are widely used. These units are well established and well documented, but they are costly and inefficient from the viewpoint of energy consumption. Moreover, these units require high maintenance and skilled operation due to the practical difficulties associated with the high operating temperature, such as corrosion and scale formation.
1.1. Solar stills

Solar stills have been thoroughly studied and tested for the production of desalinated water using solar energy. Many studies discuss in detail the effect of different factors such as solar input, ambient temperature, water depth, and wind velocity, on the performance of the still. For most cases, even under optimized operating conditions, the reported efficiency of the single basin solar still was in the range of 30-45%, with less than 5L/m2.d of fresh water production. This low efficiency is due mainly to the complete loss of latent heat of condensation of water vapor on the glass cover of the solar still [1, 2]. Efforts have been focused for some time on recovering the latent heat of condensation. Preheating of the feed water by passing it over the glass cover allowed only partial utilization of the latent heat resulted in a limited increase in still production [3,4]. More significant improvements in solar still design were achieved through the multiple use of the latent heat of condensation within the still. In a multiple-effect unit consisting of several cells, heat is supplied only to the first cell (effect). Water is evaporated in the second effect as it trickles over a metallic surface heated by the condensation of the vapor from the first effect. This allows the utilization of the latent heat of condensation at different levels. Some investigators [5] reported a

maximum efficiency of 57% for a double effect still, while others reported a production ranging upward to 13 L/m2.d [6]. This increased productivity was achieved by using mirrors as solar reflectors to increase the solar energy received so that the operating temperature could reach 85 C in the multi-effect still. However, if the reflector area is included in the calculation, productivity is still not better than 6 L/m2.d. Multi-effect solar stills may be used for efficient production of desalinated water but only in small capacities because the condenser is an integral part of the still. The low heat and mass transfer coefficients in this type of still require operation at relatively high temperatures and thus the use of large, expensive, metallic surfaces for evaporation and condensation.
1.2. Solar desalination with a humidificationdehumidification cycle

Extensive research has been carried out by different institutions, especially in Germany, to develop an efficient means of utilizing solar energy for water desalination [7]. Small plants, based on the multi-effect humidification process (MEH) were constructed and tested in different countries. In these plants, heat is recovered by air circulation between the humidifier and the condenser using natural draft or forced draft circulation. Heschl and Sizmann [8] designed and tested an MEH unit with an integrated collector, evaporator, and condenser. Air was circulated through the unit by natural draft. Textile heat exchangers were used for efficient evaporation and condensation, with minimum pressure drop. A maximum value of the gained output ratio (GOR), defined as the water temperature increase in the condenser unit over the water temperature increase in the solar collector unit (GOR=3) was reported. This is significantly higher than that of a single still (GOR<I). Plants with induced air circulation have been tested by different investigators using various

S. A l-Hallaj et al. / Desalination 120 (1998) 273-280

275

designs [9]. Although a GOR of 2-8 was reported, the auxiliary energy demand for operating the air fan was appreciable due to high levels of pressure drop in different sections of the unit. Some investigators employed an open-air cycle and obtained good productivity [10]. However, the power required for air circulation was always significant. In this work two sizes of solar desalination units with a humidification-dehumidification cycle were constructed and tested [ 11 ]. The units were at both the bench-scale and a pilot-scale with a 2m 2 solar collector. The units were tested indoors (steady-state mode) as well as outdoors (transient modes). The indoor testing was performed to study the effect of operating parameters such as air velocity, condenser area and design, humidifier area and design, water feed rate and operating temperature on the performance of the unit.

Flat plate solar collector

-.q

Heat input QR

"~

HUMID Warm air

Evaporator

4
g., ~lNE to the sea

i i I
DRY

Condl

~, 1
"~ from the Distillate

Q,

Qr

Fig. I. Schematic diagram of the unit.

2. Experimental The two MEH units were similar in design but different in sizes and material of construction. The pilot unit was constructed from steel plates and the bench unit from Plexiglas sheets. As shown in Fig. I, the two units consisted basically of two vertical rectangular ducts connected at the top through an air fan and the bottom with two separate tanks for desalinated water and brine. This arrangement provided an air closed cycle and open water cycle. Each duct has a cross sectional area of 0.3 m0.05 m and 1.0 m height in the bench unit. The corresponding dimensions in the pilot were 1.0mx0.18m and 2.0m height. The condensers were made of 1.0 m x 0.3 m and 2.0 m x 1.On galvanized steel plates for the bench and pilot units, respectively. A copper tube was welded to the condenser plate in a helical shape. The tube length and outside diameter were 3.0m and 8ram, respectively, in the bench unit and 18.0m and 11 mm, respectively, in the pilot unit. One or two condensers were fixed vertically in one of the two ducts in both units.

The humidifier was a typical cooling tower built of wooden structure and fixed in the second duct. Wooden slats were fixed with a 45 oinclined angle as rows along the tower and were supported on a wooden frame. This arrangement was selected to provide minimum pressure drop with reasonable surface area. The total surface area of the wooden structure per unit volume was 87m2/m 3 and 14m2/m 3 for the bench and pilot units, respectively. The preheated water leaving the condenser tubes was further heated by passing it through a 2.5 kW heater connected to an electrical variac to provide different rates of heating during the indoor steady-state operation. In the normal outdoor operation, a 1.02.0m tubeless fiat plate solar collector was used to supply the pilot desalination unit with the required heat; performance of the solar collector is described elsewhere [12]. The hot water leaving the solar collector (or the electrical heater in the indoor operation) was sprayed over the packing in the humidifier section.

276

S. AI-Hallaj et al. / Desalination 120 (1998) 273-280

A uniform initial water distribution was necessary to insure good water wetting to the packing especially at low water flowrates. A variable speed air fan was fixed at the top of both units to provide different air flowrates through the units in the forced draft operation. Water and air temperatures were measured at different locations in the two units using ChromelAlumel thermocouples connected to a multichannel programmable recorder. Water flowrate was measured using rotameters, while air velocity was measured using an electrical fan speedometer (type Airflow 100). However, the air velocity was too small to be measured, especially in the natural draft operation. It was decided to calculate the air velocity from the energy balance of the unit [13]. Solar radiation was measured with a Kipp and Zenon pyrometer connected to an integrator to give the daily and hourly solar radiation. In the indoor steady-state operations, the unit was usually left running for 3-4h to reach steady state before measurements were taken. Condensate was measured after each hour of operation. In the outdoor operation, condensate and temperature measurements were started early in the morning and stopped late in the evening.

If the feed water temperature, T~, is taken as a reference, then the first two quantities can be calculated easily. Evaluation of the heat losses from the unit requires an accurate estimate of the heat loss coefficient Uloss.The loss coefficient for both units was theoretically calculated from the summation of all the thermal resistances as usually done in heat transfer textbooks [14]. In order to measure Ulo~s experimentally, a continuous flow of hot water was introduced to the condenser to heat the unit for a period o f 3-4 h to reach steady state. The feed water was then bypassed from the condenser out of the unit. The heat lost from the water as it flows through the condenser is equal to the heat losses through the unit walls as described by the following equation:

mwCpw(rl-r2):U,o,,A[(rs+T6)/2-r ]

(1)

3. Overall energy balance in the desalination unit

In order to check out the accuracy of the measurements, an overall energy balance was made on both units. The heat input by the electrical heater or the useful energy obtained from the solar collector (Qu) must equal to the summation of: 1. Heat rejected with the brine leaving the humidifier. 2. Heat rejected with the condensate leaving the condenser (usually negligible). 3. Heat losses from the unit to the ambient through the walls o f the unit. 4. Heat stored in the walls of the unit (only in the outdoor unsteady-state operation).

Air was assumed to transfer the heat uniformly from the condenser to all the unit walls. Air temperature was assumed to vary linearly from the top to the bottom of the unit, and hence the arithmetic average of T5 and T6 was used. The measured and the theoretically calculated Ulosswere 1.0 and 0.6 W/m2.K, respectively, for the pilot unit and 1.2 and 0.8 W/m2.K for the bench unit. The difference between the experimental and the theoretical values of Ulossmay due to the nonlinear uniform thickness of the insulation and the assumption used in Eq. (1). The heat stored in the unit walls was calculated from the estimated heat capacity of the unit walls and the measured variation in their temperature with time. In the steady-state operation this quantity is equal to zero; however, it is very significant in the outdoor operation due to the transient nature of the solar energy. It is usually positive in the morning, an indication of heat storage, and negative in the afternoon, indicating a heat release. In the early morning, most of the useful solar energy is stored as a sensible heat in the unit walls.

S. Al-Hallaj et al. / Desalination 120 (1998) 273-280

277

The energy balance was found correct within 10% for the 72 runs made with both units operated at steady state. The same was found in the outdoor operations. This is considered to be reasonable, particularly for the pilot unit which has large heat capacity walls that may introduce significant error in the energy balance.

4. Performance of the desalination units with indoor steady-state operation

The indoor operation was used to study the performance of the units under steady-state operation. Both the bench and the pilot units were supplied with the required heat from an electrical heater at a preset power. Operation with either natural or forced draft was carried out in both units. Effects of heat load, air velocity, and humidifier operating temperature were studied. Measurements included the use of single and double condensers in both units. Two performance factors were used to define the efficiency o f the desalination units: 1. The performance factor P F 1 which is the ratio of energy utilized for the production of desalinated water to the useful energy input to the unit:
PF1 = m w l ~'/ Qu

water leaving the humidifier may be utilized as a source of hot water for domestic uses. Fig. 2 shows the effect of the ratio of heat load to condenser area (Qu/Ac) on the performance factors PF1 and P F 2 for both units when operated with natural and forced air draft. All the data belong to humidifier inlet temperature, T3, in the range of 60-63 C. The condenser area was varied from 0.6m 2 in the bench unit with single condenser to 8.0 m 2 in the pilot unit with a double condenser. Heat load, Q,, was varied by changing water flowrate and the heating power to maintain almost the same water temperature, T3. Values of P F 1 were found to increase from 0.25 at very large values of Q,/Ac in the bench unit to 1.25 at small values of Q J A c in the pilot unit. The corresponding values o f P F 2 increased from 0.3 to 1.61. Fig. 3 shows no significant improvement in the performance of the desalination units when forced circulation was used. Humidifier inlet water temperature was maintained close to 60C. Fig. 4 shows a significant effect of using forced air

1.8

1.6 1.4

PF1 PF2

(2)
13 1.2
1.0

2. The brine usually leaves the humidifier at moderate temperatures and hence may be used as a source o f energy for night operation for further production. The performance factor may be then defined as the ratio of energy utilized for the production of desalinated water during the two cycles to the energy input to the unit:
PF 2

=a o oo 1

E o.s
a. 0.6

o rlA

o o o D

0.4
o

0.2

Forced
q

Natural

=(mw]+mw2)~./Q.

(3)

o:o

& Qu/Ac(kW/m ';o )

210

The use o f P F l with P F 2 is recommended even if night operation is not practical as the warm

Fig. 2. Performance factors PF~ (day operation) and PF2 (night and day operation) vs. to the ratio o f the heat load to the condenser area (Q,,Mc) (indoor, 7"3=60 C).

278
1.6

S. Al-Hallaj et al. /Desalination 120 (1998) 273-280

r~

Natural Forced

1.4
1.2
I.L

(3.

n rl 1.0
r~

O~

An [3 0.6-

u. 0.8,
G) L)

lii

i
[]

0.4.

4L
D
D

0.2.
I I ' t] Q./A=(kW, m2 )

70C. At low temperature the evaporation rate becomes low, and hence air velocity is expected to have a strong effect on evaporation rate. The strong effect of the temperature of the water sprayed in the humidifier is shown in Fig. 4. This is due to the non-linear increase in the water vapor pressure with temperature. However, it is incorrect to draw a strong conclusion on such effect from the indoor experiment. In the real outdoor operation, the solar collector efficiency is expected to drop with the increase of its temperature. This may prevent or limit the expected increase in the performance factor, as discussed in the following section.

Fig. 3. Effect of using natural or forced air draft on the performance factor PF~ (indoor, single condenser, T3=60C).

5. Performance of the pilot desalination unit with outdoor operation

Typical solar radiation measurements for two days in the month of October in the city oflrbid in Jordan are shown in Fig. 5. The calculated solar radiation on the inclined collector is also shown. Fig. 6 presents the measured hourly production of the unit for the same two days. The production of a single-basin solar still under similar solar radiation is also given [ 15]. The productivity of the MEH unit was significantly higher than that of the single-basin still under the same solar radiation. The MEH was also capable of producing desalinated water soon after the unit was started, unlike that which is usually observed in the singlebasin still. Fig. 7 shows the effect of solar radiation on the daily production of the desalination unit. The production is significantly higher than that of the single-basin still. Production of the MEH unit increased with night operation using the rejected hot water from the humidifier, as shown in the same figure. Extrapolating the results to a typical summer day in Jordan shows that a production of 8 L/m 2.d can be expected. A computer program was developed and used to study the effect of different operating para-

2.0 1.8 1.6

rl t3

T3= 70C Ta= 50C

~ " 1.4 1.2

1:3 LI.

1.0 0.8

C t_

n 3

1~ 0.6 Q)
o. 0.4

0.2 0.0

Forced

Natural
D

olo '012 '01, '016 '018 '11o '112 '114

Qu/A( kW/m z)

1.6

Fig. 4. Effect of the humidifier water inlet temperature (T3) on the performance factor PFI (indoor). circulation at lower temperatures around 50C compared to higher operating temperatures at

S. Al-Hallaj et al. / Desalination 120 (1998) 273-280

1.0 1000. 0.9.

279

0.8"E
800 l

(Irbid,Jordan, O c t 5 , 1 9 9 3 ) (Irbid, Jordan, O c t 2 6 , 1 9 9 3 ) S o l a r Still (Irbtd, Jordan, Oct. 2 6 , 1 9 9 3 )

0.7-

06-"
600.

.~
n, o
tO

~" O.5
0 0 0

~ 0.4
400 0.3

o=
200
--- Horlzordal [ (Id~ld, Jordan, Oct. 2l~ t99:J1

0.2 0.1
0.0
a

Inclined 45 ~ , , i , ' r

17

'

'

'

'

10

11

12

13

14

15

16

17

18

Time(hr)

Time(hr)

Fig. 5. Horizontal and inclined (45 from the horizontal) hourly solar radiation in Irbid, Jordan.

Fig. 6. Hourly fresh water production measurements of the pilot unit. (outdoor).

8.5

s.o7.5-" 7.0-~ 6.5-' 6.0 5.5

Day operation. (C~tobar)

Day end nightoperatiort(Oetober) ~ ~ t "

. "

Solar

etil.(October)

- - - EJq~ectnl results lo t July a n d A u g u s t .

/ / / i / i / / / i / / /"

~E

s.o ~= 4.52 4.0:

% 3.5 ~ 2.52.0 ~
I

/ / / // ~

~"

0""

other hand, the productivity could be increased, though not significantly, by increasing the surface area of the humidifier (5. m 2 for the pilot unit). The effect of feed water temperature (to the humidifier) on the production of the unit was only noticed at temperatures below 50C. This was related to the low heat and mass transfer coefficients at low temperatures and to the nonlinear increase in the water vapor pressure with temperature [ 17].

C~ 3.0-

6. Conclusions
!
!

14

16 18 210 22 Solar Radiation(Mj/mZ.day)

24

26

Fig. 7. Measured and expected daily fresh water production of the pilot unit vs. daily average solar radiation. meters on the performance of the unit [16]. Simulation results showed that doubling the condenser area ( 4 m 2 in the pilot unit) could increase production by 50% (>9L/m2.d). On the

A strong effect of water flow rate on unit production was observed in the outdoor experiments The unit production first increases upon increasing the flow rate to an optimum value. Beyond that value the unit production decreases with increasing water flow rate. This is because increased water flow rate increases both heat and mass transfer coefficients as well as the solar collector efficiency. At the same time it lowers the operating water temperature in the unit and hence,

280

S. A I - H a l l a j e t al. / D e s a l i n a t i o n

120 (1998) 273-280

lowers the evaporation and condensation efficiency. The effect o f air velocity was only noticed at low operating temperatures, while it has a smaller effect at high temperatures. This suggests that these units can be operated with natural air circulation at high temperatures, while at low temperatures a moderate forced air circulation is advantageous. In outdoors experiments the mass o f the unit proved to be another factor that negatively affects the unit performance. A delay of 3 h was noticed between sunrise and the start of production o f fresh water. It was noticed that most of the energy received in these early hours is used as sensible heat to warm up the large mass of the unit, which was about 300 kg. This lag time could be avoided by using a lighter material than galvanized steel for construction. Night operation could increase the unit production and serve to keep the unit at elevated temperature, which enables unit production to start earlier the next day.

T5 T6

---

Uloss - ~, --

Air temperature at the bottom o f the unit, K Air temperature at the top o f the unit, K Overall heat loss coefficient, W/m 2. K Latent heat o f condensation, J/kg

References
[1] M.A. Malik, G.N. Tiwari, A. Kumar and M.S. Sohda, Solar Distillation, Pergamon Press,Oxford, 1982. [2] M.M. Farid and F. Hamad, Renewable Energ, 3(1) (1993) 75. [3] R.A. Akhtamov, B.M. Achilov, O.S. Kamilov and S. Kakharov, Geliotekhnika, 14(4) (1978) 51. [4] G.N. Tiwari, N. Madhuri and H.P. Garg, Energy Conver. Mgmt., 25(3) (1985) 315. [5] H.M. Yeh, L. Ten and L. Chen, Energy, 12(12) (1987) 1251. [6] M.S. Hassan, S. Toyma, K. Murase and M.A. Wahhab, Desalination, 71 (1989) 347. [7] J.H. Graef, Desalination, 82 (1991) 187. [8] O. Heschl and R. Sizmann, Tagungsbericht 5. Intern. Sonnenforum, Berlin, Vol. 1, 1984, p. 375. [9] P. LeGoff, J. LeGoffand M.R. Jedy, Desalination, 71 (1989) 347. [10] Y. Assouad and Z. Lavan, J. Solar Enging., 110 (1988) 14. [11] S. A1 Hallaj, Solar desalination with humidificationdehumidification cycle. M.Sc. Thesis, Jordan University of Science and Technology, Irbid, Jordan, 1994. [12] A. Tamimi and K. Rawajfeh, Energy, 15 (11) (1990) 963. [13] M. Farid, S. AI Hallaj, N. Nawayseh, and A. Tamimi, Solar 95, Conference Proceedings of the 33rd Annual Meeting, Hobart, Tasmania, 1995, pp. 293-306. [14] J.P. Holman, Heat Transfer, 5th ed., McGraw-Hill, New York, 1981. [15] A. Shawaqfeh, Solar desalination of ethanol-water system. M.Sc. Thesis, Jordan University of Science and Technology, Irbid, Jordan, 1993. [16] N.K. Nawayaseh, M.M. Farid, A. Omar, S. AI Hallaj, and A. Tamimi, Desalination, 109 (1997) 277. [17] M.M. Farid, N.K. Nawayaseh, S. AI Hallaj and A. Tamimi, SOLAR95, Hobart, Tasmania, 1995, 293306.

7. Symbols
A
A c

Cp~ mw
mwl mw2

-------

PF 1

--

PF 2

--

Qu Ta T1 T2 T3 T4

-------

Unit surface area, m z Condenser surface area, m 2 Water heat capacity, W/kg.K Inlet saline water flow rate, kg/s Water condensate in day operation, kg Water condensate in night operation, kg Hourly performance factor, day operation Hourly performance factor, night operation Energy input to the unit, W Ambient temperature, K Inlet water temperature (condenser inlet), K Condenser water outlet temperature (heater or collector inlet), K Humidifer water inlet temperature (heater or collector outlet), K Humidifer water outlettemperature, K