Case Studies in Thermal Engineering 56 (2024) 104265

Contents lists available at ScienceDirect

Case Studies in Thermal Engineering

journal homepage: www.elsevier.com/locate/csite

Heat and water recovery from gaseous stream using a flat-sheet

ceramic membrane as a transport membrane condenser
Van Cong Le a, 1, Jun Cong Ge b, 1, Suk-Min Seo b, Dae-Hae Kim c, Chan Woo Park a, b, d, *
a Department of Energy Storage and Conversation Engineering, Jeonbuk National University, 567 Baekje-daero, Deokjin-gu, Jeonju-si, Jeollabuk-do,
54896, Republic of Korea
b Division of Mechanical Design Engineering, Jeonbuk National University, 567 Baekje-daero, Deokjin-gu, Jeonju-si, Jeollabuk-do, 54896, Republic of

Korea
c Thermalchemical Energy System R&D Group, Korea Institute of Industrial Technology, Cheonan, Chungcheongnam, 31056, Republic of Korea
d Advanced Mechanical Components Design & Research Center, Jeonbuk National University, 567, Baekje-daero, Deokjin-gu, Jeonju-si, Jeollabuk-do,

561-756, Republic of Korea

HIGHLIGHTS

• A flat-sheet membrane condenser was investigated to recover heat and water from saturated gas.
• The effects of operating conditions on recovery performance were studied.
• The highest water recovery rate obtained is 84.1%.
• Thermal resistances were analyzed to estimate the heat transfer performance.
• Computational fluid dynamics were used to further study the recovery mechanism.

ARTICLE INFO ABSTRACT

Keywords: Ceramic-membrane technology has been employed to recover waste heat and vapor from moist
Flat-sheet ceramic membrane gases. In addition to the hydrophilicity, pore size, and porosity of separation materials, a mem-
Heat recovery brane condenser can achieve a high recovery rate. In this study, a transport membrane condenser
Water recovery (TMC) was successfully constructed using a flat-sheet ceramic membrane (FCM) with a large pore
Transport membrane condenser
size of 45.54 μm. The membrane characteristics, including the mean pore size, pore size distribu-
Computational fluid dynamics
tion, and porosity, were investigated to determine their recovery mechanism. Their water and
heat fluxes, recovery rates, and thermal resistance were compared under several experimental
conditions. The inlet conditions significantly affect the TMC recovery performance. Higher recov-
ery fluxes were obtained at lower water temperatures, higher gas temperatures, and higher gas
flow rates. However, when the Reynolds number of water increased, the heat flux increased
rapidly, and the recovered water decreased slightly. The highest TMC water recovery rate was
84.1%. In addition, to further analyze the separation mechanisms, a two-dimensional model was
developed using ANSYS Fluent code. The simulation results were compared with the experimen-
tal data to verify the accuracy of the numerical method. The comparison results showed a high
degree of consistency.

* Corresponding author. Division of Mechanical Design Engineering, Jeonbuk National University, 567 Baekje-daero, Deokjin-gu, Jeonju-si, Jeollabuk-do, 54896,
Republic of Korea.
E-mail address: cw-park@jbnu.ac.kr (C.W. Park).
1 Co-first author.

https://doi.org/10.1016/j.csite.2024.104265
Received 20 December 2023; Received in revised form 22 February 2024; Accepted 14 March 2024
Available online 23 March 2024
2214-157X/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
V.C. Le et al. Case Studies in Thermal Engineering 56 (2024) 104265

1. Introduction
Fresh water is a critical environmental requirement, the supply of which has continuously decreased with increases in population
and the demands of everyday living [1,2]. Meanwhile, wastewater is steadily increasing, and can be considered as a valuable water
source [3]. Wastewater contributes to more than 50% of water streams in the United States [4]. Approximately 47.2% of water usage
is for power generation, and a large amount of this water is discharged into the environment as flue gas [5,6]. A typical 2 × 660 MW
power plant requires a supply of approximately 4200 m3 of water per hour, with a specific consumption of approximately
3.2 m3/MWh; however, up to 29% of the water requirement can be supplied by efficiently extracting moisture from flue gases [7].
Moreover, approximately 20% of the total energy lost during the power production process is through flue gas, which is mainly dissi-
pated by the convection of a high-temperature gas mixture and the condensation of water vapor [8,9]. Additionally, the energy effi-
ciency of a fuel-cell power plant is considerably enhanced by recovering waste vapor from the exhaust gas [10].
There are several technologies for recovering water vapor from moist gases, including condensation-based methods [11], deep
cooling towers [12], desiccant or compression dehumidification [13,14], and membrane separation [15]. In a flue-gas condenser sys-
tem, a large amount of heat is dissipated through non-condensable products, and the recovered water is relatively dirty [16]. A desic-
cant or compressor dehumidifier requires additional refrigeration or a compressor, which consumes initial operational power and re-
duces the energy efficiency of the system [17]. An optimized cooling tower can reduce the vapor content of the flue gas by more than
80%; however, the tower system requires a complex setup and a deep-water source. Membrane separation technology has been
demonstrated in huge applications and exhibited high performance by employing suitable membrane and system arrangements
[18,19].
Polymer membranes were initially developed in the late 20th century to separate vapor from air [20]. In addition to gas separa-
tion applications, gas dehydration technologies using polymer membranes have been widely and commercially applied in several in-
dustrial fields [21,22]. Polymer membrane technology comprises a massive group of materials of varying shapes, and characteristics
and has created a $100 million market in recent years [23]. A circular ceramic membrane was proposed by the Gas Technology Insti-
tute for heat and water vapor recovery, which combined a membrane condenser and heat exchanger to form a membrane heat ex-
changer or transport membrane condenser (TMC) based on the condensation and transport of vapor to a coolant flow [24]. Wang et
al. [25] developed a TMC using multilayer hydrophilic membranes under various recovery conditions. Ceramic-membrane con-
densers have been demonstrated to be highly efficient, cost effective, and easily operated at various scales (laboratory, pilot, and in-
dustry) with short payback periods [26,27]. A wide range of membrane condensers have been studied using asymmetric membranes
with nanoporous selective layers [28,29].
However, although the capillary effect offers high vapor condensation, it is important to consider the low permeability of water
owing to the nanoscale pore size. Zhang et al. [30] illustrated the effects of the pore size, porosity, and structural parameters on the
permeation rate of condensed water. Even though the condensation rate under proper conditions was very high, the permeation rate
was of limited value because of the small pore size. Moreover, microporous and mesoporous layers increase the manufacturing cost,
significantly contributing to the total cost of the system [31]. Recently, macroporous membranes have been investigated to improve
water permeability and reduce membrane manufacturing costs. Cheng et al. [32] considered a membrane with average pore size of up
to 1 μm and demonstrated that membranes with large pores have excellent potential for membrane condenser applications. Gao et al.
[33] utilized a 1 μm porous membrane and demonstrated that the macroporous membrane provided a greater water recovery effi-
ciency. Li et al. [34] demonstrated the prospects of macroporous TMCs using a pilot-scale device with a lifespan of 15 years and a pay-
back period of 4.7 years in a 330-MW power plant unit. Furthermore, Huang et al. [35] demonstrated that TMC water and heat recov-
ery are improved by a membrane with large pores; however, the pore size of the circular membrane is restricted to avoid transmem-
brane gas permeation.
Notably, all the investigated membranes for transport membrane condenser applications were tubular membranes. Plate mem-
branes have been widely used in filtration or desalination for water treatment applications [36,37]. Gunder et al. [38] conducted an
experimental comparison of a membrane separation system and concluded that the plate membrane module obtained higher flux
rates while consuming less energy than the hollow fiber and tubular modules. In addition, plate membranes have been used in air hu-
mification [39] and dehumidification processes [40]. Several types of plate membranes can be used in vacuum membrane dehumidi-
fier systems with high dehumidification efficiencies [41]. Wang et al. [42] studied different composite membranes in a TMC to re-
cover waste heat and water from hot vapor-CO2 gas. A plate-type membrane should be studied as a TMC with the expectation of high
water permeability for better water recovery performance.
In this study, a flat-sheet ceramic membrane (FCM) was employed in a TMC for the simultaneous recovery of heat and vapor from
moisture gas. An alumina membrane with a pore size of 45.54 μm was employed to improve the permeability of condensed water. A
steam generator was used to supply moisture gas for small-scale experiments. The effects of different conditions on the moisture gas
and cooling water were investigated. The heat and water recovery performance and membrane efficiencies were estimated using the
recorded experimental data. The inlet mass flow rates were low; therefore, the flows were considered to be laminar in all operating
conditions. The thermal resistance system was analyzed to study the heat transfer in the TMC. Furthermore, a two-dimensional model
was developed based on computational fluid dynamics (CFD) codes to verify the experimental results and investigate the condensa-
tion mechanisms of the FCM condenser. Lin et al. [43] considered the condensation/evaporation of vapor as a two-step reaction in a
membrane-selective layer. The volume of fluid (VOF) multiphase Lee model is widely used for evaporation–condensation prediction
[44]. Liu et al. [45] showed that the Lee model reduced the computational resource requirements and retained acceptable accuracy
for liquid–vapor phase change capture. Therefore, the Lee VOF evaporation–condensation model was used to evaluate the condensa-
tion of vapor in the moisture–gas mixture.

2
V.C. Le et al. Case Studies in Thermal Engineering 56 (2024) 104265

2. Membrane preparation and methods

2.1. Membrane preparation
A square FCM with dimensions 100 mm × 100 mm × 10 mm (L × W × H) was purchased from Zhengzhou Mission Ceramic
Products Co., Ltd., China. To investigate the water and energy recovery characteristics (i.e., from waste gas) of the commercial mem-
brane, an FCM was applied without any treatment. Polylactic acid (PLA) filaments (1.75 mm) from the Sangsangisang 3Dprinter
Company, Korea, were used to print the TMC box. In the experimental test, the FCM was embedded in the TMC; thus, the final effec-
tive size of the FCM membrane was 90 mm × 90 mm × 10 mm (L × W × H). Table 1 lists the main parameters of the FCM.
Field-emission scanning electron microscopy and energy dispersive spectrometry (ZEISS SUPRA 40VP) were used to investigate
the surface morphology and elemental composition of the FCM, respectively. The pore size distribution of the FCM membrane was
measured using the mercury intrusion method (AutoPore IV 9500 Version 2.01; Micromeritics Instrument Corporation, USA). The
crystalline phase composition of the FCM was determined by an X-ray diffractometer (XRD, Bruker, D8 Advance) with a Cu Kα radia-
tion source at λ = 0.154 nm in the 2θ range of 5–90. The functional molecular groups of the FCM were characterized using a Fourier
transform infrared spectrometer (FTIR, PerkinElmer, Frontier) in the 400–4000 cm−1 range.

2.2. Experimental setup

A schematic of the experimental setup is shown in Fig. 1. An FCM was enclosed in a TMC box consisting of two plastic blocks and
two acrylic covers. The plastic blocks were prepared using PLA in a 3D printer and were designed with several microchannels to dis-
tribute the flows of humid gas and water. The detailed configurations of the TMC box and microchannel are shown in Fig. 2. The
moisture gas flow was generated by an air compressor circulating through a steam generator into the upper block. Cooling water
flowed through the bottom block in the opposite direction as a counterflow heat exchanger. Air was controlled using a regulator and
maintained at 0.5 bar in operating conditions. A mass flow controller (MFC) and an MFC readout unit were used to control the gas
flow rate. The gas temperature was verified using a heater in the steam generator, and because of the low flow rate, the gas mixture
easily achieved a saturation state at the specified temperature. For the cooling water conditions, a water pump was utilized to control
the water velocity, and a low temperature was achieved using a cooling water bath and a plate heat exchanger. A sonic level sensor
was attached to the bottom of the water tank to track the mass changes on the water side. The water height values were then con-
verted into masses of permeated water on the cooling-water side. Thermocouples were installed in the inlet and outlet manifolds to
obtain the gas and cooling water temperatures. A humidity sensor was used to confirm the saturation conditions of the humid gas. The
experimental data (temperature, flow meter, water level, and humidity signal) were recorded using a data acquisition device con-
nected to a computer using LabVIEW. The detailed conditions and experimental variables are listed in Table 2. All manifolds and de-
vices were covered with insulation materials to prevent heat transfer to the surrounding environment.
Initially, the cooling water temperature was gradually increased until it reached a steady state. Therefore, the experimental sys-
tem was operated for 10–30 min to ensure stable conditions. The experimental variables were recorded for at least 60 min. To mini-
mize the effects of ambient conditions, the experiments were performed three times to obtain accurate average data.

Table 1
Properties of the FCM.

Properties Contents

Membrane type Flat-sheet membrane

Dimensions (L × W × H) 90 mm × 90 mm × 10 mm
Materials Al2O3 ceramic
Average pore size 45.54 μm
Porosity 52.97%
3
Density 2.0 g/cm

Fig. 1. Scheme of experimental devices and setup.

3
V.C. Le et al. Case Studies in Thermal Engineering 56 (2024) 104265

Fig. 2. FCM condenser: a) assembly; b) microchannel arrangement at inlet/outlet cross sections.

Table 2
Experimental operating conditions.

Parameter Set point Unit

Gas temperature (Tg,i) 45, 50, 55, 60, 65 °C

Gas flow rate (Vg,i) 4, 8, 12, 16, 20 LPM
Water temperature (Tw,i) 12.5, 15, 17.5, 21, 30 °C
Water flow rate (Vw,i) 0.1, 0.2, 0.4, 0.6, 0.8 LPM

2.3. Membrane performance evaluation

The heat and vapor recovery were evaluated based on the temperature and mass changes on the water side during the experiment.
The water flux (Jw, [kg/(m2h)]) was calculated based on the increase in the water mass over time, using the water level variation over
time. Water flux is expressed as follows:

ΔW
Jw = . (1)
ΔtA

Because the condensed water was transported to the cooling side, the heat transfer included the sensible heat and heat of vapor
condensation. The heat flux (qw, [kJ/(m2h)]) of the process was obtained from the increase in the water temperature and specific en-
thalpy of the recovered water as:

Cmw,i ΔT + mtr e
qw = , (2)
A

where ΔW [kg] is the mass of water increased during Δt [h], A is the effective membrane area [m2], C is the specific heat capacity of
the water [kJ/kgC], mw is the inlet flow rate of cooling water [kg/h], ΔT [°C] is the temperature difference between inlet and outlet,
mtr is the average water recovery over time [kg/h], and e is the specific enthalpy liquid water [kJ/kg] at the outlet temperature Tw,o
[°C].
To determine the effectiveness of the membrane condenser for water recovery performance under specific conditions, the water
recovery rate (γ, [%]) is defined as [24]:

Jw
𝛾= × 100, (3)
Jmax
( )
mg,i × ag,i − aw,i
Jmax = , (4)
A

Here, Jmax [kg/(m2h)] is the maximum flux of water that can be recovered with a gas flow rate of mg [kg/h]; ag,i [kg/kg] and aw,i [kg/
kg] are the water content in the gas at the specific temperatures Tg,i [°C] and Tw,i [°C], corresponding to the inlet temperature of gas
and water, respectively.
The rate of heat recovery (η, [%]) was estimated from the maximum heat flux (qmax, [kJ/(m2h)]) as:
qw
η= × 100, (5)
qmax

4
V.C. Le et al. Case Studies in Thermal Engineering 56 (2024) 104265

mg,i × (𝐻 𝑔,𝑖 − 𝐻 ′ 𝑔,𝑖 )

(6)
𝑞𝑚𝑎𝑥 = ,
𝐴
where Hg,i [kJ/kg] is the specific enthalpy of the gas at the gas inlet temperature [kJ/kg] and H'g,i [kJ/kg] is the specific enthalpy of
the gas at the water inlet temperature, which is the temperature to which the gas flow can be reduced.

2.4. Heat transfer coefficient calculation

To analyze the heat transfer in the FCM, the heat transfer coefficients (HTCs) of both the gas and water sides were evaluated. The
overall HTC (U) was calculated as:

q′ w
U= . (7)
LMTD

In addition, the overall HTC can be expressed based on thermal resistance as

1 1 t 1
= + +
UA hw A km A hg A
. (8)

Here, q'w is the heat flux in [W/m2]. the counterflow logarithm of the mean temperature difference (LMTD) between the hot hu-
mid gas and cooling water [46] is defined as:
   
Tg,i − Tw,o − Tg,o − Tw,i
LMTD =
(9)
  .
T −T
ln Tg,i −Tw,o
g,o w,i

The HTC on the water side (hw, [W/(m2K)]) was obtained from the Nusselt number (Nuw) and thermal conductivity (kw, [W/
(mK)]) of water, as:

Nuw kw
hw = , (10)
Dh

where Dh is the characteristic length of the water flow. In this experiment, the waterside length was too short relative to the length of
the hydrodynamic entrance region; therefore, the water flow was assumed to be a developing laminar flow, and the water Nusselt
number was calculated based on the Reynolds number (Re) and Prandtl number (Pr) of the water flow [47], as:
( )
D
0.03 Lh Re.Pr
Nuw = 5.51 + (( ) ) 2∕3 , (11)
Dh
1 + 0.016 L
Re.Pr

uDh
Re = . (12)
𝜈

Here, u [m/s] and ν [m2/s] are the speed and kinematic viscosity of the water flow, respectively. Generally, when condensation
occurs, the membrane voids are assumed to be filled with water. Therefore, the membrane thermal conductivity could be obtained
from the porosity (ε) and thermal conductivity of the liquid water and solid membrane (ks, [W/(mK)]) as:

km = A kw + (1 − A) ks . (13)

The HTC on the gas side (hg, [W/(m2K)]) could then be estimated as:

1
hg = 1
− t
− 1
. (14)
U km hw

The thermal resistances of the cooling water, membrane, and humid gas [K/W] were calculated as:

1
Rw = , (15)
hw A
t
Rm =
km A
, (16)

1
Rg =
hg A
. (17)

5
V.C. Le et al. Case Studies in Thermal Engineering 56 (2024) 104265

2.5. Experimental uncertainty

The method proposed by Kline and McClintock [48] was used to estimate experimental uncertainties.

  2  2  2 
(18)
 𝜕R 𝜕R 𝜕R
ER =  E1 + E2 + · · · + En ,
𝜕X1 𝜕X2 𝜕Xn

where ER is the experimental uncertainty of variable R. When R is a function of (X1; X2; …; Xn), the uncertainty of variable X1 is E1,
that of variable X2 is E2, and so on. The experimental uncertainties of the measured variables provided by the manufacturers are listed
in Table 3.

2.6. Numerical modeling

In this study, CFD codes in ANSYS Fluent were used to estimate the condensation mechanisms of the FCM condenser. The experi-
mental data were used to develop a two-dimensional model comprising a gas mixture, water, and membrane. A schematic of the three
fluid zones and meshing of the FCM condenser model are shown in Fig. 3. The mass flow inlet and pressure outlet were used as bound-
ary conditions for the gas and water flows, respectively. Membrane properties, including porosity, pore diameter, and thermal con-
ductivity, were used to define the flow in the membrane zone. The VOF multiphase model was solved using a primary phase of an air
and vapor mixture and a secondary phase of liquid water. The species transport equation in a multiphase flow was used to determine
the concentration of water vapor. A pressure-based solver using the PRESTO algorithm was implemented using a second-order
schemes. An explicit formulation was used to calculate the volume fraction in the VOF. Hence, a time step of 10−4 s was used to main-
tain the Courant numbers under a limit of 2.5 [49]. Model validation was guaranteed using condensation rates obtained under differ-
ent operating conditions. The physical model was simplified based on the experimental system with the following approximations
and assumptions.
• Gas and water flows are uniform at the inlet and outlet.
• The air–vapor mixture is homogeneous with a specific vapor mass fraction, and air and vapor behave as incompressible fluids in
the mixture.
• The membrane condenser is perfectly insulated.

Table 3
Experimental uncertainty.

Parameter Device Unit Uncertainty

Temperature 4-wire RTDs (PT100, NEWINS) 0–90 °C ±0.1 °C

Water flow rate Flow meter (M2000 Badger Meter, Germany) 0–10 LPM ±0.1%
Gas flow rate MFC readout-unit (SERO2000, SEHWA Hightech, Korea) 0–20 LPM ±0.1%
Water height Sonic level sensor (SL-40 level indicator, Picosonic, Korea) 10–10000 mm 0.1 mm
Water flux 10%
Heat flux 12.6%

Fig. 3. (a) Boundary conditions and (b) membrane condenser model.

6
V.C. Le et al. Case Studies in Thermal Engineering 56 (2024) 104265

• The membrane pore size and porosity are symmetric and isotropic, respectively, whereas the thermal properties of the membrane
are held constant.
Based on these assumptions, the following numerical model was developed: The mass transfer between vapor and liquid water
was evaluated based on the Lee evaporation–condensation model [44,50]. In the model, the phase change phenomena were consid-
ered to be the mass transfer between the liquid and vapor phases, depending on the saturation and phase temperatures. The liq-
uid–vapor mass transfer was estimated using the following transport equation:

𝜕𝜌
(19)
( )
+ ∇ ⋅ 𝜌v = m lv − m vl ,
𝜕t

where ṁlv , [kg/(s.m3)] represents the rate of mass transfer from the liquid phase to the vapor phase (evaporation), and ṁvl
[kg/(s.m3)] is related to the mass transfer from the vapor to the liquid phase (condensation). Hence, the evaporation–condensation
coefficient can be evaluated as

(20)
( )
me−c = m lv − m vl Vc .

In Eq. (20), Vc, [m3] is the volume of cell (gas, membrane, and water zones). As defined from the liquid to vapor phase, a positive
mass transfer represents liquid vaporization, whereas a negative value represents vapor condensation. Mass transfer between the two
phases can be described according to the temperature difference between phases and the specified saturation temperature Tsat [K].
When the liquid temperature Tl [K] is greater than the saturation temperature, evaporation is calculated as:
( )
Tl − Tsat
m lv = coef f evap ∗𝛼 l 𝜌l . (21)
Tsat

In contrast, when the vapor temperature Tv [K] is below the saturation temperature, the vapor is transformed into liquid water as:
( )
Tv − Tsat
m vl = coef f cond ∗𝛼 v 𝜌v . (22)
Tsat

The evaporation–condensation frequency coef f can be determined and validated based on experimental data in various operating
conditions. The condensation frequency was adjusted repeatedly to obtain a relevant condensation rate for comparison with the ex-
perimental water recovery. Thus, a value of 0.245 s−1 is established for condensation and a default value of 0.1 s−1 is used for evapo-
ration frequency.
A user-defined function (UDF) is employed for the saturation vapor temperature with reference to the polynomial approximation
given by Lowe [51]:
[ ( )]
17.2694 ⋅ Tsat − 273.15
Psat = 6.1078 ⋅ exp , (23)
Tsat − 35.86

where Psat [mb] is the saturated vapor pressure.

From the vapor condensation, the source term Se-c is added to the energy equation:

𝜕
(24)
( )
(𝜌E) + ∇ ⋅ v (𝜌E + p) = ∇ ⋅ (k∇T) + Se−c .
𝜕t

A porous media model was applied to the fluid flowing through the membrane zone. The momentum equation in the membrane
zone can be expressed as

𝜕  
  
(25)
  T
𝜌v + ∇ ⋅ 𝜌vv = −∇p + ∇ ⋅ 𝜇 ∇v + ∇v + 𝜌g+Si .
𝜕t

The momentum sink Si in the membrane is composed of viscosity loss (first term) and inertial loss (second term) [52], as:
(𝜇 )
1
Si = − vi + C2 𝜌 |v| vi , (26)
𝛼 2

where Si is the source term for the ith momentum equation (x, y, z). The flow resistance coefficients were calculated based on the
porosity and pore diameter of the experimental membrane as:

D2p A3
𝛼= , (27)
150 (1 − A)2

7
V.C. Le et al. Case Studies in Thermal Engineering 56 (2024) 104265

3.5 (1 − A)
C2 = . (28)
Dp A3

In the above, μ is the viscosity of the fluid, α and C2 are the permeability and inertial loss coefficients, respectively, and Dp and ε
are the pore diameter and the porosity of membrane, respectively.
In this study, a hybrid mesh was used with rectangular cells in the membrane cell zone, as shown in Fig. 3b. Inflation mesh layers
were created at the interfaces between the membrane and the gas/water to capture the fluid flow near the porous solid surfaces. The
mesh quality was varied to maintain the accuracy and computational cost of the independent mesh tests. Five different elemental
sizes were used in the membrane cell zone: 1 mm, 0.5 mm, 0.2 mm, 0.1 mm, and 0.05 mm. The gas outlet temperatures and conden-
sation rates during the first 20 s are compared in Table 4. A temperature deviation of less than 0.1% was obtained between element
sizes 0.2 mm and 0.5 mm, however, the condensation rate deviation was 3%. A small deviation in the condensation rate was observed
for an element size of 0.1 mm. Therefore, an element mesh size of 0.1 mm was selected for the numerical model.
Table 5 lists the parameter and boundary conditions set in the simulation model.

3. Results and discussions

3.1. Membrane characteristics
Fig. 4 shows scanning electron microscopy images and the elemental composition of the FCM. As shown in Fig. 4a, the ceramic
membrane had many pore structures; moreover, the surface was smooth without cracks, breaks, or defects. The presence of porous
structures indicates that FCMs have great potential for water–vapor separation. In addition, most of the apertures in the FCM are
greater than 1 μm, indicating that the separation process for the water–vapor mainly depends on the surface condensation and per-
meation mechanisms [53]. As shown in Fig. 4b, the energy-dispersive X-ray spectroscopy (EDX) results show that the FCM mainly
consists of C, O, Al, Si, and other elements (main elemental composition: C: 20.36%, O: 43.78%, Al: 20.65%, and Si: 9.06%). These

Table 4
Independent mesh tests.

Mesh element size (mm) Gas outlet temperature (°C) Condensation rate (kg/s)

1 325.074 1.432E-05
0.5 323.587 1.318E-05
0.2 323.298 1.28E-05
0.1 323.223 1.264E-05
0.05 323.22 1.259E-05

Table 5
Simulation boundary condition parameters.

Parameters Value

Water inlet temperature (K) 288.15

Gas inlet temperature (K) 333.15
Membrane porosity (%) 52.97
Membrane pore size (μm) 45.54
Water mass flow rate (g/s) 6.667
Gas mass flow rate (g/s) 0.075–0.375
Inlet vapor mass fraction 0.13
−1
Condensation frequency, (s ) 0.245

Fig. 4. (a) SEM images and (b) EDX analysis results of the FCM.

8
V.C. Le et al. Case Studies in Thermal Engineering 56 (2024) 104265

are the typical elements in ceramic materials. According to the EDX mappings of the FCM shown in Fig. 5, Al, O, Si, C, Na, K, Ti,
Ca, and Fe were homogeneously dispersed on the surface and internal channels of the FCM. Fig. 6 shows the pore size distribution
of the FCM verify the effects of the pore diameter on the water–vapor separation process. The average pore diameter of the FCM
was 45.98 μm. The average pore diameter of this membrane is significantly greater than 1 μm, reaffirming that surface condensa-
tion and permeation play major roles in the water–vapor separation process. In addition, the pore size distribution was narrow, in-
dicating that the apertures of the FCMs were homogeneous.
Fig. 7 shows the XRD patterns and FTIR spectra of the FCM. As shown in Fig. 7a, the diffraction peaks are well-matched to the
hexagonal phase corundum (α-Al2O3, PDF # 10–0173), based on an XRD data analysis using JADE software. The characteristic peaks
at approximately 2θ = 25.58°, 35.14°, 37.78°, 43.36°, 52.55°, 57.52°, 61.16°, 66.55°, 68.20°, 76.88°, and 84.38° correspond to the
(012), (104), (110), (113), (024), (116), (122), (214), (300), (1010), and (223) Bragg diffractions of the α-Al2O3. The sharp diffrac-
tion peaks indicate that the products were well crystallized. The higher the peak, the greater the crystallinity. No impurity crystalline

Fig. 5. EDX mappings of the FCM.

Fig. 6. FCM pore size distribution.

9
V.C. Le et al. Case Studies in Thermal Engineering 56 (2024) 104265

Fig. 7. (a) XRD patterns and (b) FTIR spectrum of the FCM.

phase is found, except for the hexagonal phase from α-Al2O3; this illustrates that the purity of the FCM is high and not polluted. Wang
et al. [54] and Zhu et al. [55] reported similar results. From Fig. 7b, it can be clearly observed that there is a wider band near
3403 cm−1, which may contribute to the presence of an amorphous structure or irregular defects. Most of the strong peaks appear be-
tween approximately 400 and 1000 cm−1 and are related to the characteristic Al–O vibration of alumina, indicating the presence of
better crystalline phases. The major peaks at 444, 594, and 642 cm−1 can be attributed to the AlO6 octahedra, which comprise the
building blocks in the α-Al2O3 structure. The peak at 1032 cm−1 is assigned to the symmetric stretching of Al–O–H [56].

3.2. Effect of cooling water

The operating conditions significantly influence the recovery behavior of the TMC [26]. These effects were considered to compare
the fluid flow in a rectangular shape with that in a conventional circular flow. Because the effective membrane area and condenser
size are different, the water flow rate ranges differ between studies. In this study, water flowed in a rectangular shape. Therefore, the
nondimensional Reynolds number of water was used as an equivalent parameter for comparison with other studies. Fig. 8 illustrates
the effects of the Reynolds number of water on the water and heat fluxes in the FCM condenser. The Reynolds number was varied
from approximately 26.4 to 185.8 by modifying the water flow from 0.1 to 0.8 LPM, respectively. It can be observed that the heat flux
increases almost twice (from 7.4 to 13.3 MJ/(m2h)) when the Reynolds number of water is increased. The enhancement in the heat
flux can be attributed to the water-side thermal resistance, as shown in Fig. 9. The heat resistance of the water flow decreases from
0.96 to 0.52 K/W. However, the water flux varies negligibly with changes in the water flow rate, and a slight decline can be observed
with a water Reynolds number of 185.8, which corresponds to a water flow rate of 0.8 LPM. This trend was also observed in a macro-
ceramic membrane condenser by Huang et al. [35] in an experimental study and by Zhang et al. [30] in a CFD numerical simulation.
The permeation rate was more likely to be influenced by the structural parameters of the membrane than by the operating parame-
ters. Based on these results, a Reynolds number of approximately 100, corresponding to a water flow rate of 0.4 LPM, was chosen for
the other experiments.
The effects of the water temperature on the heat and water fluxes are shown in Fig. 10. When the water temperature increases
from 12.5 to 30 °C, the heat flux decreases reliably from 13.2 MJ/(m2h) to 4.9 MJ/(m2h). These results were consistent with those re-
ported by Huang et al. [35]. Reducing the water inlet temperature increases the temperature difference between the moist air and

Fig. 8. Effect of Reynolds number of water (water temperature: 15 °C, gas flow rate and temperature: 8 LPM and 60 °C, respectively).

10
V.C. Le et al. Case Studies in Thermal Engineering 56 (2024) 104265

Fig. 9. Heat resistance of the water side (water flow rate: 0.4 LPM, gas flow rate and temperature: 8 LPM and 60 °C, respectively).

Fig. 10. Effect of water temperature (water flow rate: 0.4 LPM, gas flow rate and temperature: 8 LPM and 60 °C, respectively).

cooling water, thereby driving the heat transfer process. However, the variation in the water flux with the inlet temperature was neg-
ligible in this case. The water flux obtains the highest and lowest values at 5.5 and 4.6 kg/(m2h), respectively. This can be attributed
to the restriction of the condensation rate, which has been demonstrated to vary marginally with changes in coolant temperature
[30].

3.3. Effect of gas inlet temperature

Fig. 11 shows the variations in heat/water flux and recovery rates at different gas temperatures. The recovery fluxes increased
continuously with increasing gas inlet temperature. The heat flux improves from 7.4 to 13.3 MJ/(m2h), whereas the water flux
rapidly gains from 2.3 kg/(m2h) to 6.9 kg/(m2h) when the gas temperature is modified from 45 to 65 °C. Evidently, when the gas
inlet temperature is increased, the moisture mixture cannot carry more vapor molecules in saturated conditions. As the moisture
content in the gas mixture increased, the water condensation improved; therefore, the recovered water increased. Similarly, in-
creasing the gas temperature enhanced the LMTD, thereby improving heat transfer performance. Similar trends have been observed
in other studies on tubular membranes through simulations and experiments [24,30]. However, as the gas temperature increased,
the water recovery efficiency decreased slightly from 71.4% to 56.4%, whereas the heat recovery decreased from 62.7% to 38.1%.
Generally, the maximum enthalpy and vapor content gains manifest with increasing temperature [25]. These theoretical values in-
creased faster than the possible recovered fluxes. In other words, the heat and mass transfer are limited by the membrane charac-
teristics rather than by the capacity of the humid gas.

11
V.C. Le et al. Case Studies in Thermal Engineering 56 (2024) 104265

Fig. 11. Variation of (a) heat flux and water flux and (b) recovery rates with different gas inlet temperatures (water flow rate and temperature: 0.4 LPM and 15 °C, re-
spectively, and gas flow rate: 8 LPM).

3.4. Effect of gas flow rate

The recovery fluxes and efficiencies for different gas flows are shown in Fig. 12. Evidently, humid gas flow has a strong impact on
recovery performance. When the flow rate is modified from 4 LPM to 20 LPM, the heat flux is enhanced from 7.1 MJ/(m2h) to 13.7
MJ/(m2h), whereas the water flux increases from 4.5 kg/(m2h) to 6.3 kg/(m2h). A high water recovery efficiency of 84.1% was ob-
tained at a low gas velocity. However, it decreased rapidly to 29.2% when the gas flow rate increased, whereas the heat recovery rate
decreased from 46.1% to 19.5%. These results demonstrate that the FCM condenser can achieve a very high recovery efficiency with a
low gas-flow rate, but is unable to maintain high performance when the gas velocity increases. The enthalpy and vapor content of the
humid gas increased linearly with increasing flow rate. Nevertheless, the duration of the heat and mass transfer processes decreased;
thus, some heat and vapor could not be recovered as quickly as the maximum amount increased.
Table 6 shows the water and heat recovery rates of the different membranes and the membrane mean pore sizes reported in
the literature. Most TMC studies have focused on multilayered tubular membranes with selective layers. However, it has been
shown that nanoporous layers restrict water permeation and increase manufacturing costs. In this study, a flat-sheet membrane
with large pores achieved a high water recovery rate of 84.1%. Wang et al. [25] developed a TMC using a tubular membrane
with a separation layer of 6–8 nm pore size and 1 μm thickness, which resulted in water recovery from 20% to 60% and heat

Fig. 12. Variation of (a) heat flux and water flux and (b) recovery rates with different gas flow rates (gas temperature: 60 °C, water flow rate and temperature: 0.4 LPM
and 15 °C, respectively).

Table 6
Comparison of water and heat recovery rate in published literature.

Subjects Highest water recovery Highest heat recovery References

Tubular membrane 60% 85% [25]

6–8 nm
Hollow ceramic membrane 60% – [57]
20–100 nm
Tubular membrane 82% 78% [32]
1 μm
Tubular membrane 79.21% 71.71% [35]
2–3 μm
Composite plate membrane 75% 78% [42]
0.2–0.3 μm
Alumina plate membrane 84.10% 62.70% This study
45.54 μm

12
V.C. Le et al. Case Studies in Thermal Engineering 56 (2024) 104265

recovery from 33% to 85% under different conditions. Chen et al. [57] used hollow micro–nano–porous composite membranes
with pore sizes ranging from 20 nm to 100 nm. They reported that the membrane module achieved the highest water recovery
rate of 60% when the membrane pore size was 100 nm. Additionally, the preparation cost of the selective layer is reduced for
membranes with larger pore sizes. Furthermore, as previously mentioned, the recovered water is limited by the membrane per-
meation, especially with a nanoporous selective layer [30]. Cheng et al. [32] reported a relatively high water recovery rate of
82% with a 1 μm pore-sized tubular membrane. However, the water flux obtained was from 0.7 to 2.6 kg/(m2h), which is sig-
nificantly improved in our study. Huang et al. [35] explored the effect of large pore size using various pore sizes from 2 to 3 μm
by varying the sintering temperature in the fabrication process. The authors reported that the water flux increased significantly
when the mean pore size increased. A water recovery rate of 79.21% was achieved with a mean pore size of 3.29 μm, corre-
sponding to a sintering temperature of 1200 °C. Additionally, Wang et al. [42] proposed a membrane heat exchanger using plate
composite membranes with mean pore sizes of 0.2–0.3 μm to recover heat and vapor from hot vapor-CO2 gas. The composite
membranes exhibited good stability with remarkable water and heat recovery performance. Consequently, the membrane con-
denser proposed in this study provides high water permeability with a large pore size in an FCM, and shows great potential for
use in membrane heat exchanger applications. However, the heat recovery rate was lower in the present study than that re-
ported in the literature. It is worth noting that while water recovery depends on the membrane properties, heat recovery is more
affected by the thermal properties of the membrane and thermal resistance.

3.5. Heat transfer performance

The heat-transfer performance of the FCM condenser was further investigated using a thermal resistance system. The effects of gas
flow rate and inlet temperature on the resistance of the water/gas and ceramic membranes are shown in Fig. 13. It can be observed
that the overall thermal resistance of the TMC decreases with increasing flow rate and/or gas temperature. In all cases, the thermal re-
sistance of the membrane is negligible, with a value of approximately 0.03 K/W, whereas the gas-side resistance is constant at ap-
proximately 0.67 K/W. By contrast, the gas thermal resistance was affected by the inlet conditions of the gas. The vapor condensation
improved when the velocity and temperature increased. Thus, the heat-transfer performance improved as the thermal resistance de-
creased on the gas side. Notably, the water-side thermal resistance can be improved based on the water flow rate, as mentioned previ-
ously.

3.6. FCM mechanism

The experimental data were used to develop and verify the reliability of the CFD model. The condensation rate and gas outlet tem-
perature predicted by the numerical model were compared with the experimental results under different operating conditions, as
shown in Fig. 14. It can be observed that the numerical condensation rate and temperature are accurately validated according to the
experimental data. The difference between the experimental and predicted results is within the acceptable range, with a maximum
value of 5.46%.
Fig. 15 presents the simulation results for the FCM condenser at a steady state. The temperatures on the side plane are shown in
Fig. 15a and b. The temperature was distributed along the gas and water flow directions as heat was exchanged. In the gas zone, the
lower temperature toward the water side illustrates heat conduction from the hot moisture gas to water. The water temperature in-
creases along the flow direction, illustrating heat recovery by the cooling water of the TMC. Fig. 15c shows the mass fraction of vapor
in the mixture with air. The mass fraction of the vapor was well distributed along the gas side of the TMC. The vapor mass fraction de-
creases along the flow direction of the humid gas. The blue color in the water zone indicates that there is no vapor phase in the liquid
water flow. However, some concentrated vapor was detected as irretrievable vapor on the upper surface at high temperatures. For
comparison, the vapor mass fraction at a gas inlet of 0.3 g/s is shown in Fig. 15d. The vapor concentration in the gas mixture was
higher at higher gas inlet velocities. This observation confirmed the water flux and recovery rate results from the previous discussion.

Fig. 13. Thermal resistance of the FCM condenser.

13
V.C. Le et al. Case Studies in Thermal Engineering 56 (2024) 104265

Fig. 14. Comparison of the experimental and simulation results: (a) condensation rate and (b) gas outlet temperature.

Fig. 15. Numerical results for the FCM condenser i steady-state conditions: (a) temperature distribution (gas inlet: 0.15 g/s), (b) temperature distribution in membrane
and water (gas inlet: 0.15 g/s), (c) vapor mass fraction (gas inlet: 0.15 g/s), and (d) vapor mass fraction (gas inlet: 0.3 g/s).

As mentioned earlier, as the gas flow rate increased, the mass of recovered water increased, but the recovery rate decreased. The de-
velopment of the gas flow is shown in Fig. S1 and S2.
In a flat-plate membrane, the vapor molecules condense on the solid membrane surface to form a laminar liquid film, and then
move downward and penetrate the cooling water flow owing to the effects of the pressure gradient, suction on the membrane surface,
and gravity (with the high-density property of the liquid phase) [58]. Using FCM, the condensed water is covered by a membrane and
fully permeates to the water side without any retentate flow. Gravity acts as an additional promoter of the mass transport of con-
densed water, whereas a large pore size extends membrane permeability. By contrast, in a tubular TMC, the water recovery perfor-
mance is limited by the permeability of the membrane; a large amount of condensed water remains outside the membrane tube and is
considered waste [30]. Therefore, the flat-sheet membrane condenser is advantageous over the conventional circular membrane con-
denser in terms of water permeability performance. This advantage can also be observed in the significantly high recovery efficiencies
obtained at low gas flow rates.

4. Conclusions
An FCM condenser was proposed for recovering heat and wastewater from moisture gas. The effect of the operating conditions on
the recovery characteristics was investigated. Mass transfer behavior with the additional effect of gravity was studied using a CFD
model. The following conclusions were drawn.
(1) The inlet water conditions had a negligible effect on water recovery, but a significant impact on heat recovery performance.
Increases in the gas temperature and flow rate promoted recovery fluxes but reduced efficiencies.
(2) The highest water recovery rate achieved was 84.1%. The flat-sheet membrane exhibited high water recovery efficiency with
a large pore size. However, more flat-sheet membranes should be studied in future research to enhance the water recovery
and heat transfer performance.
(3) The water thermal resistance decreases with increasing water velocity whereas the gas thermal resistance decreases with
increasing gas flow rate and temperature. Membrane resistance was negligible in all operating conditions.

14
V.C. Le et al. Case Studies in Thermal Engineering 56 (2024) 104265

(4) The predicted condensation rate is in good agreement with the experimental results. The simulation results presented the
temperature and vapor mass fraction distributions.

Acknowledgements
The authors would like to appreciate the research funds for their support with this project. This research was supported by the Ba-
sic Science Research Program and Regional Innovation Strategy (RIS) through the NRF (National Research Foundation of Korea, grant
No. RS-2023-00209123, 2023RIS-008). The authors also appreciate the Technology Development Program funded by the MSS (Min-
istry of SMEs and Startups, grant No. S3276045) for their assistance. And this work was also supported by the Technology Innovation
Program (20018237) funded by the Ministry of Trade, Industry & Energy (MOTIE, Republic of Korea).

CRediT authorship contribution statement

Van Cong Le: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Software, Validation. Jun
Cong Ge: Writing – original draft, Methodology, Formal analysis, Investigation, Visualization, Writing – review & editing. Suk-Min
Seo: Software, Visualization, Writing – original draft. Dae-Hae Kim: Project administration, Conceptualization, Methodology, Re-
sources. Chan Woo Park: Writing – review & editing, Supervision, Project administration, Methodology, Conceptualization, Funding
acquisition, Investigation, Resources.

Declaration of Competing interest

All authors of this article declare that there is no conflict of interest.

Data availability
No data was used for the research described in the article.

Nomenclature
A Membrane effective area [mm2]
coef f Evaporation–condensation frequency [s−1]
C Specific heat capacity [J/(kgK)]
CFD Computational fluid dynamics
D Characteristic length [−]
Dp Mean pore size diameter [m]
e Water specific enthalpy [kJ/kg]
FCM Flat-sheet ceramic membrane
FTIR Fourier transform infrared spectrometer
H Enthalpy [kJ/kg]
h Heat transfer coefficient [W/(m2K)]
J Water flux [kg/(m2h)]
k Thermal conductivity
LMTD Logarithm mean temperature difference
LPM Litter per minute
m Mass flow rate [kg/s]
MFC Mass flow control
Nu Nusselt number [−]
PLA Polylactic acid
Pr Prandtl number [−]
R Thermal resistance [W/K]
Re Reynolds number [−]
T Temperature [°C]
t Membrane thickness [m]
TMC Transport membrane condenser
U Overall heat transfer coefficient [W/(m2K)]
VOF Volume of fluid model
XRD X-ray diffractometer

Greek symbols
δ Membrane thickness (m)
γ Water recovery rate [%]
η Heat recovery rate [%]
ε Membrane porosity [−]

15
V.C. Le et al. Case Studies in Thermal Engineering 56 (2024) 104265

α Permeability loss coefficient [−]

ΔW Mass change in water side [kg/s]
ΔT Temperature difference [K]
Δt Time variation (h)
ρ density [kg/m3]

Subscripts
g gas
i inlet
m membrane
o outlet
s solid
sat saturation
tr transfer
evap evaporation
cond condensation
v Water vapor
w Water liquid

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.csite.2024.104265.

References
[1] M. Miloloža, D. Kučić Grgić, T. Bolanča, Š. Ukić, M. Cvetnić, V. Ocelić Bulatović, D.D. Dionysiou, H. Kušić, Ecotoxicological assessment of Microplastics in
freshwater sources—a review, Water 13 (1) (2021) 56.
[2] A.J. Lynch, S.J. Cooke, A.H. Arthington, C. Baigun, L. Bossenbroek, C. Dickens, I. Harrison, I. Kimirei, S.D. Langhans, K.J. Murchie, People need freshwater
biodiversity, Wiley Interdisciplinary Reviews: Water (2023) e1633.
[3] R.R.Z. Tarpani, A. Azapagic, Life cycle sustainability assessment of advanced treatment techniques for urban wastewater reuse and sewage sludge resource
recovery, Sci. Total Environ. 869 (2023) 161771.
[4] J. Rice, P. Westerhoff, High levels of endocrine pollutants in US streams during low flow due to insufficient wastewater dilution, Nat. Geosci. 10 (8) (2017)
587–591.
[5] H. Ritchie, M. Roser, Water Use and Stress, Our World in Data, 2023.
[6] L. Ma, W. Wei, F. Sun, Y. Shi, Research on formation mechanism of typical low-temperature fouling layers in biomass-fired boilers, Case Stud. Therm. Eng. 35
(2022) 102076.
[7] R.M. Singh, P. Shukla, P. Singh, Environmental Processes and Management: Tools and Practices, Springer Nature, 2020.
[8] W. Zhang, S. Wang, L. Mu, H. Jamshidnia, X. Zhao, Investigation of the forced-convection heat-transfer in the boiler flue-gas heat recovery units employing the
real-time measured database, Energy 238 (2022) 121715.
[9] N. Fedorova, P. Aziziyanesfahani, V. Jovicic, A. Zbogar-Rasic, M.J. Khan, A. Delgado, Investigation of the concepts to increase the dew point temperature for
thermal energy recovery from flue gas, using aspen, Energies 12 (9) (2019) 1585.
[10] L. Yuan-Hu, J. Kim, S. Lee, G. Kim, H. Han, Efficiency improvement of a fuel cell cogeneration plant linked with district heating: construction of a water
condensation latent heat recovery system and analysis of real operational data, Appl. Therm. Eng. 201 (2022) 117754.
[11] A. Zajacs, R. Bogdanovics, A. Borodinecs, Analysis of low temperature lift heat pump application in a district heating system for flue gas condenser efficiency
improvement, Sustain. Cities Soc. 57 (2020) 102130.
[12] Z. Cui, Q. Du, J. Gao, R. Bie, Optimum design of a deep cooling tower for waste heat and water recovery from humid flue gas, Case Stud. Therm. Eng. 49 (2023)
103317.
[13] W. Li, Y. Yao, Thermodynamic analysis of internally-cooled membrane-based liquid desiccant dehumidifiers of different flow types, Int. J. Heat Mass Tran. 166
(2021) 120802.
[14] X. Zhang, J. Wu, Z. Li, Y. Chen, A hybrid flue gas heat recovery system based on vapor compression refrigeration and liquid desiccant dehumidification, Energy
Convers. Manag. 195 (2019) 157–166.
[15] B. Li, B. Qi, Z. Guo, D. Wang, T. Jiao, Recent developments in the application of membrane separation technology and its challenges in oil-water separation: a
review, Chemosphere (2023) 138528.
[16] J. Selivanovs, E. Vigants, V. Priedniece, I. Veidenbergs, D. Blumberga, Flue gas treatment multi-criteria analysis, Energy Proc. 128 (2017) 379–385.
[17] M.S. Büker, H. Parlamış, M. Alwetaishi, O. Benjeddou, Experimental investigation on the dehumidification performance of a parabolic trough solar air collector
assisted rotary desiccant system, Case Stud. Therm. Eng. 34 (2022) 102077.
[18] H. Sijbesma, K. Nymeijer, R. van Marwijk, R. Heijboer, J. Potreck, M. Wessling, Flue gas dehydration using polymer membranes, J. Membr. Sci. 313 (1) (2008)
263–276.
[19] W. Zhao, H. Lu, C. Li, Composite hollow fiber membrane dehumidification: a review on membrane module, moisture permeability and self-cleaning
performance, Int. J. Heat Mass Tran. 181 (2021) 121832.
[20] R.W. Baker, Membranes for vapor/gas separation, Membr. Technol. Res. Inc 25 (2006).
[21] S. Majumdar, B. Tokay, V. Martin-Gil, J. Campbell, R. Castro-Muñoz, M.Z. Ahmad, V. Fila, Mg-MOF-74/Polyvinyl acetate (PVAc) mixed matrix membranes for
CO2 separation, Separation and Purification Technology 238 (2020) 116411.
[22] Z. Tu, P. Liu, X. Zhang, M. Shi, Z. Zhang, S. Luo, L. Zhang, Y. Wu, X. Hu, Highly-selective separation of CO2 from N2 or CH4 in task-specific ionic liquid
membranes: facilitated transport and salting-out effect, Separation and Purification Technology 254 (2021) 117621.
[23] R.W. Baker, Membrane Technology and Applications, John Wiley & Sons, 2023.
[24] D. Wang, Transport Membrane Condenser for Water and Energy Recovery from Power Plant Flue Gas, Gas Technology Institute, Des Plaines, IL (United States),
2012.
[25] T. Wang, M. Yue, H. Qi, P.H.M. Feron, S. Zhao, Transport membrane condenser for water and heat recovery from gaseous streams: performance evaluation, J.
Membr. Sci. 484 (2015) 10–17.
[26] D. Wang, R. Ciora, C. Lin, Z. Ma, Flue Gas Water Vapor Latent Heat Recovery for Pressurized Oxy-Combustion, Gas Technology Inst., Des Plaines, IL (United
States), 2018.
[27] S. Zhao, S. Yan, D.K. Wang, Y. Wei, H. Qi, T. Wu, P.H.M. Feron, Simultaneous heat and water recovery from flue gas by membrane condensation: experimental
investigation, Appl. Therm. Eng. 113 (2017) 843–850.

16
V.C. Le et al. Case Studies in Thermal Engineering 56 (2024) 104265

[28] J.F. Kim, E. Drioli, Transport membrane condenser heat exchangers to break the water-energy nexus—a critical review, Membranes 11 (1) (2020) 12.
[29] Z. Li, H. Zhang, H. Chen, D. Gao, Advances, challenges and perspectives of using transport membrane condenser to recover moisture and waste heat from flue
gas, Separation and Purification Technology 285 (2022) 120331.
[30] H. Zhang, J. Zhang, Z. Liu, Z. Li, H. Chen, Simulation study of using macroporous ceramic membrane to recover waste heat and water from flue gas, Separation
and Purification Technology 275 (2021) 119218.
[31] S.K. Hubadillah, M.H.D. Othman, T. Matsuura, A.F. Ismail, M.A. Rahman, Z. Harun, J. Jaafar, M. Nomura, Fabrications and applications of low cost ceramic
membrane from kaolin: a comprehensive review, Ceram. Int. 44 (5) (2018) 4538–4560.
[32] C. Cheng, H. Zhang, H. Chen, Experimental study on water recovery from flue gas using macroporous ceramic membrane, Materials 13 (3) (2020) 804.
[33] D. Gao, Z. Li, H. Zhang, J. Zhang, H. Chen, H. Fu, Moisture recovery from gas-fired boiler exhaust using membrane module array, J. Clean. Prod. 231 (2019)
1110–1121.
[34] Z. Li, H. Zhang, H. Chen, Application of transport membrane condenser for recovering water in a coal-fired power plant: a pilot study, J. Clean. Prod. 261 (2020)
121229.
[35] J. Huang, H. Zhang, Y. Zhang, D. Liang, H. Chen, Recycle coal fly ash for preparing tubular ceramic membranes applied in transport membrane condenser,
Separation and Purification Technology 282 (2022) 119972.
[36] E. Obotey Ezugbe, S. Rathilal, Membrane technologies in wastewater treatment: a review, Membranes 10 (5) (2020) 89.
[37] U.M. Aliyu, S. Rathilal, Y.M. Isa, Membrane desalination technologies in water treatment: a review, Water Pract. Technol. 13 (4) (2018) 738–752.
[38] B. Günder, K. Krauth, Replacement of secondary clarification by membrane separation — results with tubular, plate and hollow fibre modules, Water Sci.
Technol. 40 (4) (1999) 311–320.
[39] C.-Y. Chen, Y.-H. Chang, C.-H. Li, C.-C. Chang, W.-M. Yan, Physical properties measurement and performance comparison of membranes for planar membrane
humidifiers, Int. J. Heat Mass Tran. 136 (2019) 393–403.
[40] B. Yang, W. Yuan, F. Gao, B. Guo, A review of membrane-based air dehumidification, Indoor Built Environ. 24 (1) (2015) 11–26.
[41] C.-H. Li, T.-F. Yang, J.-B. Jhang, W.-K. Li, W.-M. Yan, Physical characteristics analysis and performance comparison of membranes for vacuum membrane
dehumidifiers, Case Stud. Therm. Eng. 26 (2021) 101213.
[42] E. Wang, Y. Shang, F. Li, M. Xiao, S. Yan, Organic-inorganic composite membrane-based transport membrane condenser: performance of waste recovery from
hot gas mixture of CO2 and water vapor, Separation and Purification Technology 330 (2024) 125306.
[43] C.-X. Lin, D. Wang, A. Bao, Numerical modeling and simulation of condensation heat transfer of a flue gas in a bundle of transport membrane tubes, Int. J. Heat
Mass Tran. 60 (2013) 41–50.
[44] Z. Tan, Z. Cao, W. Chu, Q. Wang, Improvement on evaporation-condensation prediction of Lee model via a temperature deviation based dynamic correction on
evaporation coefficient, Case Stud. Therm. Eng. 48 (2023) 103147.
[45] Z. Liu, Z. Yu, W. Sun, K. Zhang, Evaluation of several liquid–vapor phase change models for numerical simulation of subcooled flow boiling, Case Stud. Therm.
Eng. 47 (2023) 103057.
[46] A.F. Mills, Heat Transfer, CRC Press, 1992.
[47] Y.A. Çengel, A.J. Ghajar, Heat and Mass Transfer: Fundamentals [and] Applications, McGraw-Hill Education, 2020.
[48] S.J. Kline, Describing uncertainties in single-sample experiments, Mech. Eng. 75 (1963) 3–8.
[49] R. Mullen, T. Belytschko, An analysis of an unconditionally stable explicit method, Comput. Struct. 16 (6) (1983) 691–696.
[50] W.H. Lee, Pressure iteration scheme for two-phase flow modeling, in: " MULTIPHASE TRANSPORT: FUNDAMENTALS, REACTOR SAFETY, APPLICATIONS,
1980, pp. 407–432.
[51] P.R. Lowe, An approximating polynomial for the computation of saturation vapor pressure, J. Appl. Meteorol. 16 (1) (1977) 100–103.
[52] U. Manual, Ansys fluent 12.0, Theory Guide (2009) 67.
[53] C. Cheng, H. Zhang, H. Chen, Experimental study on water recovery from flue gas using macroporous ceramic membrane, Materials 13 (3) (2020).
[54] C. Wang, J. Sun, C. Ge, H. Tang, P. Wu, Synthesis, characterization and lubrication performance of reduced graphene oxide-Al2O3 nanofluid for strips cold
rolling, Colloids Surf. A Physicochem. Eng. Asp. 637 (2022) 128204.
[55] C. Zhu, M. Shi, S. Yang, H.u. Zhang, Z. Yang, B. Ding, Growth of radial arranged TiO2 nanorods by direct oxidation of Al-coated In-738 superalloy, J Alloy Comp
485 (1–2) (2009) 328–332.
[56] K. Djebaili, Z. Mekhalif, A. Boumaza, A. Djelloul, XPS, FTIR, EDX, and XRD analysis of Al2O3 scales grown on PM2000 alloy, Journal of Spectroscopy 2015
(2015).
[57] H. Chen, Y. Zhou, X. Su, S. Cao, Y. Liu, D. Gao, L. An, Experimental study of water recovery from flue gas using hollow micro–nano porous ceramic composite
membranes, J. Ind. Eng. Chem. 57 (2018) 349–355.
[58] S.-A. Yang, C.o.-K. Chen, Laminar film condensation on a finite-size horizontal plate with suction at the wall, Appl. Math. Model. 16 (6) (1992) 325–329.

17