Abstract

Two rectangular modules with a total interior membrane surface area of 13.53 m2 were consecutively combined to evaluate the use of heat recovery in an air-gap membrane distillation (AGMD) system. Several operating inlet parameters including feed water temperature, mass water flow rate and salinity were investigated. The experimental results revealed that the performance of the system was improved by virtue of efficient heat recovery resulting from combining two AGMD membrane modules in series. Under optimal inlet operating parameters of cooling water temperature of 20 °C, salinity of 0.05% and flow rate of 3 l/min, the system productivity (Pp) increased up to 192.9%, 179.3%, 176.5% and 179.2%, and the thermal efficiency (ηth) by 261.5%, 232.6%, 239.4% and 227.3% at feed water temperatures of 45 °C, 55 °C, 65 °C and 75 °C, respectively. Concurrently, the specific waste heat input (Ew.h.i) decreased by 6.7%, 4.7%, 5.6% and 2.7% due to the efficient heat recovery. The results confirmed that heat recovery is an important factor affecting the AGMD system that could be improved by designing one of the two AGMD modules with polytetrafluoroethylene (PTFE) hollow fibers with a flow length shorter than the other one having a salt rejection rate of 99%.

NOMENCLATURE

     
  • AGMD

    Air-gap membrane distillation

  •  
  • MS-AGMD

    Multi-stage air-gap membrane distillation

  •  
  • PTFE

    Polytetrafluoroethylene

  •  
  • Sf

    Salinity of feed (%)

  •  
  • Sp

    Salinity of productivity (%)

  •  
  • CPw

    Specific heat of feed bulk (J/kg.°C)

  •  
  • EC

    Specific energy consumption (kWh/m3)

  •  
  • mproductivity

    Productivity rate (kg/hr)

  •  
  • mhot

    Hot bulk flow rate (l/min)

  •  
  • mcold

    Cold bulk flow rate (l/min)

  •  
  • Mw

    Mass water flow rate (l/min)

  •  
  • Pp

    Permeate flux (Productivity) (l/m2 h)

  •  
  • Ehot

    Energy flux from the hot bulk across the PTFE hollow fiber membrane and from the thermal insulation layer to productivity layer (kJ/min)

  •  
  • Ecold

    Energy flux from the productivity layer to cold water bulk (kJ/min)

  •  
  • EL.H.,productivity

    Evaporative latent heat (kJ/min)

  •  
  • Ew.h.i

    Specific waste heat input (MJ/m3)

  •  
  • ESen., t.i.l

    Sensible heat loss through thermal insulation layer (kJ/min)

  •  
  • ESen., productivity

    Sensible heat loss through productivity (kJ/min)

  •  
  • Tw

    Feed water temperature (oC)

  •  
  • Tc

    Cooling water temperature (oC)

  •  
  • t

    Time (minute)

  •  
  • ΔTcross

    Actual driving force across the hollow fiber membrane (oC)

  •  
  • ΔT2-1 (non-porous)

    Temperature difference between the inlet and outlet of the non-porous hollow fiber membrane in module (L) (oC)

  •  
  • ΔT3-4 (micro-porous)

    Temperature difference between the inlet and outlet of the micro-porous hollow fiber membrane in module (L) (oC)

  •  
  • ΔT3-1(HRR)

    Temperature difference between the inlet of non-porous and inlet of micro-porous membrane in module (L) (oC)

  •  
  • ΔT2-1(HRR)

    Temperature difference between the outlet and inlet of non-porous membrane in module (L) (oC)

  •  
  • ΔT3-2′(cross)

    Temperature difference across the hollow fiber membrane in module (L) (oC)

  •  
  • ΔT2-2′ (non-porous)

    Temperature difference between the inlet and outlet of the non-porous membrane in module (S) (oC)

  •  
  • ΔT3-3′(micro-porous)

    Temperature difference between the inlet and outlet of the micro-porous membrane in module (S) (oC)

  •  
  • ΔT3-2′(HRR)

    Temperature difference between the inlet of non-porous and inlet of micro-porous membrane in module (S) (oC)

  •  
  • ΔT2-2′(HRR)

    Temperature difference between outlet and inlet of non-porous membrane in module (S) (oC)

  •  
  • ΔT3-2(cross)

    Temperature difference across the hollow fiber membrane in module (S) (oC)

  •  
  • ΔHV

    Latent heat of evaporation (≈2326 kJ/kg)

  •  
  • WD

    Weight of the productivity (kg)

  •  
  • ρw

    Density (kg/m3)

  •  
  • ηth

    Thermal efficiency (%)

  •  
  • SRR

    Salt rejection rate (%)

  •  
  • RR

    Recovery ratio (dimensionless)

  •  
  • PR

    Performance ratio (dimensionless)

  •  
  • HRR

    Heat recovery ratio (%)

INTRODUCTION

Potable water for human consumption is among the major challenges encountered in the world due to population growth, deterioration of fresh water sources and industrialization (Swaminathan et al. 2016). The gap between availability and demand for drinkable water has driven the need for seawater desalination using membrane distillation (MD) to become a promising separation process in the 21st century (Deshmukh & Elimelech 2017; Andrés-Mañas et al. 2020). Membrane distillation (MD) may be defined as a thermal separation process driven by vapor pressure difference across the membrane (Ruiz-Aguirre et al. 2017; Nagy 2019). Water evaporation in an MD process occurs at the hot membrane surface. Vapor molecules then diffuse through the dry pores of the membrane until reaching the cold surface where they condense into a pure permeate flux (Zhao et al. 2013). Membrane distillation (MD) can be categorized into four basic configurations: direct contact membrane distillation (DCMD), air-gap membrane distillation (AGMD), sweep gas membrane distillation (SGMD) and vacuum membrane distillation (VMD) (Zare & Kargari 2018).

Recently, natural seawater has been successfully desalinated using a single stage air-gap membrane distillation (AGMD) system (El Mokhtar et al. 2019). The motivation behind introducing the air-gap membrane distillation (AGMD) system is to provide a thermal insulation layer between the hot micro-porous and the cold non-porous membranes (Al-Obaidani et al. 2008). The thickness of the layer has a profound influence on resistance to water vapor molecules, productivity (Guijt et al. 2005), conductive heat transfer through it and thermal efficiency (Chernyshov et al. 2005). Some investigators (e.g. Duong et al. 2015) demonstrated heat recovery for enhancing the thermal efficiency of the AGMD system; however, this enhancement comes at the expense of the productivity, which challenges the design of a reliable system at higher fluxes.

The first AGMD system with internal heat recovery was introduced by Henderyckx (1971). Woldemariam et al. (2016) emphasized the vital role of heat recovery in reducing the specific heat demands to around 105 kWh/m3, while they ranged from 692 kWh/m3 to 875 kWh/m3 without heat recovery. A heat recovery factor up to 55% of the thermal energy input into the MD system was reported by Guillén-Burrieza et al. (2011). In another study, Yao et al. (2012) stated that the continuous-effect AGMD module system with heat recovery could output approximately 80% of thermal efficiency (ηth) and 13.8 of gained output ratio (GOR) as well as about 5.3 kg/m2.h of productivity (Pp) and 5.7 of gained output ratio (GOR) as reported by Geng et al. (2014). More interestingly, increasing the number of MD modules enhanced the overall system performance (Guillén-Burrieza et al. 2011). In a recent paper, Lee et al. (2019) reported a very high GOR reaching up to 24.4 by using a novel multi-stage AGMD reversal design combined with a natural/forced cooling system in comparison with only 3.89 without a natural (passive) cooling system. Pangarkar & Deshmukh (2015) developed a multi-effect AGMD module for water treatment and reported about 3.2–3.6 times increase in the productivity and thermal efficiency of a four multi-effect AGMD module higher than that of the flux of a single-stage AGMD module. In another experimental investigation carried out by Geng et al. (2015), a four multi-stage AGMD process was designed using hollow fiber membranes for supplementary reverse osmosis (RO) brine and higher water recovery was obtained. The results showed an improvement of the multi-stage AGMD performance in comparison with the single-stage module with an increased gained output ratio (GOR) of 7.1 and productivity (Pp) of 6.8 kg/m2·h. Likewise, Khalifa et al. (2017) demonstrated that the productivity of the multi-stage AGMD module outperformed the single-stage AGMD module by around 2.6 and 3 times for series and parallel stage connections, respectively. In addition, the performance ratio (PR) of 0.38 resulted from running three modules compared with only 0.17 and 0.25 where one and two modules were used, respectively. Additionally, the specific thermal energy consumption (STEC) could be saved at different values of cooling water temperature and feed speed. Generally, increasing the number of modules linked in series improved the productivity (Pp), recovery ratio (RR) and performance ratio (PR) and diminished the STEC (Guillén-Burrieza et al. 2012).

Previous studies have demonstrated the vital role of a multi-stage process with an internal heat recovery factor in promoting AGMD performance. However, the trade-off between energy (evaporation) efficiency and productivity (i.e. high heat recovery and low productivity) limits the development of the multi-stage process due to preheating the feed via latent heat recovered in the form of sensible heat, hence increasing the heat recovery and decreasing the driving force and productivity.

The main aim of this study was to overcome the restriction of the multi-stage AGMD process with high heat recovery and low productivity. Therefore, a system of two AGMD membrane modules integrated in series was designed. The system was examined systematically with several inlet operating parameters including feed water temperature (Tw), water mass flow rate (Mw) and salinity (Sw). The two membrane modules L and S were designed based on the vacuum air-gap membrane distillation (V-AGMD) process, as declared in our previous work (Abu-Zeid et al. 2016). Both distillation and heat recovery were accomplished at the cold non-porous membranes inside the modules. Accordingly, the outlet temperature of the cold feed is gradually increased due to efficient heat recovery and conduction heat transfer. The fluids were exchanged in mass and heat across the membrane due to the temperature difference (ΔT) between the hot feed and the cold permeate sides. Parameters such as the productivity (Pp), recovery ratio (RR), heat recovery ratio (HRR), thermal efficiency (ηth), evaporative latent heat (EL.H., productivity), specific waste heat input (Ew.h.i), specific energy consumption (EC), heat losses through the thermal insulation layer (ESen.,t.i.l) and productivity (ESen., productivity) were used to assess the effect of varying lengths of membrane modules on the overall heat recovery and productivity of the AGMD system.

MATERIALS AND METHODS

Set-up of the AGMD module system

The key steps involved in the proposed two-stage AGMD module system in this experimental work are illustrated in the schematic diagram shown in Figure 1. The proposed distillation system consists of two AGMD modules designed with various dimensions: a long module L of 1.50 m (length) × 0.30 m (width) × 0.11 m (thickness), a short module S of 1.30 m (length) × 0.30 m (width) × 0.11 m (thickness), heating and cooling exchangers (DONGDA Water Industry Group Co., China), pumps, energy meter (DTS634, CHINT Electrics Co., Ltd, China), electric heater, digital flow meter (LZT-1002M-V, MBLD Instrument Company, China), cooler (Panasonic 2V47W225AUA, Panasonic Wanbao Appliances Compressor (Guangzhou) Co., Ltd, China), feeding and heating reservoirs. The width of the thermal insulation layer between the hot micro-porous and the cold non-porous membranes was around 0.004 m. The dimensions of the hot micro-porous and the cold non-porous PTFE hollow fiber membranes are listed in Table 1, and the properties of the hot micro-porous membrane are described in Table 2.

Table 1

The dimensions of the hot micro-porous and the cold non-porous PTFE hollow fiber membranes in modules L and S

Type of hollow fiber membraneNumber of hollow fibersLength of hollow fiber membrane (m)Interior/exterior diameter (m × 10−3)Area of interior membrane surface (m2)
Cold non-porous 1,626 1.4 (L) 0.6/1 4.29 (L) 
1.2 (S) 3.67 (S) 
Hot micro-porous 854 1.4 (L) 0.8/1.6–1.8 3.00 (L) 
1.2 (S) 2.57 (S) 
Type of hollow fiber membraneNumber of hollow fibersLength of hollow fiber membrane (m)Interior/exterior diameter (m × 10−3)Area of interior membrane surface (m2)
Cold non-porous 1,626 1.4 (L) 0.6/1 4.29 (L) 
1.2 (S) 3.67 (S) 
Hot micro-porous 854 1.4 (L) 0.8/1.6–1.8 3.00 (L) 
1.2 (S) 2.57 (S) 
Table 2

The properties of the hot micro-porous PTFE hollow fiber membrane

CharacterDDPT-S1612 style
Thickness (mm) 0.40 
Bubble point (MPa) 0.10–0.12 
Pore size (μm) 0.20–0.25 
Porosity (%) 50–60 
Hydrostatic pressure (MPa) 0.20–0.25 
CharacterDDPT-S1612 style
Thickness (mm) 0.40 
Bubble point (MPa) 0.10–0.12 
Pore size (μm) 0.20–0.25 
Porosity (%) 50–60 
Hydrostatic pressure (MPa) 0.20–0.25 
Figure 1

Schematic diagram of the two-stage AGMD modules arranged consecutively in a vertical direction.

Figure 1

Schematic diagram of the two-stage AGMD modules arranged consecutively in a vertical direction.

AGMD experiments

In the present experiment, the salty water was withdrawn from a lower point of the feeding reservoir by using a circulation pump and forced towards the cooling exchanger for lowering the temperature at the permeate side of the membrane. Then, it entered module L from the bottom in the upward direction, and then it entered module S from the bottom in the same direction. After exiting from the top of module S, water moved towards the heating exchanger for raising the temperature at the feed side of the membrane. In the opposite direction, the salty water returned back as a hot feed stream after being warmed up by the electric heater to the required feed water temperature (45 °C, 55 °C, 65 °C and 75 °C) through a heating exchanger coming into the top of module S in a downward direction, then the top of module L in the same direction. The brine leaving the bottom of module L was returned back to the feeding reservoir for a new distillation and heat recovery. The tested water mass flow rates (1.5, 3, 4.5 and 6 l/min) were adjusted by using a digital flowmeter. Finally, the distilled waters were collected from modules L and S and separately weighed for evaluating the heat recovery and subsequently the overall performance of the AGMD system.

Distillation and heat recovery were done through three major cycles as illustrated in Figure 1. (1) Heating cycle: in which a heating exchanger and hot tap water were used for warming up the feed salt water circulated into the hot micro-porous membrane. The required feed water temperature was controlled by a temperature controller (XMTD-3001, Easey Commercial Building, Hennessy Road, Wanchai, Hongkong, China) linked to a thermocouple immersed inside the heating reservoir. Temperatures through and between the two modules L and S were continuously monitored by thermocouples at the inlets and outlets of the hot micro-porous and the cold non-porous membranes. At the end of this cycle, a waste heat input was supplied by an electric heater and continued until reaching the required hot feed water temperature. (2) Cooling cycle: where a cooler and cooling exchanger were used for keeping the permeate side temperature constant at Tc = 20 °C throughout the experiment. (3) Distillation and heat recovery cycle: where different water salinities (0.05%, 1.5%, 3% and 5%) were distilled and the latent heat was recovered on the cold non-porous membrane through both modules.

The specific energy consumption in each test was measured using an energy meter. The value of energy consumption presented in this experiment was the sum of cooling, heating and circulation energy consumption jointly. Prior to each measurement, modules L and S were left running for 60 minutes to remove dissolved gases from the feed and to reach a steady-state condition. The electrical conductivity of the pure permeate flux (productivity) and feed salinity (0.05%, 1.5%, 3% and 5%) were measured by using a conductivity meter (DDS-11A) to observe any membrane pore wetting that might occur. Each experiment was repeated three times and the average values were presented.

Experimental measurements and calculations

Assessment of the influence of membrane modules L and S with different hollow fiber flow lengths on the overall performance of AGMD included the determination of productivity (Pp), recovery ratio (RR), heat recovery ratio (HRR), thermal efficiency (ηth), evaporative latent heat (EL.H., productivity), specific waste heat input (Ew.h.i), specific energy consumption (EC), heat loss through thermal insulation layer (ESen.,t.i.l) and sensible heat loss through productivity (ESen., productivity). Heat recovery ratio (HRR) was calculated here according to the variations of temperature ratios at inlets and outlets of each membrane module. To do that as described schematically in Figure 2, productivity (Pp) and different temperatures (T) through and between the membrane modules L and S were measured every ten minutes for 60 minutes experimental time and various temperature differences (ΔTs) at different inlet operating parameters as listed in Table 3. The measurements were then used in the different equations presented in Table 4 to calculate various energy (E) fluxes across the micro-porous membrane, non-porous membrane and thermal insulation layer for membrane modules L and S as illustrated in Figure 3. The experimental productivity (Pp) along with data errors was computed and is exhibited in Table 5. The values of water density (ρw) and specific heat (Cpw) for inlet water depend basically on the salinity (Sw) and the ambient air temperature. The values used in the experimental works of this study are listed in Table 6 at 25 °C of ambient air temperature.

Table 3

Various temperature differences (ΔTs) recorded through and between membrane modules (L) and (S) at different feed water temperatures (Tw), water mass flow rates (Mw) and salinities (Sw)a

Constant operating inlet parameters
Mw = 3 l/min, Tc = 20 °C, Sw = 0.05%
Feed water temperature (oC) Module (L)ΔT2-1 (non-porous)ΔT3-4 (micro-porous)ΔT3-1 (HRR)ΔT2-1 (HRR)ΔT3-2(cross)
45 6.2 9.2 22.3 6.2 16.4 
55 8.3 11.6 25.2 8.3 19.0 
65 11.7 16.1 32.0 11.7 23.0 
75 13.1 17.9 32.9 13.1 25.2 
Module (S)ΔT2-2(non-porous)ΔT3-3(micro-porous)ΔT3-2(HRR)ΔT2-2(HRR)ΔT3-2 (cross)
45 4.8 6.8 14.6 4.8 15.3 
55 6.6 9.2 17.4 6.6 18.1 
65 8.7 11.9 19.9 8.7 21.7 
75 10.6 14.1 23.8 10.6 24.5 
Water mass flow rate (l/min) Tw = 65 °C, Tc = 20 °C, Sw = 0.05% 
Module (L)      
1.5 12.0 19.9 29.2 12.0 20.1 
11.7 16.1 32.0 11.7 23.0 
4.5 11.0 14.0 33.3 11.0 25.7 
9.8 12.2 33.4 9.8 28.0 
Module (S)      
1.5 9.8 13.2 18.8 9.8 19.0 
8.7 11.9 19.9 8.7 21.7 
4.5 7.9 10.4 21.3 7.9 23.4 
7.1 9.3 22.7 7.1 24.7 
Salinity (%) Tw = 75 °C, Tc = 20 °C, Mw = 3 l/min 
Module (L)      
0.05 13.1 17.9 32.99 13.1 25.2 
1.5 17.7 19.5 33.71 17.7 24.0 
3.0 14.6 23.2 34.55 14.6 26.0 
5.0 14.0 26.6 34.6 14.0 27.3 
Module (S)      
0.05 10.6 14.1 23.8 10.6 24.5 
1.5 14.4 15.9 24.3 14.4 23.7 
3.0 12.9 17.2 25.8 12.9 23.9 
5.0 11.0 17.0 26.3 11.0 26.3 
Constant operating inlet parameters
Mw = 3 l/min, Tc = 20 °C, Sw = 0.05%
Feed water temperature (oC) Module (L)ΔT2-1 (non-porous)ΔT3-4 (micro-porous)ΔT3-1 (HRR)ΔT2-1 (HRR)ΔT3-2(cross)
45 6.2 9.2 22.3 6.2 16.4 
55 8.3 11.6 25.2 8.3 19.0 
65 11.7 16.1 32.0 11.7 23.0 
75 13.1 17.9 32.9 13.1 25.2 
Module (S)ΔT2-2(non-porous)ΔT3-3(micro-porous)ΔT3-2(HRR)ΔT2-2(HRR)ΔT3-2 (cross)
45 4.8 6.8 14.6 4.8 15.3 
55 6.6 9.2 17.4 6.6 18.1 
65 8.7 11.9 19.9 8.7 21.7 
75 10.6 14.1 23.8 10.6 24.5 
Water mass flow rate (l/min) Tw = 65 °C, Tc = 20 °C, Sw = 0.05% 
Module (L)      
1.5 12.0 19.9 29.2 12.0 20.1 
11.7 16.1 32.0 11.7 23.0 
4.5 11.0 14.0 33.3 11.0 25.7 
9.8 12.2 33.4 9.8 28.0 
Module (S)      
1.5 9.8 13.2 18.8 9.8 19.0 
8.7 11.9 19.9 8.7 21.7 
4.5 7.9 10.4 21.3 7.9 23.4 
7.1 9.3 22.7 7.1 24.7 
Salinity (%) Tw = 75 °C, Tc = 20 °C, Mw = 3 l/min 
Module (L)      
0.05 13.1 17.9 32.99 13.1 25.2 
1.5 17.7 19.5 33.71 17.7 24.0 
3.0 14.6 23.2 34.55 14.6 26.0 
5.0 14.0 26.6 34.6 14.0 27.3 
Module (S)      
0.05 10.6 14.1 23.8 10.6 24.5 
1.5 14.4 15.9 24.3 14.4 23.7 
3.0 12.9 17.2 25.8 12.9 23.9 
5.0 11.0 17.0 26.3 11.0 26.3 

aThe symbol (Tc) presented in the table refers to the cooling water temperature.

Table 4

Different equations used in calculating various energy (E) fluxes across micro-porous membrane, non-porous membrane and thermal insulation layer in modules L and S

VariableEquationUnit
Ew.h.i  MJ/m3 (L) 
 MJ/m3 (S) 
Ehot  kJ/min (L)
kJ/min (S) 
Ecold  kJ/min (L) 
 kJ/min (S) 
HRR  % (L) 
 % (S) 
ESen., t.i.l  kJ/min 
ESen., productivity  kJ/min 
EL.H., productivity  kJ/min 
RR  dimensionless 
EC  kWh/m3 
Pd  L/m2.hr 
ηth  
  
VariableEquationUnit
Ew.h.i  MJ/m3 (L) 
 MJ/m3 (S) 
Ehot  kJ/min (L)
kJ/min (S) 
Ecold  kJ/min (L) 
 kJ/min (S) 
HRR  % (L) 
 % (S) 
ESen., t.i.l  kJ/min 
ESen., productivity  kJ/min 
EL.H., productivity  kJ/min 
RR  dimensionless 
EC  kWh/m3 
Pd  L/m2.hr 
ηth  
  
Table 5

The experimental productivity (Pp) with data errors at different feed water temperatures (Tw), water mass flow rates (Mw) and salinities (Sw)

Feed water temperature (oC)Module (L)Module (S)
45 0.73 ± 0.03 1.42 ± 0.22 
55 1.01 ± 0.18 1.81 ± 0.15 
65 1.30 ± 0.26 2.30 ± 0.10 
75 1.50 ± 0.21 2.69 ± 0.13 
Water mass flow rate (l/min)Module (L)Module (S)
1.5 0.94 ± 0.13 1.28 ± 0.25 
1.30 ± 0.26 2.30 ± 0.10 
4.5 1.86 ± 0.09 2.92 ± 0.19 
2.58 ± 0.15 3.60 ± 0.24 
Salinity (%)Module (L)Module (S)
0.05 1.50 ± 0.21 2.69 ± 0.13 
1.5 1.31 ± 0.17 2.28 ± 0.12 
3.0 1.02 ± 0.12 1.51 ± 0.16 
5.0 0.62 ± 0.13 0.88 ± 0.20 
Feed water temperature (oC)Module (L)Module (S)
45 0.73 ± 0.03 1.42 ± 0.22 
55 1.01 ± 0.18 1.81 ± 0.15 
65 1.30 ± 0.26 2.30 ± 0.10 
75 1.50 ± 0.21 2.69 ± 0.13 
Water mass flow rate (l/min)Module (L)Module (S)
1.5 0.94 ± 0.13 1.28 ± 0.25 
1.30 ± 0.26 2.30 ± 0.10 
4.5 1.86 ± 0.09 2.92 ± 0.19 
2.58 ± 0.15 3.60 ± 0.24 
Salinity (%)Module (L)Module (S)
0.05 1.50 ± 0.21 2.69 ± 0.13 
1.5 1.31 ± 0.17 2.28 ± 0.12 
3.0 1.02 ± 0.12 1.51 ± 0.16 
5.0 0.62 ± 0.13 0.88 ± 0.20 
Table 6

Density (ρw) and specific heat (CPw) of feed inlet water at different salinity

Salinity (%)Water density (kg/m3)Specific heat (J/kg.oC)
0.05 996.9 4,186.5 
1.5 1,002.5 4,172.7 
1,019.6 4,152.2 
1,034.9 4,098.3 
Salinity (%)Water density (kg/m3)Specific heat (J/kg.oC)
0.05 996.9 4,186.5 
1.5 1,002.5 4,172.7 
1,019.6 4,152.2 
1,034.9 4,098.3 
Figure 2

Different temperature-recording regimes through and between membrane modules (L) and (S).

Figure 2

Different temperature-recording regimes through and between membrane modules (L) and (S).

Figure 3

Different energy fluxes across the cold non-porous membrane, hot micro-porous hollow fiber membrane and thermal insulation layer.

Figure 3

Different energy fluxes across the cold non-porous membrane, hot micro-porous hollow fiber membrane and thermal insulation layer.

RESULTS AND DISCUSSION

It is important to indicate that the two membrane modules L and S used in this investigation differed only in the length of hollow fibers employed (i.e. those of module S are only 15% shorter than those of module L). Theoretically, this little difference seems ineffective on the performance of the multi-stage air-gap membrane distillation (MS-AGMD) process. However, when both modules were connected in series and experimentally tested under various inlet operating parameters, they gave promising results in terms of productivity and heat recovery, as will be explained in detail in the following sections.

Effect of the length of small hollow fibers at different feed water temperatures (TW)

The resulting data presented in Figure 4 show the impact of the small hollow fiber length difference between the two membrane modules L (1.4 m) and S (1.2 m) integrated in series on the whole performance of the MS-AGMD process for different bulk feed input temperatures, keeping all other operating inlet parameters stable. It can be seen from Figure 4(a) and 4(b) that the integrated module S (shorter flow length) in series with module L (longer flow length) notably increased the whole productivity (Pp) and recovery ratio (RR) of the MS-AGMD process by about 192.9%, 179.3%, 176.5% and 179.2% at feed water temperatures of 45 °C, 55 °C, 65 °C and 75 °C, respectively. Similarly, the heat recovery ratio (HRR) was enhanced greatly by about 118.3%, 115.5%, 119.4% and 111.8% (Figure 4(c)). The primary reasons are linked to the fact that the average hot bulk feed temperature was maintained at a higher level for a shorter flow length, thus ensuring a larger driving force for mass transfer in the upcoming stages of the MS-AGMD modules and then the productivity eventually. The main cause for the coming down of productivity in module L (longer flow length) has been explicated by Liu et al. (2016) as follows: the evaporation of water and vapor molecules transferring from the hot feed solution to the thermal insulation layer is accompanied simultaneously by the carriage of an amount of heat from the hot feed stream to the insulation layer which inevitably reduces the outlet hot feed temperature. Thus, the average bulk hot feed temperature in module L was brought down and the productivity decreased correspondingly. The low values of Ehot, Ecold and high values of EL.H, productivity listed in Table 7 for membrane module S and compared with module L support our explanation.

Table 7

The various energy fluxes (E) calculated at different feed water temperatures (Tw)

Feed water temperature (oC)EhotEcoldEL.H, productivityESen., t.i.lESen.,productivityEw.h.i
Module L 
 45 115.1 77.6 28.5 49.0 37.5 205.3 
 55 148.7 103.9 39.1 64.8 44.8 237.8 
 65 201.6 146.4 50.6 95.8 55.2 287.9 
 75 223.9 164.0 58.2 105.7 59.9 315.5 
Module S 
 45 85.1 60.2 55.0 5.1 24.9 191.5 
 55 114.6 82.6 70.2 12.4 32.0 226.6 
 65 148.6 109.2 89.4 19.8 39.4 271.6 
 75 176.6 133.3 104.4 28.9 43.3 306.7 
Feed water temperature (oC)EhotEcoldEL.H, productivityESen., t.i.lESen.,productivityEw.h.i
Module L 
 45 115.1 77.6 28.5 49.0 37.5 205.3 
 55 148.7 103.9 39.1 64.8 44.8 237.8 
 65 201.6 146.4 50.6 95.8 55.2 287.9 
 75 223.9 164.0 58.2 105.7 59.9 315.5 
Module S 
 45 85.1 60.2 55.0 5.1 24.9 191.5 
 55 114.6 82.6 70.2 12.4 32.0 226.6 
 65 148.6 109.2 89.4 19.8 39.4 271.6 
 75 176.6 133.3 104.4 28.9 43.3 306.7 
Figure 4

Effect of the length of small hollow fibers at different feed water temperatures (Tw) on the: (a) productivity, (b) recovery ratio, (c) heat recovery ratio and (d) thermal efficiency of the MS-AGMD process (Mw = 3 l/min, Tc = 20 °C, Sw = 0.05%).

Figure 4

Effect of the length of small hollow fibers at different feed water temperatures (Tw) on the: (a) productivity, (b) recovery ratio, (c) heat recovery ratio and (d) thermal efficiency of the MS-AGMD process (Mw = 3 l/min, Tc = 20 °C, Sw = 0.05%).

Effect of the length of the small hollow fibers at different mass water flow rates (MW)

The role of integrating short module S in series with long module L on the performance of the MS-AGMD process at a stable inlet feed temperature of 65 °C, cooling water temperature of 20 °C, and feed salinity of 0.05% was clarified by examining the productivity (Pp) and recovery ratio (RR) as a function of mass water flow rate (Mw) as presented in Figure 5. Figure 5(a) and 5(b) reveal that the productivity (Pp) and recovery ratio (RR) of the MS-AGMD process were increased remarkably by about 136.3%, 176.5%, 156.5% and 139.3% and the heat recovery ratio (HRR) by around 126.7%, 119.4%, 112.4% and 106.8% as reported in Figure 5(c). For the long module L alone, increases of 73.3%, 56.6%, 63.8% and 71.7% were only obtained in the productivity (Pp) and recovery ratio (RR) of the MS-AGMD process and increases up to 78.8%, 83.7%, 88.9% and 91.6% were achieved in the values of heat recovery ratio (HRR). The positive influence of short module S compared with long module L on the MS-AGMD process performance is due to two main reasons: firstly, a small reduction in outlet feed temperature of module S with shorter flow length especially at higher mass feed flow rates. This in turn maintains a larger trans-membrane temperature difference in subsequent stages and thus Pp and RR eventually. Secondly, a reduction of the temperature polarization effect (i.e. low thermal boundary layer thickness) on the hot membrane side while increasing the feed flow (Khalifa & Lawal 2015) improved Pp and RR.

Figure 5

Effect of the length of the small hollow fibers at different mass water flow rates (Mw) on the: (a) productivity, (b) recovery ratio, (c) heat recovery ratio, and (d) thermal efficiency of the MS-AGMD process (Tw = 65 °C, Tc = 20 °C, Sw = 0.05%).

Figure 5

Effect of the length of the small hollow fibers at different mass water flow rates (Mw) on the: (a) productivity, (b) recovery ratio, (c) heat recovery ratio, and (d) thermal efficiency of the MS-AGMD process (Tw = 65 °C, Tc = 20 °C, Sw = 0.05%).

It is known that better energy recovery within a module gives rise to a greater temperature of the preheated feed, which results in a lower external waste heat input and hence larger thermal efficiency (ηth). Figure 5(d) displays that the thermal efficiency (ηth) of the MS-AGMD process greatly improved to about 205.8%, 239.4%, 210.9% and 182.6% by virtue of short module S corresponding to increases only up to 48.5%, 41.7%, 47.4% and 54.7% in the case of long module L. This was ascribed to keeping sensible heat loss through the insulation layer (ESen., t.i.l) and the productivity (ESen., productivity) at minimal levels for a shorter flow length compared with module L of longer flow length. This resulted in lower values of ESen., t.i.l and ESen., productivity (Table 8) for membrane module S and higher values for module L. The tendency of results for the specific waste heat input (Ew.h.i) and the specific thermal energy consumption (STEC) at different mass water flow rates was found to be similar to that of the feed water temperature parameter. Module S with shorter flow length reduced the Ew.h.i of the MS-AGMD process by about 5.4%, 5.6%, 8.9% and 11.7%. In addition, the increases in STEC determined as 0.28, 0.30, 0.09 and 0.11 kWh/m3 are meaningless and can be disregarded especially when compared with the significant increase in the Pp reaching up to 1.67, 2.34, 2.62 and 2.68 m3/(m2·h) × 10−3.

Table 8

The various energy fluxes (E) calculated at different water mass flow rates (MW)

Water mass flow rate (l/min)EhotEcoldEL.H, productivityESen., t.i.lESen.,productivityEw.h.i
Module L 
 1.5 124.6 75.1 36.6 38.5 49.4 125.8 
 3 201.6 146.4 50.6 95.8 55.2 287.9 
 4.5 262.9 206.5 72.4 134.1 56.3 482.6 
 6 30.5.5 245.3 100.2 145.1 60.1 701.1 
Module S 
 1.5 82.6 61.3 49.9 11.4 21.2 118.9 
 3 148.6 109.2 89.4 19.8 39.4 271.6 
 4.5 195.3 148.3 113.3 35.0 46.9 439.5 
 6 232.9 177.7 139.7 38.0 55.1 618.5 
Water mass flow rate (l/min)EhotEcoldEL.H, productivityESen., t.i.lESen.,productivityEw.h.i
Module L 
 1.5 124.6 75.1 36.6 38.5 49.4 125.8 
 3 201.6 146.4 50.6 95.8 55.2 287.9 
 4.5 262.9 206.5 72.4 134.1 56.3 482.6 
 6 30.5.5 245.3 100.2 145.1 60.1 701.1 
Module S 
 1.5 82.6 61.3 49.9 11.4 21.2 118.9 
 3 148.6 109.2 89.4 19.8 39.4 271.6 
 4.5 195.3 148.3 113.3 35.0 46.9 439.5 
 6 232.9 177.7 139.7 38.0 55.1 618.5 

Effect of the length of the small hollow fibers at different levels of salinity (SW)

Figure 6 illustrates the role of the integrated modules S and L in series with different hollow fiber flow lengths in promoting the performance of the MS-AGMD process. Figure 6(a)–6(c) demonstrate at various salinity levels of 0.05%, 1.5%, 3% and 5% that the corresponding values of Pp and RR were increased by 179.2%, 173.4%, 147.7% and 142.2% and HRR by 111.8%, 112.7%, 118.4% and 103.2%, respectively. Also, it was perceived that ηth was raised by 227.3%, 212.9%, 199.2% and 223.6% as shown in Figure 6(d). The augmentation in these parameters was due partially to the aforementioned reasons in the case of feed temperature and mass flow rate parameters as well as lower concentration polarization effect (i.e. lower concentration boundary layer thickness) within the short module S compared with the long module L. Similarly, it was observed that the Pp, RR and ηth for each module dropped gradually with salinity owing to the increasing thickness of the concentration and thermal boundary layers at the hot membrane surface, thus reducing the driving force for mass transfer and activity coefficient of the water (Khalifa et al. 2015).

Figure 6

Effect of the length of the small hollow fibers at different salinities (Sw) on the: (a) productivity, (b) recovery ratio, (c) heat recovery ratio and (d) thermal efficiency of the MS-AGMD process (Tw = 75 °C, Tc = 20 °C, Mw = 3 l/min).

Figure 6

Effect of the length of the small hollow fibers at different salinities (Sw) on the: (a) productivity, (b) recovery ratio, (c) heat recovery ratio and (d) thermal efficiency of the MS-AGMD process (Tw = 75 °C, Tc = 20 °C, Mw = 3 l/min).

The shorter flow length of the membrane module S decreased the specific waste heat input (Ew.h.i) by 2.7%, 1.2%, 8.1% and 7.3% (Table 9). The specific thermal energy consumption (STEC) had the same behavior of feed water temperature and mass flow rate parameters except at 5% salinity where membrane module S largely reduced the STEC due to greater water productivity. The experimental results agree with those of Alsaadi et al. (2015) and Duong et al. (2015). At feed salinities of 0.05%, 1.5% and 3%, the membrane module S consumed a little more energy than membrane module L. However, these further increases were small and could be omitted corresponding to the larger in Pp and RR. The further slight differences between the two membrane modules were nearly 0.13, 0.49 and 0.71 kWh/m3 while perceptible increases in Pp amounted to 3.53, 3.42 and 3.75 m3/(m2·h) × 10−3. The high feed salinity (Sw) does not affect the quality of productivity and salt rejection rate under all inlet operating parameters as it was greater than 99%. This means that there was no occurrence of membrane wetting or leakages as the membranes remained dry during the whole testing time.

Table 9

The various energy fluxes (E) calculated at different salinities (Sw)

Salinity (%)EhotEcoldEL.H, productivityESen., t.i.lESen.,productivityEw.h.i
Module L 
 0.05 223.9 164.0 58.23 105.7 59.9 315.5 
 1.5 244.1 221.8 51.17 170.6 22.2 296.3 
 3 290.4 179.8 39.81 140.0 110.6 320.2 
 5 333.0 170.7 24.0 146.6 162.3 332.9 
Module S 
 0.05 176.6 133.3 104.4 28.9 43.3 306.7 
 1.5 199.0 180.2 88.7 91.5 18.7 292.6 
 3 215.3 161.5 58.8 102.7 53.8 294.3 
 5 212.8 134.1 34.2 99.9 78.7 308.5 
Salinity (%)EhotEcoldEL.H, productivityESen., t.i.lESen.,productivityEw.h.i
Module L 
 0.05 223.9 164.0 58.23 105.7 59.9 315.5 
 1.5 244.1 221.8 51.17 170.6 22.2 296.3 
 3 290.4 179.8 39.81 140.0 110.6 320.2 
 5 333.0 170.7 24.0 146.6 162.3 332.9 
Module S 
 0.05 176.6 133.3 104.4 28.9 43.3 306.7 
 1.5 199.0 180.2 88.7 91.5 18.7 292.6 
 3 215.3 161.5 58.8 102.7 53.8 294.3 
 5 212.8 134.1 34.2 99.9 78.7 308.5 

In brief, improving the performance of the MS-AGMD process with an internal heat recovery system could be successfully accomplished by linking two AGMD membrane modules in series as a result of the recovering of the latent heat of condensation and the energy associated with the saline flowing out of one module being utilized in the next module (Warsinger et al. 2015). To achieve high heat recovery and good performance of the multi-stage configuration, the present research work recommends the design of one of two AGMD membrane modules with hollow fiber flow length shorter than the other one.

CONCLUSIONS

The results of this study demonstrated that the use of heat recovery realized a significant improvement in the air-gap membrane distillation (AGMD) system performance. The limitations of the multi-stage air gap membrane distillation (MS-AGMD) process, which is a trade-off between energy (evaporation) efficiency and productivity (i.e., high heat recovery and low pure permeate), were solved. It causes perceptible increases in the productivity, recovery ratio, thermal efficiency and evaporative latent heat. Also, a noticeable decrease in specific energy consumption, specific waste heat input, sensible heat loss through the thermal insulation layer and productivity were obtained with several operating inlet parameters. At the optimal operating inlet parameters of cooling water temperature of 20 °C, salinity of 0.05% and mass water flow rate of 3 l/min, the system's productivity (Pp) is totally augmented significantly by 192.9%, 179.3%, 176.5% and 179.2%, and also the thermal efficiency (ηth) by 261.5%, 232.6%, 239.4% and 227.3% at feed water temperatures of 45 °C, 55 °C, 65 °C and 75 °C, respectively. Correspondingly, the specific waste heat input (Ew.h.i) is decreased by 6.7%, 4.7%, 5.6% and 2.7%. Also, no important difference resulted in the specific energy consumption (EC) between modules S and L, both having almost the same values except a little increase reaching 0.12, 0.51, 0.30 and 0.13 kWh/m3. However, this small increase leads to increases of the total productivity (Pp) of about 0.68, 1.36, 2.34 and 3.53 m3/(m2·h) × 10−3. Due to its simplicity and robustness, the AGMD system outlined in this work is able to provide reasonable water demands for small areas.

ACKNOWLEDGEMENTS

The authors would like to gratefully acknowledge the National Basic Research Program of China (No. 2015CB655303), the National Natural Science Foundation of China (No. 21076176), the National Key Technologies R&D Program (2015ZX07406006), and the Distinguished Scientist Fellowship Program (DSFP) King Saud University for the financial support of this work. Also, the authors are grateful to Prof. Paul McNulty (School of Biosystems and Food Engineering, University College Dublin, Ireland) for careful proofreading of the manuscript.

REFERENCES

REFERENCES
Abu-Zeid
M. A. E. R.
,
Zhang
L.
,
Jin
W. Y.
,
Feng
T.
,
Wu
Y.
,
Chen
H. L.
&
Hou
L.
2016
Improving the performance of the air gap membrane distillation process by using a supplementary vacuum pump
.
Desalination
384
,
31
42
.
Al-Obaidani
S.
,
Curcio
E.
,
Macedonio
F.
,
Di Profio
G.
,
Al-Hinai
H.
&
Drioli
E.
2008
Potential of membrane distillation in seawater desalination: thermal efficiency, sensitivity study and cost estimation
.
Journal of Membrane Science
323
,
85
98
.
Alsaadi
A. S.
,
Francis
L.
,
Maab
H.
,
Amy
G. L.
&
Ghaffour
N.
2015
Evaluation of air gap membrane distillation process running under sub-atmospheric conditions: experimental and simulation studies
.
Journal of Membrane Science
489
,
73
80
.
Andrés-Mañas
J. A.
,
Roca
L.
,
Ruiz-Aguirre
A.
,
Acién
F. G.
,
Gil
J. D.
&
Zaragoza
G.
2020
Application of solar energy to seawater desalination in a pilot system based on vacuum multi-effect membrane distillation
.
Applied Energy
258
,
114068
.
Chernyshov
M. N.
,
Meindersma
G. W.
&
de Haan
A. B.
2005
Comparison of spacers for temperature polarization reduction in air gap membrane distillation
.
Desalination
183
,
363
374
.
Duong
H. C.
,
Chivas
A. R.
,
Nelemans
B.
,
Duke
M.
,
Gray
S.
,
Cath
T. Y.
&
Nghiem
L. D.
2015
Treatment of RO brine from CSG produced water by spiral-wound air gap membrane distillation: a pilot study
.
Desalination
366
,
121
129
.
El Mokhtar
I.
,
Boubakri
A.
,
Bouguecha
S. A.
&
Hafiane
A.
2019
Modeling and experimental study of air gap membrane distillation unit: application for seawater desalination
.
Desalination and Water Treatment
154
,
72
81
.
Geng
H. X.
,
Wu
H. Y.
,
Li
P. L.
&
He
Q. F.
2014
Study on a new air-gap membrane distillation module for desalination
.
Desalination
334
,
29
38
.
Geng
H. X.
,
Wang
J.
,
Zhang
C. Y.
,
Li
P. L.
&
Chang
H. Y.
2015
High water recovery of RO brine using multi-stage air gap membrane distillation
.
Desalination
355
,
178
185
.
Guijt
C. M.
,
Meindersma
G. W.
,
Reith
T.
&
de Haan
A. B.
2005
Air gap membrane distillation 2. Model validation and hollow fibre module performance analysis
.
Separation and Purification Technology
43
,
245
255
.
Guillén-Burrieza
E.
,
Blanco
J.
,
Zaragoza
G.
,
Alarcon
D.-C.
,
Palenzuela
P.
,
Ibarra
M.
&
Gernjak
W.
2011
Experimental analysis of an air gap membrane distillation solar desalination pilot system
.
Journal of Membrane Science
379
,
386
396
.
Guillén-Burrieza
E.
,
Zaragoza
G.
,
Miralles-Cuevas
S.
&
Blanco
J.
2012
Experimental evaluation of two pilot-scale membrane distillation modules used for solar desalination
.
Journal of Membrane Science
409–410
,
264
275
.
Henderyckx
Y.
1971
Evaporation-condensation recovery of a solution component using vapor-permeable wall spaced from a cold wall. US Patent 3563860
.
Khalifa
A. E.
&
Lawal
D. U.
2015
Performance and optimization of air gap membrane distillation system for water desalination
.
Arabian Journal for Science and Engineering
40
,
3627
3639
.
Khalifa
A. E.
,
Alawad
S. M.
&
Antar
M. A.
2017
Parallel and series multistage air gap membrane distillation
.
Desalination
417
,
69
76
.
Liu
Z. Y.
,
Gao
Q. J.
,
Lu
X. L.
,
Zhao
L. H.
,
Wu
S.
,
Ma
Z.
&
Zhang
H.
2016
Study on the performance of double-pipe air gap membrane distillation module
.
Desalination
396
,
48
56
.
Nagy
E.
2019
Membrane distillation
. In:
Basic Equations of Mass Transport through a Membrane Layer
, 2nd edn.
Elsevier
,
Amsterdam
,
The Netherlands
, chapter 9, pp.
483
496
.
Pangarkar
B. L.
&
Deshmukh
S. K.
2015
Theoretical and experimental analysis of multi-effect air gap membrane distillation process (ME-AGMD)
.
Journal of Environmental Chemical Engineering
3
(
3
),
2127
2135
.
Ruiz-Aguirre
A.
,
Andrés-Mañas
J. A.
,
Fernández-Sevilla
J. M.
&
Zaragoza
G.
2017
Modeling and optimization of a commercial permeate gap spiral wound membrane distillation module for seawater desalination
.
Desalination
419
,
160
168
.
Swaminathan
J.
,
Chung
H. W.
,
Warsinger
D. M.
&
Lienhard
J. H.
2016
Membrane distillation model based on heat exchanger theory and configuration comparison
.
Applied Energy
184
,
491
505
.
Warsinger
D. E. M.
,
Swaminathan
J.
,
Maswadeh
L. A.
&
Lienhard
J. H.
2015
Superhydrophobic condenser surfaces for air gap membrane distillation
.
Journal of Membrane Science
492
,
578
587
.
Woldemariam
D.
,
Kullab
A.
,
Fortkamp
U.
,
Magner
J.
,
Royen
H.
&
Martin
A.
2016
Membrane distillation pilot plant trials with pharmaceutical residues and energy demand analysis
.
Chemical Engineering Journal
306
,
471
483
.
Zare
S.
&
Kargari
A.
2018
Membrane properties in membrane distillation
. In:
Emerging Technologies for Sustainable Desalination Handbook
(
Gude
V. G.
, ed.),
Butterworth-Heinemann
,
Oxford
,
UK
, pp.
107
156
.
Zhao
K.
,
Heinzl
W.
,
Wenzel
M.
,
Büttner
S.
,
Bollen
F.
,
Lange
G.
,
Heinzl
S.
&
Sarda
N.
2013
Experimental study of the memsys vacuum-multi-effect-membrane-distillation (V-MEMD) module
.
Desalination
323
,
150
160
.