Solar diffusion driven desalination for decentralized water production

Desalination 289 (2012) 35–44

Contents lists available at SciVerse ScienceDirect

Desalination journal homepage: www.elsevier.com/locate/desal

Solar diffusion driven desalination for decentralized water production Fadi Alnaimat ⁎, James F. Klausner Department of Mechanical & Aerospace Engineering, University of Florida, Gainesville, Florida, 32611, USA

a r t i c l e

i n f o

Article history: Received 23 August 2011 Received in revised form 27 December 2011 Accepted 28 December 2011 Available online 1 February 2012 Keywords: Desalination Humidification and dehumidification Solar energy Transient Evaporation Condensation

a b s t r a c t This paper examines the operation of the solar diffusion driven desalination process (humidification–dehumidification) under dynamic operating conditions. The solar heat input is recycled in a unique dynamic mode so that it does not require an external source of cooling water. A detailed analytical investigation that is based on numerical simulation suggests that this process can potentially produce 100 L/day distilled water with an average specific electric energy consumption as low as 3.6 kWh/m3 using a total of eight 2 m 2 flat plate solar collectors. Water production and energy consumption have been investigated under various design and operating conditions. A unique operating mode has been explored to reduce the specific energy consumption. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Most conventional desalination plants are large scale centralized units that typically serve urban populations. In recent years, there is considerable interest in developing decentralized desalination technologies. The utilization of solar energy to drive desalination should focus on decentralized units since the energy source is spread out and is prohibitively expensive to concentrate. An environmental advantage of decentralized desalination is that the brine discharge is spread out over a large area, and thus the environmental impact is considerably less than that associated with large scale centralized desalination plants. In rural arid regions, populations are distributed over a large land mass. For such cases, it is more economical to install and operate decentralized water production units that serve the local population in lieu of large centralized water production where water must be transported long distances. The rural arid regions typically have excellent solar resources, and thus solar driven desalination is appealing. Nelson [1] argued that the development of decentralized water treatment technologies is essential to assuring an adequate future water supply and healthy ecology. Gocht et al. [2] observed that there is a noticeable lack of experience operating decentralized desalination units, especially those powered by renewable energy. In order for such technologies to gain traction and become viable in the marketplace, it is essential that the operating performance and

⁎ Corresponding author. Tel.: + 1 352 392 (1086); fax: + 1 352 392 (1071). E-mail address: falnaimat@ufl.edu (F. Alnaimat). 0011-9164/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.desal.2011.12.028

maintenance experience be gained with such units, especially when driven by a highly transient energy source such as solar. In order to meet the demand for decentralized desalination, there have been numerous investigations into solar desalination with Photovoltaic Reverse Osmosis (PV-RO) systems [3]. Mohamed et al. [4] reported that small-scale seawater reverse osmosis (SWRO) systems typically do not employ energy recovery devices due to their expense, and the specific energy consumption runs as high as 20 kWh/m 3. With an energy recovery system and no fouling, the specific energy consumption can be as low as 3.3 kWh/m 3. Ghermandi et al. [3] have suggested that long term reliability and performance of PV-RO plants have not been adequately tested, especially since extensive pretreatment and maintenance is required. The water cost reported for such systems ranges from 8 to 25 $/m 3. Therefore, there is interest to investigate less expensive and less maintenance intensive options for solar driven sea and brackish water desalination. The solar still is one of the oldest and by far the simplest water desalination method. A solar still consists of a structural element to collect solar energy for evaporation of saline water. Solar stills can be passive or active, when water circulation is needed. The main advantages of the passive solar still are that it does not require electrical energy for pumping (passive solar collector), it is simple, and it is easy to operate. However, the main drawback of the solar still is that it typically has low water production due to the loss of latent heat of condensation through the solar still glass cover. There exist many types of solar stills, including single slope, double slope, single and double basin, inverted, tubular, spherical, double effect multi wick, and greenhouse integrated solar stills. Many investigators [5–10] have attempted various methods to improve the efficiency of solar stills. Yadav et al. [5] considered water

36

F. Alnaimat, J.F. Klausner / Desalination 289 (2012) 35–44

flow through the basin instead of a single fill per day. It was concluded that the fresh water production rate increases with increasing flow rate in the basin up to an optimum value. Tiwari et al. [6] considered the effect of cooling water flowing over the still glass cover and found that distillate production increases with cooling water flow rate. Other researchers [7,8] considered the use of materials such as black rubber, gravel and aluminum sheet to store a higher amount of thermal energy in the basin. These studies suggested an improvement in distillate production when materials are added to the basin. Distillate production improvement in the Valsaraj [7] study ranged from 6.5 to 50% and in the Sakthivel et al. [8] study it ranged from 17 to 20%. Sodha et al. [9] showed that the overall efficiency of a multiple wick solar still is 4% higher than that of the simple basin configuration. Kumar et al. [10] investigated the performance of a double basin solar still and showed that its performance is 52% higher than the single basin due to better utilization of latent heat of vaporization. Solar driven humidification–dehumidification (HDH) desalination has been investigated as an alternative to common desalination systems for decentralized water production. Due to the dynamic operation of solar HDH desalination systems and the dependence of the performance on the available solar flux and collector area, it is difficult to directly compare the performance of different HDH systems. Younis et al. [11] studied the performance of an HDH system with feed water extracted from a solar pond. It was demonstrated that the fresh water production rate strongly depends on the air mass flow rate and is weakly dependent on the feed water mass flow rate. Farid et al. [12] studied the performance of a solar driven humidification–dehumidification system and reported a daily fresh water production rate of 12 L/m 2collector.day. No information on the incident solar flux was reported other than experiments were done on a “typical day” in Basrah, Iraq. Al-Hallaj et al. [13] studied a solar driven HDH desalination system operating in Irbid, Jordan during the month of October, and the daily fresh water production rate ranged from approximately 2.25 to 5.0 L/ m 2collector.day, depending on the average daily solar flux. Muller-Holst et al. [14] fabricated a solar HDH desalination system for operation in Munich, Germany. The system performance showed a strong seasonal variation. The average daily fresh water production in June was approximately 7.5 L/m 2collector.day while that in January was approximately 1.2 L/m 2collector.day. Dai et al. [15] tested a solar driven HDH desalination system where the saline water is partially re-circulated and the air is drawn from and discharged to the environment. Rather than using solar heating, an auxiliary heater was used to heat the feed water. Fresh water production rates as high as 64 kg/h are reported. Orfi et al. [16] described a unique solar driven HDH desalination system where both the saline water and air were heated within solar collectors. The evaporator consists of water and air channels separated by spongy absorbent material, and the air and water flow counter-currently through the channels. The experimental performance is difficult to judge since the water solar collector was replaced by an electric water heater. Nafey et al. [17] also investigated a HDH desalination facility where both the water and air use solar heating in Suez City, Egypt. A parabolic trough concentrator is used for water heating and a flat plate collector is used for air heating. A fresh water production rate of 10.7 L/day was achieved in July, and the production rate dropped to 5 L/day in November. While all of these prior investigations have made significant contributions toward the development of solar driven HDH desalination, the electric specific energy consumption was not reported. The electric specific energy consumption is an important performance measure, especially when comparing with PV-RO desalination systems. Yamali et al. [18] theoretically investigated the effect of different system operating conditions on a solar humidification dehumidification desalination system. The system used is based on a closed water open air cycle, where the air is heated using a double-pass solar air heater. It was suggested that increasing humidifier water

and air flow rate increases water production. Higher initial water temperature and water volume in a saline storage tank resulted in a significant improvement in water production. It was also found that increasing the cooling water flow rate and decreasing its temperature results in enhancement of water production. Ettouney [19] investigated different humidification dehumidification process configurations for fresh water production enhancement. In the study, the implementation of a conventional condenser, vapor compressor, desiccant air dryer, and membrane dryer were investigated. It was concluded that the main drawback of the humidification dehumidification desalination process is the presence of a large air/ water vapor stream that needs to be dried. This results in reduced system efficiency. In a recent theoretical study by Mohamed et al. [20], the operation of a humidification dehumidification process using a concentrating solar parabolic trough was investigated. A condenser cooling water temperature of 10 °C was assumed. The study covered 4 seasons of the year. A comparison study was presented to show the effect of the different parameters (collector outlet water temperature, collector thermal efficiency) for different seasons of the year on the system productivity. It was found that the highest water production is in the summer season and the maximum predicted solar collector outlet temperature is 92 °C. It was concluded that the daily water production is highest in the summer, not only due to the high solar flux, but also due to the longer duration of daytime. This work considers solar diffusion driven desalination, which is potentially low cost, low power consumption, and low maintenance and can be driven entirely from solar resources. Diffusion driven desalination (DDD), a humidification–dehumidification process described by Klausner et al. [21] and Li et al. [22] uses direct contact evaporation and condensation and is adapted for transient operation with solar heating. A laboratory scale facility was fabricated to explore the dynamic thermal and fluid transport associated with the desalination process, and a rigorous transient model for the heat and mass transfer during the desalination process was developed by Alnaimat et al. [23]. The model was exhaustively tested using the laboratory scale facility. The heat and mass transfer model developed by Alnaimat et al. [23] is used to investigate a unique mode of operating the DDD facility where there is no heat discharge during operation so that no external cooling water is necessary for operation. Heat is only discharged to the environment during night hours when the system is idle. It will be demonstrated that the average specific electric energy consumption can be as low as 3.6 kWh/m 3. Thus, there is significant interest to further explore solar diffusion driven desalination as a potentially low cost and low maintenance alternative to PV-RO systems. 2. Dynamic operation of solar diffusion driven desalination A process flow diagram for the solar diffusion driven desalination process is shown in Fig. 1. Saline water is pumped from one of two saline water tanks through solar collectors, and heated saline water is sprayed over high surface area packing material within the evaporator. Countercurrent air is blown through the evaporator where it makes direct contact with a falling liquid film. The heated water evaporates and humidifies the air stream. The humidified air stream is directed counter-currently through the condenser where it makes contact with a falling liquid film of distilled water. The air stream is dehumidified and returned to the evaporator. During normal operation with solar heating, the heat exchanger is by-passed. The discharge saline water from the evaporator is returned to the saline water tank for recirculation. Likewise, the fresh water discharge from the condenser is returned to the freshwater tank where it is re-circulated through the condenser. Since the system is fully closed, it operates without any heat discharge during operation, except for the heat loss from the system components. Heat is recycled through

F. Alnaimat, J.F. Klausner / Desalination 289 (2012) 35–44

Heated Saline Water

37

Cooled Freshwater

Humidified Air

Solar Collector

Condenser

Evaporator

Saline Water Pump Air Blower

Tank Fill

Re-Circulated Saline Water

Saline Water Tank 1 Brine Discharge

Distillate

Freshwater Tank

Tank Fill Freshwater Pump

Saline Water Tank 2

Brine Discharge

Bypass

Heat Exchanger

Saline Water Fill

Fig. 1. Process flow diagram.

the system, and as the whole system increases in temperature, the fresh water production rate increases. Since the evaporator temperature will always be higher than the condenser temperature, there is always a net production of distilled water. 3. Mathematical modeling of dynamic system operation 3.1. Evaporator and condenser The evaporation and condensation process within the evaporator and condenser strongly depends on the inlet water and air temperature, inlet humidity ratio, and water to air flow ratio. Since the distillation process is driven by solar heating, the operation is inherently transient, and a transient heat and mass transfer analysis is required to evaluate its operation. In order to simulate the performance of the system, the transient heat and mass transfer model described by Alnaimat et al. [23] is used. The model accounts for dynamic variations of the water, air, and packed bed temperatures. The assumptions that have been made are: (1) there are no thermal losses from the evaporator to the environment during operation, and (2) both air and water vapor may be treated as perfect gases. The governing equations for the evaporator are given as,

 U L aw T L −T pack ∂T L L ∂T L ∂ω G hfg −hL Ua ðT −T a Þ − − w L ð1Þ ¼ − ρL α L CpL ρL α L ∂z ρL α L CpL ∂t ∂z ρL α L CpL

 hfg ðT L Þ−hv ðT a Þ G ∂ω ∂T a −G ∂T a U G ða−aw Þ þ ¼ − T pack −T a ρa α a ð1 þ ωÞCpmix ∂z ρa α a ð1 þ ωÞCpmix ρa α a ∂z ∂t þ

Uaw ðT −T a Þ ρa α a ð1 þ ωÞCpmix L

ð2Þ ∂T pack ∂t

¼

1 ρpack α pack Cppack




 U L aw T L −T pack −U G ða−aw Þ T pack −T a ð3Þ

  ∂ω kG aw M v P sat ðT i Þ ω P − ¼ Ti 0:622 þ ω T a G R ∂z

ð4Þ

where TL, Ta, Tpack and ω are the water, air, and packed bed temperatures and humidity ratio, respectively. G and L are the respective air and water mass fluxes that are based on the evaporator and condenser cross sectional area. Psat(Ta) is the water saturation pressure corresponding to the air temperature Ta, and Ti is the liquid/vapor interfacial temperature. Eq. (1) describes the dynamic variation of water temperature within the evaporator. It is derived from the conservation of energy equation applied to the water side of the control volume depicted in Fig. 2. The first term on the right hand side of Eq. (1) reflects the change in sensible heat along the z-direction, the second term accounts for the heat transport through evaporation, the third term accounts for the convective heat transport to the packing, and the fourth term accounts for the convective heat transport to the gas/vapor mixture. In a similar manner, Eqs. (2) and (3) are derived to describe the dynamic temperature variation of the air and packed bed in the evaporator. Eq. (4) is derived from the definition of mass transfer coefficient applied to the control volume. See Nomenclature for a description of symbols. Similar governing equations for the condenser are given as,

 ∂T L L ∂T L ∂ω G hfg −hL Uaw ðT a −T L Þ U L aw T pack −T L þ þ ð5Þ ¼ − ρL α L ∂z ρL α L CpL ρL α L CpL ∂t ∂z ρL α L CpL

 hfg ðT L Þ−hv ðT a Þ G ∂ω ∂T a −G ∂T a U G ða−aw Þ T −T ¼ − − ρa α a ∂z ρa α a ð1 þ ωÞCpmix ∂z ρa α a ð1 þ ωÞCpmix a pack ∂t Uaw − ðT −T Þ ρa α a ð1 þ ωÞCpmix a L ð6Þ

∂T pack ∂t

¼

1 ρpack α pack Cppack




 U G ða−aw Þ T a −T pack −U L aw T pack −T L ð7Þ

38

F. Alnaimat, J.F. Klausner / Desalination 289 (2012) 35–44

also used to solve for the temperature history of the fresh water tank. An energy balance applied on the solar collector gives,
 _ w;solar C p T solar;out −T solar;in ; Isolar Asolar ηsolar ¼ m

ð10Þ

where Isolar is the solar irradiation, ηsolar is the flat plate solar collector efficiency and is assumed to be 85%, and Asolar is the solar collector surface area. Tsolar,in and Tsolar,out are the respective temperatures of water entering and exiting the solar collector. It should be emphasized that the solar collector modeling given in Eq. (10) is used for simplicity, and the Hottel–Whillier–Bliss equation can be used for more detailed modeling of the solar collector efficiency. The evaporator inlet water temperature is equal to the water temperature exiting the solar collector, and Eq. (10) gives I Asolar ηsolar T L;evap;in ¼ T solar;in þ solar . _ C m w;solar

p

3.3. System performance measures

Fig. 2. Evaporator or condenser differential control volume depicting liquid/vapor/solid interactions.

In order to evaluate the economic viability of a solar diffusion driven desalination process compared with other solar driven desalination options, it is of interest to evaluate the daily fresh water yield normalized by the solar collector area, as well as the average electric specific energy consumption. The fresh water production rate for the solar DDD process is easily evaluated as _ fw ¼ m _ a ðωin −ωout Þ; m


 ∂ω ∂T a P 2 ω b−2cT a þ 3dT a : ¼ ∂z P−psat ðT a Þ ∂z

ð8Þ

Since the evaporator and condenser governing partial differential equations were formulated in one dimension, closure relations are required for the heat and mass transfer coefficients. Expressions for Ti, Psat(T), Cpmix, kG, kL, UL, and UG are listed in Appendix A. The initial and boundary conditions are applied as follows: Initial conditions: T L;evap ðz; t ¼ 0Þ ¼ T amb ; T a;evap ðz; t ¼ 0Þ ¼ T amb ; T pack;evap ðt ¼ 0Þ ¼ T amb T L;cond ðz; t ¼ 0Þ ¼ T amb ; T a;evap ðz; t ¼ 0Þ ¼ T amb ; T pack;evap ðt ¼ 0Þ ¼ T amb ωevap ðz; t ¼ 0Þ ¼ ωðT amb ; Φ ¼ 100% Þ; ωcond ðz; t ¼ 0Þ ¼ ωðT amb ; Φ ¼ 100% Þ

ð11Þ

_ fw is the fresh water produced in the condenser, m _ a is the air where m flow rate, and ωin and ωout are the respective humidity ratios of air entering and exiting the condenser. The total daily fresh water production is obtained by integrating the fresh water production rate over the duration of daily operation. The energy required to operate the solar DDD consists of both thermal and electrical energy. The thermal energy is obtained exclusively from the solar collectors. Electrical energy is required to circulate the water and air through the DDD facility. The electrical energy consumption rate to pump the water to the top of the evaporator or condenser is computed as, _ L gH ¼ PwL ¼ m

LAc ρ gH: ρL L

ð12Þ

The energy consumption rate required for circulating the air through the evaporator or condenser is calculated as, Boundary conditions: PwG ¼ V G ΔP G ¼

T L;evap ðz ¼ H; t Þ ¼ T L;evap;in ; T a;evap ðz ¼ 0; t Þ ¼ T a;evap;in T L;cond ðz ¼ H cond ; t Þ ¼ T L;cond;in
; T a;cond ðz ¼ 0; t Þ ¼ T a;cond;in

¼ T a;evap z ¼ H evap ; t ωevap ðz ¼ 0; t Þ ¼ ωcond;out ; ωcond ðz ¼ 0; t Þ ¼ ωevap;out :

3.2. Saline water tank, fresh water tank, and solar collector For simplicity, a perfectly insulated saline water tank with a uniform temperature field is assumed. The application of the conservation of mass and energy results in, dT w;tank dt

 _ w;in
m ¼ T w;tank;in −T w;tank ; Mw

_ a ð1 þ ωin Þ m GAc ð1 þ ωin Þ ΔP G ¼ ΔP G ; ρG ρG

ð13Þ

where VG is the gas/vapor volumetric flow rate, and ΔPG is the pressure drop across either the evaporator or condenser, which is due to frictional forces between the air and the packed bed. A correlation provided by the packing manufacturer is calibrated to experimental data and is used to predict the pressure drop across the packed bed, "  2  4 4 # ΔP G G1:4 L G 7 L ¼ 0:054 þ 654:48 þ 1:176  10 ; ρl ρl ρ2G H ρG

ð14Þ

ð9Þ

_ w;in is the where Mw is the total mass of water in the storage tank, m water flow rate entering the water tank. Eq. (9) is solved to capture the temperature history of the saline water storage tank. Eq. (9) is

where ρG (kg/m 3) is the air density, ρL (kg/m 3) is the liquid water density, and H (m) is the height of the packed bed. A comparison between predicted and measured pressure drop across HD Q-PAC polypropylene packing manufactured by Lantec is shown in Fig. 3. The experimental DDD facility reported by Alnaimat et al. [23] is

F. Alnaimat, J.F. Klausner / Desalination 289 (2012) 35–44

150

39

80

15 Solar Heat Input

Exp L=1.12 kg/m 2 .s 2

Exp L=1.97 kg/m .s Model L=1.12 kg/m 2 .s

100

Model L=1.63 kg/m 2 .s Model L=1.197kg/m 2 .s

50

Vsw,tank =V fw,tank =0.75 m 3

10

60

5

40

TL,cond,in

0

0

0

0.5

1

1.5

8

10

TL,evap,in

12

2

Temperature (oC)

Solar Heat Input (kW)

Pressure Drop (Pa/m)

Exp L=1.36 kg/m 2 .s

14

16

18

20 20

Time (hr)

Gas Flux (kg/m2.s)

Esp ¼

PwL þ PwG : _ fw m

ð15Þ

Here the electric specific energy consumption only accounts for the major losses (including circulation through solar collector) through the system but does not include minor losses. 4. Solar diffusion driven desalination performance In order to elucidate the operating characteristics and performance of the solar DDD process, a specific example is considered here. Table 1 lists specific operating conditions for the solar DDD system shown in Fig. 1. It should be noted that the mass fluxes shown are based on the flow cross sectional area. To make the simulation tractable, the following assumptions are made, 1) the temperature in the saline and fresh water tanks is uniform, and mixing occurs instantaneously and 2) at the start of each desalination cycle the entire system is at equilibrium with the ambient surroundings, which has a temperature of 25 °C. With suitable insulation, heat loss to the surroundings can be reduced, and after several cycles of operation the system may start off at a higher temperature than 25 °C, which would improve performance. So assumption (2) is conservative. As Table 1 Operating conditions for solar diffusion driven desalination simulation. Symbol

Description

Value

Levap Lcond G Aevap Acond Hevap Hcond Vsw(t = 0) Vfw(t = 0) Tsw(t = 0) Tfw(t = 0) Asolar N ηsolar _L m _ fw m _a m

Liquid mass flux through evaporator Liquid mass flux through condenser Air mass flux through evap and cond Cross sectional area of evap Cross sectional area of cond Height of evaporator packing Height of condenser packing Initial volume of saline water Initial volume of fresh water Initial saline water temperature Initial fresh water temperature Solar collectors area Number of solar collectors Solar collector efficiency Saline water mass flow rate Fresh water mass flow rate Air mass flow rate

0.75 kg/m2.s 0.75 kg/m2.s 0.75 kg/m2.s 0.36 m2 0.36 m2 1.0 m 1.0 m 0.75 m3 0.75 m3 25 °C 25 °C 16 m2 8 85% 0.27 kg/s 0.27 kg/s 0.27 kg/s

already mentioned, the distillation process is inherently dynamic due to the dynamic nature of the solar heat flux. The solar thermal power used for this simulation is for the city of Jacksonville, Florida, U.S. on a clear day in the month of June with no cloud cover. It is located at 30.23 latitude and 81.68 longitude [24]. The solar thermal power used is a typical one for cities in the southern United States, North Africa, and the Middle East. With respect to the process flow diagram shown in Fig. 1, only one saline water tank (Tank 1) is used during operation, and Fig. 4 shows the variable solar power input to the system over a 12 h period, the saline water temperature response into the evaporator, and the corresponding fresh water temperature response into the condenser. The solar thermal input to the system is based on eight solar collectors; each has an aperture area of 2 m 2 and efficiency of 85%. The total solar thermal input into the system over 12 h of operation is 98.6 kWh. As the solar heat flux decreases, the system still has capacity for desalination due to the stored heat in the system. Further desalination is accomplished by recirculating the fresh water through the heat exchanger. At the same time, ambient saline water (25 °C) is pumped through the heat exchanger and fills the empty saline water tank (Tank 2). Heat is transferred from the fresh water to the saline water which fills the idle tank. During this process, the condenser cools down, and additional fresh water is produced. As shown in Fig. 4, after 8.5 h from the start of operation, saline fill water is directed through the heat exchanger to the idle tank. The feed saline water temperature to the evaporator (from Tank 1) drops as the system is cooled down by the fill water flowing to the idle tank. Operating the solar DDD facility in this manner is denoted as the standard mode of operation. Fig. 5 shows the total fresh water 50

100 Fresh Water Production Rate Fresh Water Produced

Vsw,tank =V fw,tank =0.75 m 3

50

25

0

8

10

12

14

16

18

0 20

Time (hr) Fig. 5. Total fresh water production and production rate; standard mode.

Fresh Water Produced (L)

used to make the measurements. The mean error is 1.7 Pa, and thus Eq. (14) is deemed adequate. The electric specific energy consumption is computed as,

Fig. 4. Solar thermal input, evaporator water inlet temperature, and condenser water inlet temperature; standard mode.

Fresh Water Production Rate (L/hr)

Fig. 3. Comparison between measured and predicted air flow pressure drop through Lantec HD Q-PAC structured packing.

F. Alnaimat, J.F. Klausner / Desalination 289 (2012) 35–44

15

80 Solar Heat Input Vsw ,tank =Vfw ,tank =0.75 m 3

10

60

5

40

TL,cond,in

0

8

10

12

14

Temperature (oC)

production and production rate for the 12 h period with solar heating operating in the standard mode. It is evident that the fresh water production rate increases as the system increases in temperature. Fig. 5 shows that there is a sudden increase in the fresh water production rate at 16.5 h, which is due to the sudden decrease in the condenser inlet water temperature when the fill period begins, as shown in Fig. 4. The sudden decrease in the condenser temperature is for an ideal case; in reality there will be a gradual drop in the condenser water temperature. The gradual drop in the real case is primarily due to the time required to lose/gain the thermal energy stored in the piping and heat exchanger. As a result, the sudden increase in fresh water production rate shown in Fig. 5 will be dampened for the real case. Similar damping is expected for the sudden increases or decreases shown in Figs. 6–9. Fig. 6 shows the electric specific energy consumption during the distillation process operating in the standard mode. Initially the fresh water production rate is low, and thus the specific energy consumption is high. For approximately an 11 h period after the first hour of operation the specific energy consumption is below 5 kWh/ m 3 of fresh water, which demonstrates reasonably good performance. From approximately hours 18–20 the specific energy consumption increases because the production rate decreases with decreasing solar flux. At hour 16.5 when the fill process begins, the specific energy consumption is very low because the fresh water production rate is high, and then the specific energy consumption increases as the fresh water production rate decreases. The average specific energy consumption is 5.2 kWh/m 3 over the 12 h desalination cycle. The average daily fresh water production is 6.3 L/m 2collector.day and compares well with that for MEH for the month of June as reported by Mueller-Holst et al. [14]. The average specific energy consumption compares well with RO systems, especially sea water PV-RO systems in which the specific energy consumption is typically 10–20 kWh/m 3. The total production rate for the solar DDD is 0.1 m 3/day while that for typical PV-RO systems is 1.7 m 3/day [4]. It is noted that the DDD system is easily scaled to achieve higher output. After approximately 12 h of operation, the system is shut down. The brine from Tank 1 is discharged, and for the subsequent distillation cycle, the saline water is fed from Tank 2. The system will sit idle for the next 12 h until the solar flux is again available. During the idle period, heat is lost from the system, and for the purpose of this analysis, it is assumed that the entire system equilibrates back to the ambient temperature, 25 °C. By insulating the system, it should be possible to sustain a higher temperature than the ambient after a number of successive cycles, and the net performance should be enhanced. In order to improve the specific electric energy consumption to drive the desalination process, it is not necessary to run the system when the solar flux is low and the specific energy consumption is

Solar Heat Input (kW)

40

TL,evap,in

16

18

20 20

Time (hr) Fig. 7. Solar heat input, saline water temperature into evaporator, and fresh water temperature into condenser; delayed mode.

high. Therefore, during the first 3 h of operation, the saline water is only circulated through the solar collectors so that it heats up before the distillation cycle begins. The distillation cycle is initiated after 3 h of heating the saline water with the solar collectors. At approximately hour 16.5, the heat exchanger is switched into service and the tank refill process is initiated. The idle saline water tank is refilled in 3.5 h, and the entire system is shut down at hour 20. Operation of the solar DDD facility in this manner is referred to as the delayed mode of operation. For operation in this delayed mode using the same operating conditions listed in Table 1, the transient solar heat input, the saline water temperature into the evaporator, and fresh water temperature into the condenser are displayed in Fig. 7. The fresh water production and production rate are shown in Fig. 8. Fig. 9 shows the corresponding specific electric energy consumption. It is observed that the total daily fresh water production is 100 L/day (6.3 L/m 2collector.day), as is the case for standard operation. However, the specific energy consumption is typically below 5 kWh/m 3, and the average specific energy consumption for the entire distillation cycle is 3.6 kWh/m 3, which is a 30% reduction compared to the standard mode of operation. The system pumps and fan operate for a shorter period of time in the delayed operating mode compared to that of the standard mode. As such, the delayed operating mode is recommended as the best operating mode for the solar diffusion driven desalination system. Clearly, the delayed operating mode provides enhanced performance since the specific energy consumption is reduced. For the operating assumptions listed in Table 1, approximately 8 solar collectors with 2 m 2 collector area will be required to achieve 100 L/

Esp (kW.hr/m3 fw)

8

6

4 Vsw,tank =V fw,tank =0.75 m 3

2

0

8

10

12

14

16

18

Time (hr) Fig. 6. Electric specific energy consumption; standard mode.

20

100

50 Fresh Water Production Rate Fresh Water Produced

Vsw,tank =V fw,tank =0.75 m 3

50

25

0

8

10

12

14

16

18

0 20

Time (hr) Fig. 8. Total fresh water production and production rate; delayed mode.

Fresh Water Produced (L)

Fresh Water Production Rate (L/hr)

10

F. Alnaimat, J.F. Klausner / Desalination 289 (2012) 35–44

41

8

10

5 Avg. Daily Production

7 6 5 4 3 2

Avg. Esp

7

4

6

3

5

Avg. Esp (kW.hr/m3)

Vsw,tank =V fw,tank =0.75 m 3

8

Esp (kW.hr/m3 fw)

Avg. Daily Production (L/m2.day)

9

2 Vsw,tank =V fw,tank =0.75 m 3

1 4

0 8

10

12

14

16

18

20

Time (hr)

Apr

Jul

day water production. For more or less production, the size of the system scales approximately linearly. This aforementioned system performance assumes that the system temperature drops back to the ambient temperature (25 °C) at the completion of each distillation cycle. When it is assumed that there is suitable insulation (especially around the saline water tank) to hold the system temperature at a temperature above ambient for the start each distillation cycle, the system performance will be enhanced. This is because thermal energy storage is achieved. Fig. 10 shows the daily fresh water production and average specific electric energy consumption for different initial system temperatures while the solar DDD system operates in the delayed mode (same operating conditions as in Table 1). It is observed that there is approximately a 10% increase in the daily fresh water production when the initial system temperature increases from 25 to 40 °C. There is also a slight increase in the specific electric energy consumption. It is not possible at this time to simulate the amount of thermal energy storage that can be achieved during the idle period of the desalination cycle. Therefore, it is of interest to experimentally determine the effectiveness of system insulation on a prototype system because it will have a significant impact on the performance. The preceding analysis has focused on the month of June where the solar flux is typically the highest. It is of interest to evaluate the system performance with seasonal variations in the

solar flux. Fig. 11 shows the daily fresh water production and average specific energy consumption for seasonal variations in the solar flux. The performance was evaluated for the months of Jan, Apr, Jul, and Oct. The peak fresh water yield occurs in July and falls to a minimum in January. It is observed that the average specific electric energy consumption does not vary significantly with variations in the season.

5. Experimental validation In order to validate the predictive capabilities of the transient heat and mass transfer model for the solar DDD facility operating with no cooling water, a transient experiment was done. The laboratory experimental facility, reported by Alnaimat et al. [23], was operated with an evaporator water mass flux of 1 kg/m 2s, condenser water mass flux of 2 kg/m 2.s, air mass flux of 1 kg/m 2.s, initial saline water tank volume of 140 L, and dynamic heat input over a 150 min period. After 150 min, the heat is shut off, and cooling water is supplied to the heat exchanger shown in Fig. 1. Fig. 12 shows the water and air temperatures into the evaporator and a comparison between the predicted and measured water and air temperatures discharging the evaporator. Fig. 13 is a similar comparison for the condenser. As observed, the comparison is quite good over the duration of the experiment for both

5

90

7

4

6

4 25

3

Vsw,tank =V fw,tank =0.75 m 3

30

35

2

1 40

Initial System Starting Temperature (oC) Fig. 10. Average daily water production and specific electric energy consumption for various initial system temperatures.

Evaporator Temperature (oC)

T w, in

Avg. E sp

Avg. Esp (kW.hr/m3)

Avg. Daily Production (L/m2.day)

Avg. Daily Production

5

1

Oct

Fig. 11. Seasonal variation of average daily water production and specific electric energy consumption.

Fig. 9. Specific electrical energy consumption; delayed mode.

8

Jan

T w, out, pred

80

T a, in

70

T a, out, pred T w, out, exp

60

T a, out, exp

50 40 30 20

20

40

60

80

100

120

140

160

180

200

220

Time (min) Fig. 12. Comparison of measured and predicted transient evaporator air and water temperatures operating in the standard mode.

42

F. Alnaimat, J.F. Klausner / Desalination 289 (2012) 35–44

90

Condenser Temperature (oC)

T w, in

80

T w, out, pred T a, in

70

T a, out, pred T w, out, exp

60

T a, out, exp

50 40 30 20

0

20

40

60

80

100

120

140

160

180

200

220

Time (min) Fig. 13. Comparison of measured and predicted transient condenser air and water temperatures operating in the standard mode.

the evaporator and condenser, thus giving credibility to the simulated performance summarized in Figs. 4–11. 6. Discussion In order to achieve improved fresh water production and reduced specific electric energy consumption with solar diffusion driven desalination, it is necessary to retain as much thermal energy within the system as possible. Operation in the dynamic mode allows the system to run for a period without any external cooling and achieve good performance. MEH desalination also is able to re-circulate the heat within the system without discharging it to ambient when operating in a dynamic mode. Thus solar MEH and DDD processes give a similar average daily fresh water production. As mentioned earlier, the electric specific energy consumption for MEH has not been reported, although when operating with natural circulation it should be on par with DDD. Al-Hallaj et al. [25] concluded that solar MEH has potential to provide the lowest cost per unit of fresh water production for small scale solar driven desalination. The main difference between DDD and MEH processes is that DDD uses a compact direct contact condenser which is inexpensive and highly efficient. It is possible to integrate the evaporator and condenser into a single enclosure. Alnaimat [26] did a cost analysis for solar DDD and concluded that the distilled water cost is on the order of $4/m 3 when using low cost solar collectors. Economically this compares well with PV-RO which can have costs on the order of $12/m 3. Clearly, there will exist a range of costs depending on the operating conditions, the application, and required maintenance for any of the competing desalination technologies. Therefore, more experience will need to be gained to better understand the cost and performance of competing solar driven desalination technologies. According to Arjunan et al. [27], economic analysis of solar stills is lacking because most researchers have been primarily concerned with increasing the solar still efficiency and productivity rather than the economic viability. Thus, the ultimate cost of product water from solar stills is uncertain. Madani [28] reported that the cost of water produced from solar stills is on the order $3/m 3. Delyannis et al. [29] suggested that the cost of water from solar stills is hardly dependent on plant size, and economies of scale are not realized for solar stills. It was suggested that water production costs for solar stills range from $3 to 12/m 3. According to the recent experimental and theoretical study of Phadatare et al. [30], the maximum distillate output of solar stills is 2.1 L/m 2.day. Gomkale [31] has studied the operation of various

large scale solar stills that range from 50 to 3110 m 2 aperture area, and found that the production rates range from 2.4 to 2.8 L/m 2.day. It is noted that the water production for the solar DDD is three times that for solar stills. While solar stills do not consume electrical energy compared to solar DDD or any other conventional desalination process, they require a larger surface area and volume per unit of fresh water produced. When operating a thermal distillation unit in a steady-state mode of operation, cooling is required to remove the latent heat of condensation. Typically air cooling is inefficient, and water cooling is required. Cooling water is not readily available in dry arid regions. In this study, it is demonstrated that when operating a solar driven thermal distillation unit in the dynamic mode, it is advantageous to contain the latent heat of condensation within the system in order to boost the system operating temperatures. The advantages of this type of operation are that external cooling water is not required and the total fresh water production increases with increasing system temperature. Examination of Figs. 4 and 5 reveals that the highest fresh water production rate coincides with the highest temperature in the saline water tank. In order to minimize the electric specific energy consumption, the distillation process should only be operational when the temperature of the saline water tank is sufficiently high to yield reasonable fresh water production rates. Clearly, this study has shown that operation in the delayed mode significantly reduces the specific electric energy consumption compared with operation in the conventional mode. The aim of the study is to examine the best operating mode for solar diffusion driven desalination systems. It should be noted that there are other parameters that impact the performance of solar diffusion driven desalination systems but were not explicitly examined in the study. For instance, the saline water tank size impacts the distillate production rate since a larger tank has greater thermal capacitance and requires a longer heating time. Proper tank sizing is important to maximize water production and reduce fouling. Air mass flux is an important parameter that impacts the operation of the solar DDD process. While a high air mass flow rate results in heat transfer enhancement, it also increases the pressure drop through the packed bed with a corresponding increase in the specific energy consumption. In general it has been found that it is preferable to operate with lower air mass flux, which can be accomplished via increasing the cross sectional area of the packed bed. This translates to greater fabrication cost, but that cost is recovered quickly with reduced specific energy consumption and operating cost. Furthermore, there are other types of packed bed materials and structures that have not been examined in this study. An interesting future study would examine the water production rate and cost with various types of packed beds. The best packed bed selection should result in the lowest pressure drop and lowest initial cost while maintaining the same fresh water production rate. 7. Conclusions An analytical study has been undertaken to examine the operation of the solar diffusion driven desalination process under dynamic operating conditions. A unique operating mode that allows the system to operate without external cooling has been proposed and explored to improve water production and reduce the specific energy consumption. The solar heat input is recycled in a unique dynamic mode which eliminates the need for external cooling. The fresh water production and specific energy consumption have been investigated under various design and operating conditions. It is concluded that the delayed operating mode provides enhanced performance by reducing the specific energy consumption. The delayed operating mode yields a fresh water production rate of 6.3 L/m 2collector.day for the month of June in Florida with an average specific energy

F. Alnaimat, J.F. Klausner / Desalination 289 (2012) 35–44

consumption of 3.6 kWh/m 3. With the low specific energy consumption and low fabrication cost, it is believed that the solar DDD is competitive with other small scale desalination units for decentralized water production.

solar sw V

Nomenclature A Total surface area (m 2) Ac Cross sectional area of the packing (m 2) a Specific area of packing material (m 2/m 3) Cp Specific heat (kJ/kg or J/kg) DG Molecular diffusion coefficient of the gas (m 2/s) DL Molecular diffusion coefficient of the liquid (m 2/s) dp Diameter of the packing material (m) dz Height of the differential volume (m) Fr Froude number G Air mass flux (kg/m 2.s) g Gravitational acceleration (m/s 2) H Evaporator or condenser height (m) h Enthalpy (kJ/kg) hfg Latent heat of vaporization (kJ/kg) KG Thermal conductivity of the gas (W/m.K) KL Thermal conductivity of the liquid (W/m.K) kG Mass transfer coefficient of the gas side (m/s) kL Mass transfer coefficient of the liquid side (m/s) L Water mass flux (kg/m 2.s) Mv Vapor molecular weight (kg/kmol) _ m Mass flow rate (kg/s) P Total pressure (Pa or kPa) R Universal gas constant (kJ/kmol.K) Re Reynold number Sc Schmidt number T Temperature (°C or °K) t Time (sec) U Heat transfer coefficient (W/m 2.K) VG Air/vapor volumetric flow rate (m 3/s) We Webber number z Distance (m)

Acknowledgment

Greek α μ ρ σL σC Ф ω

Volume fraction (m 3/m 3) Dynamic viscosity (kg/m.s) Density (kg/m 3) Surface tension of liquid (N/m) Critical surface tension of the packing material (N/m) Relative humidity Humidity ratio

Subscript a amb cond evap fw G i in L mix out pack sat

Air Ambient Condenser Evaporator Fresh water Air/vapor mixture Interface Inlet parameter Liquid phase Mixture (gas and vapor) Exit parameter Packed bed Saturate state

43

Solar collector Saline water Vapor phase

This paper was prepared with the support of the Middle East Desalination Research Center (MEDRC) under Award No. 04-AS-003 and the Florida Energy Systems Consortium. Appendix A All parameters given in the appendix are defined similar to that in Alnaimat et al. [23].
 Ti ¼

TLþ

UG U

Ta


L  ;

1þ

UG UL

Psat(T) = 0.611379exp(bT − cT 2 + dT 3), b = 0.0723669, c = 2.78793 × 10 − 4, d = 6.76138 × 10 − 7 _ Cpmix ¼ m_ amþ_ am_ V Cpa þ m_ amþVm_ V CpV ;
1=2 U L ¼ kL ρL CpL KDLL ;
2=3 ; U G ¼ kG ðρG C PG Þ1=3 KDGG U = (UL− 1 + UG− 1) − 1,

 0:4 hμ L gi1=3 ; kL ¼ 0:0051ReLw 2=3 ScL −0:5 adp ρ L

0.7 1/3 kG = CReGA ScG (adp) − 2aDG   C = 2


0:75 aw σc −0:05 Re0:5 We0:2 ; LA Fr L L a ¼ 1− exp −1:45 σ L

ReLW ¼ awLμ ; ReGA ¼ aμG ; ReLA ¼ aμL L

G

L

ScL ¼ ρμDL L ; ScG ¼ ρ μDG G ; Fr L ¼ ρL 2ag ; WeL ¼ ρ Lσ L a 2

L

G

L

2

L

* Onda recommends C = 5.23 for dp > 15 mm. References [1] V. Nelson, New approaches in decentralized water infrastructure, Coalition for Alternative Wastewater Treatment Report, 2008. [2] W. Gocht, A. Sommerfeld, R. Rautenbach, T. Melin, L. Eilers, D.H. Neskakis, M.K. Hortsmann, A. Muhaidat, Decentralized desalination of brackish water by a directly coupled reverse-osmosis-photovoltaic-system—a pilot plan study in Jordan, Renewable Energy 14 (1998) 287–292. [3] A. Ghermandi, R. Messalem, Solar-driven desalination with reverse osmosis: the state of the art, Desalin. Water Treat. 7 (2009) 285–296. [4] E.S. Mohamed, G. Papadakis, E. Mathioulakis, V. Belessiotis, An experimental comparative study of the technical and economic performance of a small reverse osmosis desalination system equipped with a hydraulic energy recovery unit, Desalination 194 (2006) 239–250. [5] Y.P. Yadav, A. Kumar, Transient analytical investigations on a single basin solar still with water flow in the basin, Energy Convers. Manage. 31 (1991) 27–38. [6] G.N. Tiwari, V.S.V. Bapeshwara Rao, Transient performance of a single basin solar still with water flowing over the glass cover, Desalination 49 (1984) 231–241. [7] P. Valsaraj, An experimental study on solar distillation in a single slope basin still by surface heating the water mass, Renewable Energy 25 (2002) 607–612. [8] M. Sakthivel, S. Shanmugasundaram, Effect of energy storage medium (black granite gravel) on the performance of a solar still, Int. J. Energy Res. 32 (2008) 68–82. [9] M.S. Sodha, A. Kumar, G.N. Tiwari, R.C. Tyagi, Simple multiple wick solar still: analysis and performance, Solar Energy 26 (1981) 127–131. [10] A. Kumar, J.D. Anand, G.N. Tiwari, Transient analysis of a double slope-double basin solar distiller, Energy Convers. Manage. 31 (1991) 129–139. [11] M.A. Younis, M.A. Darwish, F. Juwayhel, Experimental and theoretical study of a humidification–dehumidification desalting system, Desalination 94 (1993) 11–24. [12] M.M. Farid, A.W. Al-Hajaj, Solar desalination with humidification dehumidification cycle, Desalination 106 (1996) 427–429. [13] S. Al-Hallaj, M.M. Farid, A.R. Tamimi, Solar desalination with humidification– dehumidification cycle: performance of the unit, Desalination 120 (1998) 273–280. [14] H. Müller-Holst, M. Engelhardt, W. Scholkopf, Small-scale thermal seawater desalination simulation and optimization of system design, Desalination 122 (1999) 255–262.

44

F. Alnaimat, J.F. Klausner / Desalination 289 (2012) 35–44

[15] Y.J. Dai, H.F. Zhang, Experimental investigation of a solar desalination unit with humidification and dehumidification, Desalination 130 (2000) 169–175. [16] J. Orfi, M. Laplante, H. Marmouch, N. Galanis, B. Benhamou, S. Ben Nasrallah, C.T. Nguyen, Experimental and theoretical study of a humidification–dehumidification water desalination system using solar energy, Desalination 168 (2004) 151–159. [17] A.S. Nafey, H.E.S. Fath, S.O. El-Helaby, A.M. Soliman, Solar desalination using humidification–dehumidification processes. Part II. An experimental investigation, Energy Convers. Manage. 45 (2004) 1263–1277. [18] C. Yamali, I. Solmus, Theoretical investigation of a humidification dehumidification desalination system configured by a double-pass flat plate solar air heater, Desalination 205 (2007) 163–177. [19] H. Ettouney, Design and analysis of humidification dehumidification desalination process, Desalination 183 (2005) 341–352. [20] A.M.I. Mohamed, N.A. El-Minshawy, Theoretical investigation of solar humidification–dehumidification desalination system using parabolic trough concentrators, Energy Convers. Manage. 52 (2011) 3112–3119. [21] J.F. Klausner, Y. Li, R. Mei, Evaporative heat and mass transfer for the diffusion driven desalination process, Heat Mass Transf. 42 (6) (2006) 528–536. [22] Y. Li, J.F. Klausner, R. Mei, J. Knight, Direct contact condensation in packed beds, Int. J. Heat Mass Transf. 49 (2006) 4751–4761.

[23] F. Alnaimat, J.F. Klausner, R. Mei, Transient analysis of direct contact evaporation and condensation within packed beds, Int. J. Heat Mass Transf. 54 (2011) 3381–3393. [24] 1991–2005 National Solar Radiation Database, http://rredc.nrel.gov/solar/ old_data/nsrdb/1991–2005/. [25] S. Al-Hallaj, S. Parekh, M.M. Farid, J.R. Selman, Solar desalination with humidification–dehumidification cycle: review of economics, Desalination 195 (2006) 169–186. [26] F. Alnaimat, Transient analysis of solar diffusion driven desalination, Phd Thesis, University of Florida, Gainesville, 2011. [27] T.V. Arjunan, H.S. Aybar, N. Nedunchezhian, Status of solar desalination in India, Renewable Sustainable Energy Rev. 13 (2009) 2408–2418. [28] A.A. Madani, Economics of desalination for three plant sizes, Desalination 78 (1990) 187–200. [29] E.E. Delyannis, A. Delyannis, Economics of solar stills, Desalination 52 (1985) 167–176. [30] M.K. Phadatare, S.K. Verma, Influence of water depth on internal heat and mass transfer in a plastic solar still, Desalination 217 (2007) 267–275. [31] S.D. Gomkale, Operational experience with solar stills in an Indian village and their contribution to the drinking water supply, Desalination 69 (1988) 177–182.