Beruflich Dokumente
Kultur Dokumente
Desalination
journal homepage: www.elsevier.com/locate/desal
School of Chemistry and Chemical Engineering, South China University of Technology, Guangzhou 510640, PR China
Institute for Sustainability and Innovation, College of Engineering and Science, Victoria University, PO Box 14428, Melbourne, Victoria 8001, Australia
Singapore Membrane Technology Centre, Nanyang Technological University, 639798 Singapore, Singapore
d
School of Civil and Environmental Engineering, Nanyang Technological University, 639798 Singapore, Singapore
b
c
H I G H L I G H T S
An implicit expression of GOR was derived to quickly evaluate the heat utilization of desalination system.
Low equivalent owrates in both sides of hollow-ber membranes are necessary for high GORs.
High GOR is accompanied by the low water productivity in integrated DCMD system.
Membranes with large heat resistances promote GOR.
Non-linearly scale-up effect reveals a higher GOR of industrial DCMD system than lab-scale one.
a r t i c l e
i n f o
Article history:
Received 20 November 2014
Received in revised form 8 January 2015
Accepted 10 January 2015
Available online 21 January 2015
Keywords:
Desalination
Direct contact membrane distillation
Heat recovery
Gain output ratio
Scale-up effect
a b s t r a c t
Aiming to optimize the system-level heat utilization, a pilot-scale direct contact membrane distillation desalination system integrated with heat recovery (DCMDHX) was studied using Aspen Plus. An implicit expression of
gain output ratio (GOR) was derived to reveal the interplay of heat utilization and process parameters including
operating conditions, module specications as well as membrane properties in the DCMDHX desalination system. Compared to operating temperatures, the feed/permeate recirculating owrates were identied as the most
inuential operational factors affecting the GOR. In the current settings, the maximal GOR of 6.0 was observed at
low and equivalent feed- and permeate-side owrates regardless of module specications. Low owrates, however, resulted in undesirable low water productivity, which was consistent with the trade-off relationship
observed between the heat utilization efciency and water recovery rate in MD. Employing membranes with
high heat-transfer resistance (low conductivity and thicker membrane wall) helped to improve the GOR up to
32%. Simulated results also showed that the GOR value increased by 1.3-fold with the preheater parameter
THX varying from 5 to 0 C. The non-linear scale-up relationship existed between the membrane area and
heat utilization (i.e., GOR) was also observed, indicating the possible uncertainty in accurately predicting the
GOR value for industrial-scale desalination systems based on lab-scale module testing.
2015 Elsevier B.V. All rights reserved.
1. Introduction
Due to the rising fresh water crisis worldwide in recent decades,
desalination technologies have drawn much attention. As a promising alternative for seawater desalination, membrane distillation
(MD) is operated at mild temperature and ambient pressure [1,2],
in which water vapor generated from the hot brine diffuses through
Correspondence to: X. Yang, Institute for Sustainability and Innovation, College of
Engineering and Science, Victoria University, PO Box 14428, Melbourne, Victoria 8001,
Australia.
Correspondence to: R. Wang, School of Civil and Environmental Engineering, Nanyang
Technological University, 639798 Singapore, Singapore.
E-mail addresses: xing.yang@vu.edu.au (X. Yang), rwang@ntu.edu.sg (R. Wang).
http://dx.doi.org/10.1016/j.desal.2015.01.013
0011-9164/ 2015 Elsevier B.V. All rights reserved.
a hydrophobic porous membrane and condensates by the cold distillate stream in direct contact MD (DCMD) mode. Compared to conventional desalination processes such as multi-stage ash distillation
(MSF), multi-effect distillation (MED) or reverse osmosis (RO) [1,3],
MD has many inherent benets: low sensitivity to salinity and high
salt rejection; low vulnerability to membrane fouling and good performance under mild operating conditions; feasibility to utilize low-grade
heat and renewable energy (e.g., geothermal heat or solar power) [4,
5]. In recent years, several pilot-scale MD desalination systems have
been developed to utilize solar energy for fresh water supply in arid regions [3,69]. Thus, such desalination technology serves dual roles in relieving global water shortage as well as energy crisis and enabling more
and more arid areas/countries to access safe desalted water [10].
81
has shown the feasibility of the user unit operation model for simulating
the module performance and evaluating process efciency in MD brine
process. Later on, further improvement was reported to establish a
more accurate transport model (user customized operation unit in
Aspen plus) in MD modeling incorporated with boundary correction
[48].
With the improved one dimensional (1-D) transport MD model reported in [48], this current work aims to explore a direct contact membrane distillation desalination system integrated with heat recovery
(DCMDHX) for leveraging the advantages of MD practicability in the
context of limited heat resource. An implicit expression of GOR was derived to conveniently correlate the DCMDHX system efciency in
terms of heat utilization with single-unit hollow ber module modeling.
A full factorial analysis was conducted to identify the operational factors
that are most inuential in system-level heat utilization in terms of
GOR. Necessary mathematical conditions were proposed for achieving
maximal GOR in a given DCMDHX desalination system. The newlydeveloped implicit GOR correlation was testied through a series of investigations such as the interplay between GOR and various process
variables (dependent or independent), including owrates, inuent
temperatures of feed and permeate streams, thermal efciency of MD
module that is strongly affected by membrane properties, as well as
water recovery rate. The concept of non-linear scale-up was proposed
for large-scale MD systems integrated with heat recovery in terms of
thermal energy evaluation.
2. Theory and methodology
2.1. DCMD hollow ber module modeling
In this study, an improved 1-D transport model was used to simulate
the heat- and mass-transfer process of DCMD modules [48], in which a
certain number of N hydrophobic PVDF hollow ber membranes with
an effective length of L are regularly packed into a cylindrical housing.
The current transport equations with boundary correction, which
showed higher accuracy in predicting the MD module performance
[48], are summarized in Table 1. In both lumen and shell sides of
DCMD module, the governing equations for mass, momentum and
energy conservation together with the wall correlation equations and
boundary conditions were solved simultaneously. Although this model
is applicable to MD module with either shell or lumen-side feeding
modes, only the latter was investigated in this study. Also, in this
model both the effects of feed concentration on the change of vapor
pressure and concentration polarization are considered negligible [50].
The current transport model has been veried previously [48], based
on an established DCMD system for a series of experimental settings, including various feed inlet temperatures, ber lengths and ow velocities. Also, the membrane properties were the same as that in previous
verication experiments. Hence, the model verication was not repeated here and the veried MD model was used as a customized unit for
Aspen owsheet simulation in the following sections.
2.2. DCMDHX desalination system
In this simulation study, an ideal heat exchanger (HX), in which the
heat transfer takes place through innitely large area and hence is not
limited by heat exchanging kinetics [42], was used as the heat recovery
unit and integrated into the DCMD desalination system to recover heat
from the returning permeate stream, namely DCMDHX. The recovered
heat could be utilized to preheat the brine feed inuent before entering
the membrane module.
A series of pilot-scale hollow ber modules were integrated into the
MD owsheet in Aspen Plus. The rst set of module specications is
given in Table 2, while three pilot-scale hollow ber modules with
various packing densities and ber lengths were used in the owsheet
simulations to correlate module performance with the GOR in the
82
Mass
Momentum
Energy
B.C.
Lumen side
Shell side
d1 v1
4
J
dz
Di;1 M
d
dv1
p 2 1
1 v21 0
dz 1
dz
dT 1
dp
4
v1 1
J
1 v1 cp;1
Di;1 H;1
dz
dz
v1 z0 v1;0
p1 z0 p1;0
T 1 z0 T 1;0
0:14
Nu 1:86Gz1=3
w
JH,1 = JMhv|TW,1 + JH,c
JH,1 = h1(T1 TW,1)
M
J H;c
T W;1 T W;2
J C M p T W;1 p
4N Do;1
d2 v2
JM
dz
D2i;2
d
dv2
p 2
v22 0
dz 2
dz
4N Do;1
dT 2
dp
v2 2
2 v2 cp;2
J H;2
dz
dz
D2i;2
v2 zL v2;0
p2 zL p2;0
T 2 zL T 2;0
1=4
Nu 0:4Re1=2 0:06Re2=3 Pr 0:4
w
JH,2 = JMhv|TW,2 + JH,c
JH,2 = h2(TW,2 TW,2)
W;1
Membrane area, Am
Packing density,
Housing Inner diameter. Di,2
Fiber length, L
Ratio of module length to housing
diameter, L/Di,2
mm
mm
#1
#2
#3
100.0
0.689
216
2153
10.0
100.0
0.502
253
2153
8.5
50.2
0.689
216
1080
5.0
W;2 T W;2
MD module, water vapor is generated from the hot feed driven by temperature difference between the feed and permeate, and transports
through the membrane wall and condenses at the cold permeate side.
As a result, heat and mass transfer in the DCMD module take place between hot feed stream S1,2 and cold permeate stream S2,1. Subsequently
the feed temperature decreases from T1,2 (stream S1,2) to T1,3 (stream
S1,3) and the permeate temperature rises from T2,1 (stream S2,1) to T2,2
(stream S2,2) along the module length. The heat gained by the permeate
stream is further utilized through the heat recovery unit HX to preheat the feedstock. In this DCMDHX process most feed concentrates
(efuent) are recycled continuously at the hot side with fresh feedstock
to maintain a given owrate W1,2; while minimal brine is discharged to
reduce environmental impacts. Similarly, at the cold side of the DCMD
module, the permeate (distillate water) is recycled to maintain a xed
permeate owrate W2,1 with continuous production of distillate. It is
assumed that no wetting occurs during operation.
2.3. Evaluation of heat utilization in MD
2.3.1. Thermal efciency of DCMD module
In the DCMDHX system, the heat transfer occurs only in the DCMD
module and HX unit respectively. The knowledge of heat transfer in the
heat exchanger has been well established to study the heat recovery in
the HX [51]. In MD, the vapor pressure difference between the feeding
and permeating sides drives the vapor to transfer across the membrane.
The overall heat ux q including latent heat (qv) of evaporation and conduction heat (qc) is accompanied with mass transfer [48]. The latent
heat is considered as the effective heat used for MD water production;
while the conductive heat through the membrane matrix caused by
transmembrane temperature difference is taken as heat loss in MD.
In DCMD, the thermal efciency , which is dened as the ratio of latent heat to the total heat input, is widely used to evaluate the effectiveness of heat utilization associated with water production [35]. Hence,
the universal expression of is given as:
dp
hv
CM
qv
J M hv
dT
dp
qv qc J h M T T
hv M
CM
M
v
1
2
dT
83
Table 3
Module specications and operating owrates for scale-up effect study of DCMDHX system (T1,2 = 80 C and T2,1 = 30 C, membrane properties: Di,1 = 0.98 mm, = 0.24 mm, kM =
0.066 kJ m1K1 and CM = 3.8 107 kg m2s1Pa1).
Module
Scale-up factor
Number of bers
Module shell diameter
Module Length
Lumen-side owrate
Shell-side owrate
Membrane area
mm
mm
kg/h
kg/h
m2
#4
#5
#6
#7
#8
#9
#10
#11
#12
#13
1
72
15
200
10
10
0.044
2
288
30
400
40
40
0.355
3
648
45
600
90
90
1.197
4
1152
60
800
160
160
2.837
5
1800
75
1000
250
250
5.542
6
2592
90
1200
360
360
9.576
7
3528
105
1400
490
490
15.21
8
4608
120
1600
640
640
22.70
9
5832
135
1800
810
810
32.32
10
7200
150
2000
1000
1000
44.33
(e.g., Reynolds number, Re) [48], the kM and are respectively the membrane conductivity and thickness, and the CM is the membrane distillation coefcient. In this current study, the averaged was determined
and the averaged temperature was used to calculate dp/dT and latent
heat in Eq. (1). When the MD system is operated at constant inuent
temperatures, the mainly depends on both the MD coefcient (CM)
and characteristics (i.e., kM and ) instead of ow conditions.
In an ideal situation, the DCMD module is treated as an adiabatic
system. The latent heat required for evaporation is provided through
the enthalpy change of the feed. As depicted in Fig. 2, a sufciently
thin element of the cross section of the DCMD module can be used to
correlate the stream property changes associated with transmembrane
water production.
The energy balance through the element can be written as:
W 1 zh1 z W 1 z zh1 z z J M hv Di;1 z qc Di;1 z:
When the limit taken as z approaches zero, Eq. (2) can be simplied and rearranged as:
dW 1 h1 J M hv qc Di;1 dz:
Substituting Eqs. (1) and (7) into Eq. (6) when the pressure drop is
negligible, the thermal efciency in the DCMD module given in Fig. 1,
can be expressed in terms of feed-side temperature change as:
W P hv
2369W P
where latent heat of 2369 kJ/kg [51] at the averaged temperature of feed
and permeate (55 C) is used in this work.
2.3.2. Calculation of gain output ratio (GOR) in DCMDHX process
simulations
As one of the most useful measures, the GOR is often used to evaluate the MD performance in terms of the specic energy required for per
kg distillate output. The benets of the DCMDHX system is to possibly
recover the thermal energy from the returning permeate stream for
raising the specic enthalpy of the feed, which is the combined stream
of fresh feedstock and brine reux, and hence signicantly reducing
Fig. 2. Heat and mass proles across a sufciently thin cross-sectional element of a DCMD
module.
84
the total heat input to the system. Subsequently, the process efciency
is greatly improved. The GOR in the DCMDHX system (Fig. 1) is dened
as the ratio of latent heat for evaporation associated with water production WP to total heat input to the heater:
GOR
W h
2369W P
P v
W 1;2 h1;2 h1;1
W 1;2 cP;1 T 1;2 T 1;1
GOR
3
!2
4L 4
hv
J
5 M
Di;1 c
G1
T
T
P;1
1;2
1;1
10
11
And the THX, reects the extent of heat recovery in the HX unit and
is also considered as an input in the given system:
T HX T 2;2 T 1;1 :
12
For an ideal heat recovery unit, an innite heat-exchanging area results in a complete recovery of sensible heat from the permeate stream.
Two scenarios were considered: 1) when the feed owrate (W1,0) is not
greater than that of the permeate stream (W2,2), the temperature of the
cold-side efuent approaches the hot-side inuent and the HICO
mode (conguration for simulating heat exchanger in Aspen Plus
[52]) was used to simulate the HX unit in the owsheet shown in
Fig. 1; 2) when W1,0 N W2,2, the temperature of the hot-side efuent
approaches the cold-side inuent in the HX unit, namely HOCI mode.
This study mainly focuses on the HICO mode for heat exchanger simulation (W1,0 = W2,2) and hence assumes a constant THX = 0 in the
following discussions, except the investigation of GOR vs. THX in
Section 3.6.
W P hv
W 1;2 cP;1 T MD T HX T 2;1 T 2;2
13
where WP/W1,2 can also be written as JM/G1, i.e., water recovery rate .
Similar to the derivation of Eq. (8), the thermal efciency can be
expressed through the permeate-side temperature change as:
W P hv
2369W P
:
W 2;1 cP;2 T 2;1 T 2;2
W 2;1 cP;2 T 2;1 T 2;2
14
1:
cP;1 T MD T HX W 1;2 GOR cP;2 T MD T HX W 2;1
15
Factor
A
B
C
D
W1,2
W2,1
T1,2
T2,1
kg/h
kg/h
C
C
Low level
High level
10
10
60
30
10000
10000
80
50
Run #
GOR
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
10
10000
10
10000
10
10000
10
10000
10
10000
10
10000
10
10000
10
10000
10
10
10000
10000
10
10
10000
10000
10
10
10000
10000
10
10
10000
10000
60
60
60
60
80
80
80
80
60
60
60
60
80
80
80
80
30
30
30
30
30
30
30
30
50
50
50
50
50
50
50
50
6.598
0.341
0.708
0.659
5.856
0.411
0.726
0.775
10.19
0.189
1.042
0.838
7.817
0.276
1.057
0.960
85
Table 5
Comparison of the GORs correlated by Eqs. (9) and (15) (HX unit and module temperature
differences THX = 0 C and TMD = 50 C, i.e., T1,2 = 80 C & T2,1 = 30 C, simulated
module #1).
W1,1 = W2,1
GOR
Relative error
(kg/h)
10
20
50
100
200
500
1000
2000
5000
10000
5.857
5.637
5.059
6.183
4.481
3.520
2.943
2.029
1.206
0.775
5.854
5.635
5.064
6.169
4.500
3.521
2.941
2.028
1.205
0.773
RMS
0.05%
0.04%
0.11%
0.23%
0.43%
0.01%
0.09%
0.03%
0.06%
0.18%
0.17%
error less than 0.17%. Therefore, using Eq. (15), the GOR of the DCMD
HX system can be accurately predicted based on the module performance under similar outputs such as water productivity WP and
thermal efciency at specied uid properties and heat exchanger
settings (TMD & THX). Thus, it is possible to evaluate the heat utilization of such system in a simpler manner to avoid performing timeconsuming owsheet simulations.
3.2. Effects of DCMD operating conditions in DCMDHX system
According to Eq. (15), the GOR in the DCMDHX system is closely
related to the operating conditions and module performance, which
largely depends on membrane characteristics and operating parameters. This section focuses on the factorial analysis to investigate the
effects of four operational factors on the heat utilization in DCMDHX
with a constant THX of 0 C.
3.2.1. Factorial analysis of factors affecting GOR
As discussed previously, in the DCMDHX system four operational
factors, i.e., T1,2 and T2,1, and W1,2 and W2,1, are the key variables affecting the total heat input and distillate output. Based on the GOR values
obtained at varying operating conditions (Table 4), factorial screening
analysis was conducted to identify the most signicant factors affecting
the GOR using statistical software Minitab.
The inuence of the four factors and their interactions (combinations) are illustrated in the Pareto diagram shown as Fig. 3, in which
the impact of each factor and interactions of factors is illustrated as
AB
B
D
AD
ABD
BD
Terms
AC
BC
Factor Name
A
W1,2
ABC
C
ACD
W2,1
ABCD
T1,2
BCD
T2,1
CD
0
Absolute effect
Fig. 3. Pareto diagram of full factorial analysis for factors affecting GOR in DCMDHX
system.
86
horizontal bars. The length of each bar is proportional to the effect of individual factor on GOR divided by its standard deviation. The bars are
also sorted by their standardized effects. Therefore, factors with the
most signicant effect on GOR can be identied. In addition, on the
Pareto chart a vertical line at 1.519 serves as a critical point to identify
factors exhibiting strong dependence (signicant effects) on GOR,
i.e., any factors with bars over the line showed statistically signicant inuence within a condence level of 95% [19].
As shown in Fig. 3, factor A (W1,2) is found to be the most inuential
parameter affecting GOR, followed by factor B (W2,1) and then AB; while
the combined factor of C (T1,2) & D (T2,1) has the least effect on GOR.
Among all factors, only the recirculating owrates A and B, and the combined AB show an absolute effect exceeding the critical line of 1.519.
Other than the ow parameters W1,2 and W2,1, the GOR values seem
to be statistically irrelevant to the operating temperatures T1,2 (factor
C) and T2,1 (factor D), which have been conrmed through the statistical
analysis in Fig. 3 provided their lower absolute effect bars than the
vertical line. This can also be explained through the GOR correlation in
Eq. (15) where a simultaneous change of TMD related terms in both
denominator and numerator (i.e., WP TMD) that weakens the impact
of temperature on GOR.
Similarly, the signicant effect of the feed-side recirculating owrate
on the GOR can be explained through Eq. (15), in which the ratio of
water productivity to recirculating owrate WP/W1,2 determines the
GOR. Dened as water recovery rate , the ratio JM/G1 is proportional
to WP/W1,2 and is critical for designing a DCMD module. Thus, the
relationship of JM and G1 was rst studied to verify the effect of the
recirculating owrate on the GOR. The simulation results are shown in
Fig. 4. Clearly, Fig. 4(A) shows an increasing trend of JM with increasing
35
30
(A)
-1
20
-2
JM (kg m h )
25
15
10
G1. This is due to the increase of MD driving force with reduced boundary layer thickness and hence lower heat-transfer resistance at higher
owrates. Subsequently, the mass/heat transfer across the membrane is
greatly enhanced. However, the rise of permeation ux JM is not proportional to the increase of mass rate G1. To further investigate the reasoning,
another graph of water recovery rate vs. G1 is given in Fig. 4(B), in which an
initial steep decline of JM/G1 at low G1 (b 103 kg m2 h1) and then a
slow decrease is observed until G1 reaches 107 kg m2 h1. This is because of the extremely low permeate rate JM associated with high transmembrane resistance at low G1; as G1 further increases, the improved
transmembrane mass and heat transfer promotes a signicant increase
of JM leading to a mild decreasing trend of water recovery rate JM/G1.
This has veried the strong impact of JM/G1 on the GOR in the DCMD
HX system.
5
0
-55
4.0
3.5
0.000
0.7325
1.465
-1
log W2,1/(kg h )
(JM/G1) x 10
2.197
3
2
2.930
3.0
3.662
4.395
2.5
5.127
5.860
2.0
1
1.5
(B)
1.0
0.0
5.0x10
1.0x10
1.5x10
1.0
1.5
2.0
2.5
3.0
3.5
4.0
-1
-2
-1
G1 (kg m h )
Fig. 4. Effect of feed-side mass rate G1 on the (A) permeation ux JM; (B) water recovery
rate JM/G1 in DCMDHX system (simulated module #1, W1,2 = W2,1, T1,2 = 80 C and
T2,1 = 30 C).
log W1,2/(kg h )
Fig. 5. The GOR colored contours at varied recirculating owrates (log scales) in DCMD
HX system (simulated module #1, TMD = 50 C with T1,2 = 80 C and T2,1 = 30 C;
HICO mode (THX = 0 C) used for HX simulations of owrate range W1,2/W2,1 1
(top-left contour) and HOCI mode for W1,2/W2,1 N 1 (bottom-right contour)).
D2i;2 nD2o;1
a lower GOR value is obtained based on Eq. (15). Similar explanation can
be used for the slightly lower curve of the system with module #3, in
which a shorter membrane length leads to a greater permeation ux
but slightly lower water productivity (WP) due to the greatly reduced
membrane area to 50 m2. Nevertheless, the GOR peaks of DCMDHX
systems with modules #1 to #3 are still considered quite similar in
shapes and heights.
Based on the above discussions for Figs. 5 and 6, the unity of owrate
ratio (W1,2 = W2,1) seems to be necessary for achieving maximum
Module #1
Module #2
Module #3
GOR
0
0.1
W1,2/W2,1
Fig. 6. GOR peaks of DCMDHX systems with modules of various geometric congurations
at varying owrate ratio W1,2/W2,1 (TMD = 50 C, simulated modules #1#3; HICO mode
(THX = 0 C) used for HX simulations of owrate range W1,2/W2,1 1 and HOCI mode for
W1,2/W2,1 N 1).
87
88
6.0
-1
(A)
5.5
(B)
5.5
-1
W12=W21=500 (kg h )
-1
5.0
GOR
W12=W21=50 (kg h )
5.0
W12=W21=5000 (kg h )
4.5
4.5
4.0
4.0
3.5
3.5
3.0
3.0
2.5
2.5
2.0
2.0
1.5
1.5
1.0
1.0
0.5
0.5
0.03
0.04
0.05
0.06
-1
0.07
0.08
0.09
0.00015
0.00020
0.00025
0.00030
0.00035
(m)
-1
kM (W m K )
Fig. 7. Effects of membrane characteristics on GOR in DCMDHX system in terms of membrane thermal resistance (A) thermal conductivity kM; (B) wall thickness (constant settings in
HX unit and module temperature differences THX = 0 C and TMD = 50 C, simulated module #1 with N = 15089, = 0.689 and L = 2153 mm).
5.5
-1
5.0
4.5
W12=W21=50 (kg h )
-1
W12=W21=500 (kg h )
-1
W12=W21=5000 (kg h )
GOR
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.6
0.7
0.8
0.9
1.0
Fig. 8. The relationship of GOR and thermal efciency in DCMDHX system by varying
thermal conductivity (solid markers) and membrane thickness (hollow markers)
(THX = 0 C and TMD = 50 C, simulated module #1 with n = 15089, = 0.689 and
L = 2153 mm, kM = 0.030.09 W m1K1, = 0.150.35 mm).
89
3.5
80
(A)
6
-1
-1
-1
-1
-1
-1
kM = 0.033 (W m K )
3.0
kM = 0.066 (W m K )
GOR
40
2.0
20
-1
2.5
WP (kg h )
kM = 0.099 (W m K )
GOR
60
1.5
0
0.0
0.5
1.0
1.5
2.0
-2
2.5
3.0
3.5
4.0
(B)
-1
JM (kg m h )
12.5
6.5%
7.5
5.0
5.5%
-1
6.0%
JM (kg h )
10.0
2.5
5.0%
0.0
10
15
20
25
30
35
40
45
AM (m )
Fig. 10. Effect of module scale-up (AM) in DCMDHX system efciency in terms of
(A) heat utilization (GOR) and water production (WP); and module performance in
terms of (B) permeation ux (JM) and water recovery rate () (THX = 0 C and
TMD = 50 C, simulated modules #4#13 in Table 3, kM = 0.066 W m1K1).
Correspondingly, Fig. 10(B) shows that the water recovery rate has
a similar increasing trend to the GOR curve a steep increment with
initial increase of membrane area AM, and the curve tends to at out
as module size increases further. Compared to the steady increase of
WP as AM increases, the owrate was varied more signicantly to maintain a constant G and hence leading to a slow increase of the recovery
rate for bigger modules. In contrast, in Fig. 10(B) the JM decreases significantly with increasing membrane area. It was well-understood that this
is mainly due to the decline of averaged transmembrane driving force as
module length and size increase simultaneously [32].
Overall, coupling the system efciency (i.e., GOR, WP in Fig. 10(A))
with module performance (i.e., JM and in Fig. 10(B)) reveals the
scale-up effect in designing a DCMDHX system. Clearly, the increase
on module size and membrane area AM greatly facilitates the improvement of system efciency as a result of higher GOR and WP. However,
the performance of the membrane module, i.e., the permeation ux, deteriorates in a larger scale module. This is consistent with the previous
experimental ndings on the effect of module size on permeation ux
[27]. In a word, different to the misconception of linear scale-up of
membrane systems, the non-linear relationship of membrane area and
heat utilization (i.e., GOR) has indicated the possible uncertainty in
accurately predicting the GOR value for industrial-scale desalination
systems based on lab-scale module testing, which usually employs
small membrane area less than 1 m2. Thus, it is anticipated that the
pilot-scale DCMDHX systems would have higher GORs than the
laboratorial ones.
90
(2)
(3)
(4)
4. Conclusions
With the aid of Aspen Plus simulations, an integrated direct contact
membrane distillation desalination system with heat recovery (DCMD
HX) was studied in the context of limited heat resources. An implicit expression of GOR was derived to conveniently correlate the DCMDHX
system efciency in terms of heat utilization using module (unit)
modeling and hence avoid complicated owsheet simulations. Factorial
analysis was conducted to identify the controlling factors for achieving
high GOR in the DCMDHX desalination system. The following conclusions can be drawn:
(5)
(6)
5.0
-1
W12=W21=50 (kg h )
-1
W12=W21=500 (kg h )
4.5
-1
W12=W21=5000 (kg h )
4.0
GOR
3.5
3.0
2.5
2.0
1.5
1.0
0
THX
Fig. 11. Effect of heat recovery efciency (THX) on heat utilization (GOR in DCMDHX system) (TMD = 50 C, DCMD module #1).
Nomenclature
AM
CM
cP
Di
Do
G
GOR
Gz
h
kM
JH
JM
L
N
Nu
p
Pr
Membrane area
Membrane distillation coefcient
Specic heat capacity
Inner diameter
Outer diameter
Mass rate in the bulk ow
Gain output ratio
Graetz number
Specic enthalpy of materials
Thermal conductivity of membrane
Heat ux
Permeation ux
Length
Hollow ber numbers
Nusselt number
Pressure
Prandtl number
m2
kg m2 s1 Pa1
kJ kg1 K1
m
m
kg m2 s1
Dimensionless
Dimensionless
kJ kg1
W m1K1
kJ m2 h1
kg m2 h1
m
Dimensionless
Dimensionless
Pa
Dimensionless
Heat ow ux of evaporation
Heat ow ux of conduction
Heat ow of conduction
Reynolds number
Temperature
Velocity
Specic volume
Mass owrate
Axial location
Greek letters
Membrane thickness
Packing density
Viscosity
Density
91
kJ m2 h1
kJ m2 h1
kJ h1
Dimensionless
K
m s1
m3 mol1
kg h1
m
Acknowledgments
m
Dimensionless
Dimensionless
Dimensionless
Pa s
kg m
Dimensionless
kJ kg1
K
92
Table A.1
Initial input settings for streams and unit blocks in DCMDHX system with module #1 in
Table 1.
Unit type
Settings
FEED
STREAM
S11
STREAM
S20
STREAM
FSLIT1
FSLIT2
CHILLER
HEATER
HX
FSLIT
FSLIT
HEATER
HEATER
HEATX
PUMP1
DCMD
PUMP
USER2
References
[1] K.W. Lawson, D.R. Lloyd, Membrane distillation, J. Membr. Sci. 124 (1997) 125.
[2] X. Yang, A.G. Fane, R. Wang, Membrane Distillation: Now and Future, Scrivener/
Wiley, 2013. 2013.
[3] M.S. El-Bourawi, Z. Ding, R. Ma, M. Khayet, A framework for better understanding
membrane distillation separation process, J. Membr. Sci. 285 (2006) 429.
[4] C. Charcosset, A review of membrane processes and renewable energies for desalination, Desalination 245 (2009) 214231.
[5] E. Curcio, E. Drioli, Membrane distillation and related operations a review, Sep.
Purif. Rev. 34 (2005) 3586.
[6] Z.W. Ding, L.Y. Liu, M.S. El-Bourawi, R.Y. Ma, Analysis of a solar-powered membrane
distillation system, Desalination 172 (2005) 2740.
[7] P.A. Hogan, Sudjito, A.G. Fane, G.L. Morrison, Desalination by solar heated membrane distillation, Desalination 81 (1991) 8190.
[8] J. Koschikowski, M. Wieghaus, M. Rommel, Solar thermal-driven desalination plants
based on membrane distillation, Desalination 156 (2003) 295304.
[9] Z.S. Wang, Z.L. Gu, S.Y. Feng, Y. Li, Applications of membrane distillation technology
in energy transformation process-basis and prospect, Chin. Sci. Bull. 54 (2009)
27662780.
[10] M. Hightower, S.A. Pierce, The energy challenge, Nature 452 (2008) 285286.
[11] L. Song, B. Li, K. Sirkar, J. Gilron, Direct contact membrane distillation-based desalination: novel membranes, devices, larger-scale studies, and a model, Ind. Eng.
Chem. Res. 46 (2007) 23072323.
[12] P. Wang, M.M. Teoh, T.-S. Chung, Morphological architecture of dual-layer hollow
ber for membrane distillation with higher desalination performance, Water Res.
45 (2011) 54895500.
[13] M. Khayet, T. Matsuura, J.I. Mengual, Porous hydrophobic/hydrophilic composite
membranes: estimation of the hydrophobic layer thickness, J. Membr. Sci. 266
(2005) 6879.
[14] B. Li, K.K. Sirkar, Novel membrane and device for vacuum membrane distillationbased desalination process, J. Membr. Sci. 257 (2005) 6075.
[15] M. Khayet, T. Matsuura, J.I. Mengual, M. Qtaishat, Design of novel direct contact
membrane distillation membranes, Desalination 192 (2006) 105111.
[16] M. Qtaishat, D. Rana, M. Khayet, T. Matsuura, Preparation and characterization of
novel hydrophobic/hydrophilic polyetherimide composite membranes for desalination by direct contact membrane distillation, J. Membr. Sci. 327 (2009) 264273.
[17] M. Su, M.M. Teoh, K.Y. Wang, J. Su, T.-S. Chung, Effect of inner-layer thermal conductivity on ux enhancement of dual-layer hollow ber membranes in direct contact
membrane distillation, J. Membr. Sci. 364 (2010) 278289.
[18] Y. Liao, R. Wang, A.G. Fane, Engineering superhydrophobic surface on poly(vinylidene
uoride) nanober membranes for direct contact membrane distillation, J. Membr.
Sci. 440 (2013) 7787.
93
[52] AspenTech, Aspen Plus V7.3 User Guide, Aspen Technology, Inc., Cambridge, MA.
USA, 2011.
[53] A.M. Alklaibi, N. Lior, Heat and mass transfer resistance analysis of membrane distillation, J. Membr. Sci. 282 (2006) 362369.