International Journal of Heat and Mass Transfer 79 (2014) 241250

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

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

Numerical investigation of ammonia falling lm absorption

outside vertical tube with nanouids
Liu Yang a,b,, Kai Du a,b, Xiaofeng Niu c, Yuan Zhang b, Yanjun Li b
a

Key Laboratory of Energy Thermal Conversion and Control of Ministry of Education, Southeast University, Nanjing 210096, China
School of Energy and Environment, Southeast University, Nanjing 210096, China
c
College of Urban Construction and Safety Engineering, Nanjing University of Technology, Nanjing 210009, China
b

a r t i c l e

i n f o

Article history:
Received 24 August 2013
Received in revised form 4 August 2014
Accepted 6 August 2014
Available online 27 August 2014
Keywords:
Ammonia water
Numerical
Falling-lm absorption
Nanouid

a b s t r a c t
In the last years the ammonia-water refrigeration cycle has been gradually improved by using nanouid
technology. In this work, a numerical model for the absorption of ammonia on a falling lm was developed for different ammonia-water mixtures containing nanoparticles and dispersants. The variation of
both falling lm thickness and physical properties of the mixture were considered to nd the best tting
model. Results show that when absorption pressure decreases or when initial concentration of mixture
increases, the relative intensity of effect on absorption rate is weakened by the variation of thermal
conductivity but enhanced by the variation of mass transfer coefcients and ow resistance, while the
variation of mixtures viscosity exhibits very low effect. When the results are compared to similar experimental data for the ammonia falling lm absorption with nanouids, it was found that the heat and mass
transfer are mainly affected by the lm drag reduction and its physical properties. The numerical model
obtained can be used for calculating the absorption rate of ammonia-water-nanouid mixtures with
acceptable accuracy, since 87% of relative errors are lower than 20%.
2014 Elsevier Ltd. All rights reserved.

1. Introduction
Ammonia water absorption refrigerators have recaptured scientists attention due to the phenomenon of global warming and
ozone depletion. However, the performance of the absorption cycle
needs to be improved because it is lower than the performance of
the vapor compression refrigeration system. Since the size and performance of the absorber can signicantly impact on the systems
overall performance, the latest research have focused on the
absorption enhancement unit. Generally, there are three methods
to enhance the efciency of heat and mass transfer of absorbers:
the mechanical treatment, the chemical treatment, and nanotechnology [1].
A nanouid is dened as a liquid suspension of particles whose
diameter are smaller than 100 nm. As a result of the limitation of
heat transfer in a working uid, solid particles are dispersed
throughout it to improve its thermal properties as well as its heat
transfer coefcient. In recent years, nanouids have been gradually
used as engineering uids because of their superior performance
Corresponding author at: Key Laboratory of Energy Thermal Conversion and
Control of Ministry of Education, Southeast University, Nanjing 210096, China. Tel.:
+86 25 83793214.
E-mail address: windy4ever@163.com (L. Yang).
http://dx.doi.org/10.1016/j.ijheatmasstransfer.2014.08.016
0017-9310/ 2014 Elsevier Ltd. All rights reserved.

on stability [2,3], thermal conduction [46], thermal convection

[710] and boiling heat transfer [1116]. Recently, the nanouids
applied in ammonia water absorption refrigeration system have
involved Cu, CuO [17], CNTs [18], Ag [19], Fe2O3, ZnFe2O4 [20],
Al2O3 [21], and even nano emulsier [22]. The performances of
those kinds of nanouids on absorbing ammonia were investigated
for bubble or falling lm absorption, and the results are
satisfactory.
According to the research results of Kang et al. [23], the mass
transfer coefcients has greater effect on systems performance
running in bubble mode than that in falling lm mode. Besides
this, the heat transfer coefcients have more signicant effects
on heat exchanger size (absorption rate) in the falling lm mode
than in the bubble mode. The uid ow and heat transfer aspects
of nanouids have been studied by many researchers. However,
the researches on absorption process of ammonia by nanouids
are relatively decient. A few existing literatures are mainly
focused on the experimental studies with a predominating portion
of bubble absorption type. A latest theoretical study on ammonia
bubble absorption of nanouids has been just found [18]. Nevertheless, to the authors best knowledge, the theoretical study on
ammonia falling lm absorption of nanouids has not been found.
For this reason, there is a great need of theoretical research on

242

L. Yang et al. / International Journal of Heat and Mass Transfer 79 (2014) 241250

Nomenclature
u
v
g
Cp
Dm
T
P
K
mab

Dhab
M
S
Re
Sc

lm velocity in the direction of falling, m  s1

lm velocity in the direction of lm thickness, m  s1
gravity acceleration, m  s2
constant-pressure specic heat, kJ  kg1  K1
diffusion coefcient, m2  s
temperature, C
absorption pressure, MPa
mass transfer coefcient
absorption amount in the interface of the control volume in unit time, kg  s1
heat of absorbing unit mass of ammonia gas, kJ  kg1
gross absorption in unit time, g  s1
mass transfer interface area of falling lm, m2
Reynolds number
Schmidt number

Greek letters
q
density, kg  m3
C
ow rate, m3  s1
n
mass concentration of uid, %
d
lm thickness, mm
k
thermal conductivity, W  (m  K)1
g
dynamic viscosity, Pa  s

ammonia-water falling lm absorption with nanouid and hereby

obtain the main variables of ammonia-water nanouid falling lm
absorption. In this work, a numerical model for an ammoniawater-nanoparticles falling lm absorption outside vertical tube
was developed. The variation of falling lm thickness along the
tube and the changes in physical properties of ammonia-water
mixture was considered when adding nanoparticles and dispersants. The inuence of physical properties of nanouids over the
efciency of ammonia-water absorption is studied in detail.
Finally, the numerical results were compared with experimental
data. It is expected that this study brings some basic ideas that help
to understand how mixtures physical properties affect on the heat
and mass transfer coefcients in the absorption process and also to
establish some theoretical foundation for further research on the
application of nanouids.

s
m

effective ow time in absorption, s

kinematic viscosity, m2  s1

Super/subscripts
i
vaporliquid interface
in
inlet of falling lm
w
cooling water
r
relative
n
nanouid
f
baseuid
ff
falling lm
s
solutions in containers and distributor
a
Fig. (a)
b
Fig. (b)
c
Fig. (c)
Abbreviations
AS
absolute slope of the tting straight line about absorption rate
RS
relative slop, dened by the ratio of absolute slope to
the absorption rate when kr (or gr, Kr, sr) = 1

absorber all through the experiment, the conguration selected

for the absorber is that of falling lm on the external surface of a
vertical tube in our previous experimental study [20]. And in this
paper, the combined and follow-up theoretical study also adopts
this selection.
The process of ammonia falling lm absorption outside vertical
tube with nanouids is shown in Fig. 1. The absorber unit consists
of the top shell and the falling lm tube. The ammonia water (or
nanouid) solution enters the absorber from the top and then it
forms a lm in the distributor, and nally it falls along the exterior
surface of the heat-transfer tube. Ammonia gas enters to the absorber from the bottom and is evenly distributed in the absorber. Thus
absorption heat is generated as a result of the absorption of ammonia vapor by the solution. Cooling water enters from the bottom of

2. Model descriptions
The physical model and mathematical model introduced herein
are similar to our previous studies on numerical model of falling
lm absorption with ammonia-water affected by a magnetic eld
[24]. The difference in this work is that the magnetic eld is
replaced by a nanouid, considering the changes in the mixtures
ow and its physical properties when nanoparticles and dispersants are added. The detailed description about the physical model
and mathematical model are presented at the end of this section.
2.1. Physical model
In conventional water-cooled absorption cooling devices, the
solution ow takes place on the external surface of horizontal
tubes in conventional falling lm conguration of the absorber.
However, wettability of the falling lm affects the heat and mass
transfer performance greatly. In order to observe, check and adjust
the wettability of the falling lm all through the experiment, the
shell of the main body of the absorber is made up of transparent
acrylic glass. Therefore, to observe the conditions of solution lm
distribution in real-time through the transparent shell of the

Fig. 1. Sketch of ammonia-water nanouid falling lm absorption.

243

L. Yang et al. / International Journal of Heat and Mass Transfer 79 (2014) 241250

the heat-transfer tube, and ows upwards in countercurrent ow

to the falling lm. It also removes the heat generated in absorption
process to ensure an optimal working condition at this stage [24].

variation in thickness of falling-lm in the absorption process.

The details of this mathematical solution can be obtained from
reference [24,25].

2.2. Mathematical model

3. Simulation results and discussion

The mathematical model was developed on the following

assumptions:

The main parameters of the falling lm and the tube the following: the outer and inner diameter of the tube is 25 and 22 mm,
respectively. The length is 1000 mm. The mixing area is evenly
divided into 2000  2000 mesh (in the direction of falling  in the
direction of lm thickness). A standard operating condition is
employed to investigate the inuence of operating conditions
and the physical properties of nanouids on the heat and mass
transfer of ammonia falling lm absorption process. Parameters
of the standard operating condition are listed in Table 1.
The standard operating condition is dened according to the
previous experimental operating conditions [20]. The experimental
operating condition is set to ensure the absorption process continuously proceed in all section of the falling lm along the tube.
Otherwise, if the solution is saturated and the absorption process
stops before the end of the tube, the enhancement of nanouids
on the absorption performance cannot be veried from the saturated solutions. When the absorption process continuously proceed in all section of the tube, the operating condition is set to
obtain a more obvious difference results between the ammonia
water and ammonia water nanouid.
To study the effect of the physical properties of nanouids on
the performance of absorption, it was dened a set of the following
relative variables: relative thermal conductivity, relative viscosity,
relative mass transfer coefcient and relative ow resistance of
nanouids. Their corresponding equations are:

(1) The ammonia water (or nanouids) cannot be compressed.

(2) The nanoparticles are dispersed evenly in the liquid, and the
inuence of bubbles is neglected.
(3) There are no molecular diffusion and heat conduction in the
direction of falling neither in radial plane of the tube.
(4) The pressure of ammonia vapor remains constant and the
heat conduction in this phase is neglected.
(5) All phases in the system are in thermodynamic equilibrium,
the heat generated at the absorption stage is completely
transferred throughout the vaporliquid interface and there
is no viscous shearing stress at this interface.
(6) The variation in the cooling waters temperature is linear
along the heat-transfer tube. The temperature difference
between cooling water and the outside wall of tube is set
as an arbitrary value.
The coordinate system is set as shown in Fig. 1. The X axis is
along the falling direction, and the Y axis is along the thickness
direction of the falling lm. The instantaneous velocity components in X and Y axis through the control volume are u and v,
respectively. In a time interval of dt, the quantities generated in
the control volume can be considered as zero. The continuity equation, momentum equation, energy equation and quality equation is
described respectively as follows [24]:

@qu @qv

0
@x
@y

qu

@u
@u
@
qv

@x
@y @y

1



@u
qg
@y



@n
@n
@
@n
qv

qDm
@x
@y @y
@y

The boundary conditions are expressed as follows:

At the inlet of the falling lm solution:

x 0; u u0 uin ;

v 0;

T T in ; n nin

At the outside wall of the heat transfer tube:

y 0; u v 0; T T w ;


@n
0
@y y0

At the vaporliquid interface:



@u
@T 
0; Pv FT i ; ni const; k 
@y yd
@y yd


qDm @n
mab Dhab jyd ; mab

1  n @yyd

gr gn =gf

K r K n =K f

10

sr sn =sf

11



@T
@T
@
@T
qC p u qC p v
k
@x
@y @y
@y

qu

kr kn =kf

y d;

To obtain the mathematical solution of these equations, the

following variables were considered: the properties of ammoniawater nanouids at different ammonia concentration, the phenomenon of convection along the thickness direction of lm and the

3.1. Inuence of nanouids thermal conductivity

Fig. 2 shows the inuences of thermal conductivity on the
absorption rate in different working conditions. It can be seen that,
whatever it is, in reference condition (a), or in 0.2 MPa absorption
pressure (b), or in 15% initial concentration of falling lm (c), the
inuence of thermal conductivity has very low impact on the
absorption since it increases less than 1% when the thermal conductivity of nanouids increases by 40%. This small effect may be
caused due to the fact that the lm is very thin and its thermal
resistance is much smaller than the entire thermal resistance.
The variation of the thermal conductivity of falling lm has little
effect on the performance of overall heat transfer coefcient and
the absorption performance. The above mentioned considerations
are applied just for the case of thin falling lm absorption outside
a vertical tube, but not for other kind of absorption processes, such
as bubble absorption, horizontal tubular absorption, etc.

Table 1
The operating conditions employed as a standard of reference.
Item

Tin (C)

Cin (m3/s)

nin (%)

Tw (C)

Cw (m3/s)

P (MPa)

Value

20

0.00001

15

0.00007

0.1

L. Yang et al. / International Journal of Heat and Mass Transfer 79 (2014) 241250

0.4504

0.734
0.4500
0.4496

0.733

0.4492
1.0

1.1

1.2

1.3

0.928

0.927

1.484
1.483

0.926
1.482
0.925

1.4

0.9

1.0

0.4050
0.2460
0.4045

0.2458
1.1

1.2

AS
-1

0.2462

1.0

Heat of absorption (kJs )

-1

Absorption rate (gs )

0.4060

0.4055

0.9

1.2

1.3

1.4

1.481

(b)

heat of absorption
bsorption rate
Linear fit of absorption rate

0.2464

1.1

(a)
0.2466

1.486
1.485

1.3

1.4

(c)

0.007

A Sa
A Sb
A Sc

RS

0.9

1.487

heat of absorption
absorption rate
Linear Fit of absorption rate

Heat of absorption (kJs-1)

Heat of absorption (kJs-1)

Absorption rate (gs-1)

0.929

0.735

heat of absorption
absorption rate
Linear fit of absorption rate

0.4508

Absorption rate (gs-1)

244

RSa
RSb
RSc

0.00 7

RSb=0.00636

RSa=0.00471

RSc=0.00451

A Sc=0.00589
0

ASa=0.00111
A Sb=0.00212
0

(d)

Fig. 2. Inuences of thermal conductivity on the absorption rate in different working conditions. (a): In reference condition; (b): in 0.2 MPa absorption pressure; (c): in 15%
initial concentration of falling lm; (d): comparisons of AS and RS.

To further compare the inuence of thermal conductivity on the

absorption performance at different operating conditions, absolute
slope (AS) was dened as the slope of the tting straight lines on
absorption rate. According to this, ASa, ASb and ASc in Fig. 2(d) represents the absolute slope of the tting straight lines of absorption
rate in Fig. 2(ac), respectively. It can be concluded that the absolute slope reects the absolute rate of change of absorption with
respect to the relative thermal conductivity. However, as a result
of different operating conditions showed in Fig. 2(ac), different
base numbers on absorption rate can be obtained. Therefore the
direct absolute slope comparison can not reect the relative intensity of the effect of thermal conductivity on the absorption performance. Consequently, it was necessary to dene the relative slope
(RS) as the ratio of absolute slope to the absorption rate when kr (or
gr, Kr, sr) = 1 for each different operating condition. Hence in
Fig. 2(d), RSa represents the ratio of ASa to the absorption rate when
kr = 1 and similar reasoning is applied for the denition of RSb and
RSc. This denition is also applied to other physical properties (viscosity, mass transfer coefcient, ow resistance) throughout this
paper. It can be concluded that RS reects the percentage of change
of absorption with respect to the relative thermal conductivity.
Fig. 2(ac) shows the trend of AS. A comparison of these two
slopes (AS and RS) is shown in Fig. 2(d), where it can be seen that
the operating condition with the highest absorption rate has also
the highest AS. This is because it has highest cardinal numbers
on absorption rate value. Therefore, the AS parameter can reveal
the rate of change of absorption with respect to the relative
thermal conductivity in each working condition, but again, it cannot be used for comparing the inuence of thermal conductivity at
different operating conditions. However, it can be seen from the
picture of Fig. 2(d) that the RSb has the largest values. This reveals
that the enhancement of thermal conductivity improves the
absorption rate at higher absorption pressure conditions. And the

initial mixtures concentration has little effect, because there is

no signicant difference between RSa and RSc.
3.2. Inuence of nanouids viscosity
Fig. 3(ac) shows the inuence of nanouids viscosity on the
absorption rate at the reference condition, 0.2 MPa absorption
pressure and 15% initial mixtures concentration, respectively. It
can be seen that the viscosity has negative impact on the absorption performance and its inuence intensity is much greater than
that is shown by thermal conductivity under all working conditions. The absorption rate decreases more than 20% when the viscosity increases by 30%. This conrms that the viscosity is an
important parameter. Fig. 3(d) shows the comparison of AS and
RS of Fig. 3(ac). It can be seen that the operating condition with
the highest absorption rate has the highest AS as a result of it
has the highest base number on absorption rate. However, it can
be seen from the picture of Fig. 3(d) that the working conditions
hardly affects the inuence of viscosity of nanouids on absorption
rate because the RS of all working conditions are almost equal.
3.3. Inuence of nanouids mass transfer coefcient
Fig. 4(ac) shows the effect of mass transfer coefcient of
nanouids on the absorption rate at the reference condition of:
0.2 MPa absorption pressure and 15% initial mixtures concentration of falling lm, respectively. It can be seen that under all operating conditions, the mass transfer coefcient has a positive impact
on the absorption performance and its inuence intensity is a little
weaker than that is shown by viscosity, but considerably higher
than that exhibited by thermal conductivity. Fig. 4(d) shows the
comparison of AS and RS of Fig. 4(ac). It can be seen that AS varies
in the same way as thermal conductivity and viscosity. However,

245

L. Yang et al. / International Journal of Heat and Mass Transfer 79 (2014) 241250

0.50
0.7

0.45
0.40

0.6
0.9

1.0

1.1

1.2

1.6
1.0

1.5
1.4

0.9

1.3
0.8
0.7

1.3

1.2
0.8

0.9

0.40

0.25

0.35
0.20
1.1

1.2

1.1

1.3

1.2

1.3

A Sc=-0.1912

-1

0.45

1.1

0.30

(c)

A Sa=-0.3526

A S b=-0.7174

AS

0.30

1.0

Heat of absorption (Js )

-1

Absorption rate (gs )

0.50

absorption rate
heat of absorption

0.9

1.0

(b)

(a)

0.8

1.7

-1

-1

1.1

-1

A Sa
A Sb
A Sc

RS

0.8

1.8

absorption rate
heat of absorption

Heat of absorption (Js )

0.8

Absorption rate (gs )

-1

0.55

0.35

1.2

0.9

absorption rate
heat of absorption

Heat of absorption (Js )

-1

Absorption rate (gs )

0.60

-1

R Sa
R Sc
R Sb

R Sa= -0.7837
R Sb=-0.7741
R Sc= -0.7766

(d)

Fig. 3. Inuences of viscosity of nanouids on the absorption rate in different working conditions. (a): In reference condition; (b): in 0.2 MPa absorption pressure; (c): in 15%
initial concentration of falling lm; (d): comparisons of AS and RS.

the variation of RS showed in the picture of Fig. 4(d) is very different form the two previous nonouids physical parameters (thermal conductivity and viscosity). RSc has the highest value, which
reveals that incrementing the mass transfer coefcient has greater
impact on the absorption performance, under the operating condition of higher initial concentration of nanouids. It also reveals
that the inuence of absorption pressure on the absorption rate
is not signicant, because RSa is just a little bigger than RSb. Therefore, in plain words, the mass transfer coefcient has greater effect
under the working condition of weaker absorption performance. A
possible reason for this circumstance can be explained as follows:
when the absorption performance is weaker, the mass transfer
coefcient is the main factor of resistance to the absorption process. The enhancement of mass transfer coefcient can play an
important role when absorption is weak. However, when the
absorption performance is higher, the mass transfer coefcient
takes a secondary role in resisting absorption. Consequently, the
enhancement of mass transfer coefcient has less importance at
that moment.

the decrease in ows velocity and its volumetric rate. Hence the
Reynolds number decreases, which eventually causes the diminution of heat and mass transfer performance.
The comparison between the slopes AS and RS, shown in
Fig. 5(ac), is resumed in Fig. 5(d). It can be seen that AS is almost
proportional to the base number of absorption rate of each
operating condition. The same behavior is observed for thermal
conductivity, viscosity and mass transfer coefcient. In the insert
picture of Fig. 5(d), it is clear that the operating condition has similar inuence over the RS to that of thermal conductivity but different to that of mass transfer coefcient. It can also be seen In
Fig. 5(d) that RSb has the highest value, which reveals that the
relative ow resistance has greater impact on absorption performance at higher pressure. The absolute value of RS is very high
and the ows resistance has more impact on absorption than
mixtures other physical properties. Therefore, the absorption performance can be more greatly improved in higher absorption pressure. Especially for some kind of nanouids which cause a drag
reduction effect when are combined with a surfactant [26].

3.4. Inuence of nanouids ow resistance

4. Comparison between simulation results and experimental

data

Fig. 5(ac) shows the inuence of nanouids ow resistance on

the absorption rate at the reference condition: 0.2 MPa absorption
pressure and 15% initial mixtures concentration. The relative ow
resistance is dened by the ratio of effective ow time of the nanouid compared to the baseuids ow time. It can be seen that the
relative ow resistance has a very negative impact on the absorption performance and its inuence intensity is the greatest on all
mixtures physical parameters. Also the absorption rate decreases
about 50% when the relative ow resistance increases by 30%.
The greatest impact of ow resistance on the absorption performance is mainly due to the high ow resistance, which causes

The numerical results were compared with some experimental

data for ammonia falling lm absorption outside vertical tube with
nanouids to verify the accuracy of the mathematical model. The
experimental apparatus, procedure, parameters, physical nanouids properties and calculation methods have been demonstrated
in our previous research [24].
Fig. 6 shows the schematic diagram of the experimental device
for NH3/H2O nanouid falling lm absorption process [20]. The
absorption takes place simultaneously at the interface of solutions
in the containers and the distributor. For this reason, the

L. Yang et al. / International Journal of Heat and Mass Transfer 79 (2014) 241250
1.10

0.75

0.44

0.70
0.65

0.40
0.8

0.9

1.0

1.1

1.2

1.3

0.90

0.8

0.9

1.0

1.1

Kr

Kr

(b)

0.40

0.36
0.22
1.1

1.2

1.3

ASa
ASb
ASc

1.3

1.3

RSa
RSb
RSc
RSa=0.45046
RSc=0.50884
RSb=0.43202

ASb=0.40034

ASa=0.20265
ASc=0.12525
0

Kr

1.2

RS

AS

-1

0.44

0.24

1.0

Heat of absorption (kJs )

-1

Absorption rate (gs )

1.4

0.85

0.60

0.26

0.9

1.5

0.95

(a)
absorption rate
heat of absorption

0.8

1.6

1.00

0.28

-1

-1

1.05

1.7

absorption rate
heat of absorption

Heat of absorption (Js )

0.48

-1

0.80

Absorption rate (gs )

0.85

absorption rate
heat of absorption

Heat of absorption (kJs )

0.52

-1

Absorption rate (gs )

246

(c)

(d)

Fig. 4. Inuence of mass transfer coefcient of nanouids on the absorption rate in different working conditions. (a): In reference condition; (b): in 0.2 MPa absorption
pressure; (c): in 15% initial concentration of falling lm; (d): comparisons of AS and RS.

experimental data should be trimmed to show only information of

the falling lm absorption process. After this, numerical and experimental results can trustily be compared.
The following equations were used to perform the trimming of
experimental data:

M Mff M s

12

M ff K ff Sff

13

M S K S SS

14

Due to the slow variation of liquid level in containers and in the

distributor (can be considered approximately as static mass
transfer). The falling lms mass transfer coefcient at very low
Reynolds Number (Res = 49), whose corresponding ow rate is also
very slow (can also be considered approximately as static mass
transfer), were used to conveniently replace the mass transfer
coefcient at the interface of the solutions in containers and distributor. As a result, the ratio of mass transfer coefcient of falling
lm and solution in container and distributor were obtained by
following two equations [27]:

K ff
D

v2

1=3

9:777  104 Re0:6804

Scff1=2
ff

For 1600 < Re < 10500

15

Ks
D

v2
g

1=3

1:099  102 Re0:3955

Sc1=2
s
s

For 49 < Re < 300

16

The gross absorption can be calculated by the nal trimming

equation, which has the following form:

Mff

9:777  104 Re0:6804

Sc1=2
ff
ff  Sff
1=2

9:777  104 Re0:6804

Scff  Sff 1:099  102 Re0:3955
Scs1=2  Ss
ff
s

17
In our previous experimental studies, to obtain the suitable
nanouids applied in ammonia absorption refrigeration system,
20 types of nanoparticles mixed pairwise orthogonally with 10
types of surfactants were added in ammonia-water, respectively
to observe the dispersion stability of suspension and hereby obtain
the functioning surfactant for each type of nanoparticles [28]. Then
three kinds of nanoparticles (Al2O3, ZnFe2O4, Fe2O3) with SDBS as
dispersant were selected and employed in the comparative experiments of ammonia absorption based on an overall consideration of
stability, viscosity and other physical properties [20]. The preparation, selection and ingredient of nanouids, as well as the comparative experimental results can be referenced in our previous
experimental studies about preparation and ammonia absorption
performance of nanouids [20,2931].
Fig. 7 shows the absorption rate at different initial ammonia
concentrations in order to make possible the comparison between
the model data, raw experimental data, and trimmed experimental
data. When is considered the absorption in containers and distributor, the raw experimental data seem to be below the model data,
and seem to be above without considering the absorption. However, when the model calculates the absorption rate, considering
the initial concentration of water, the data results are concordant
with raw and trimmed experimental data. As a result of the models assumptions and the errors in the experiments, measurements
and calculations, the disparity between the model and the experiments is unavoidable. However, the model is deemed to be used
for approximately design the absorber vessel since the relative
errors are within 20%.

247

L. Yang et al. / International Journal of Heat and Mass Transfer 79 (2014) 241250

0.6

0.8

0.5
0.6

0.4
0.3

0.4
0.8

0.9

1.0

1.1

1.2

-1

absorption rate
heat of absorption

1.4
1.2

1.6
1.0
1.2

0.8
0.6

1.3

0.8

0.9

1.0

(a)

0.20

0.3

0.15

1.1

1.2

0.8

1.3

-1

ASb= -1.6532

RSc= -1.4351
-1

ASa
ASb
ASc

0.2

1.3

RSa
RSc
RSb

RS

0.4

0.25

ASa= -0.7501

AS

0.5

0.30

1.0

1.2

ASc= -0.3533

-1

absorption rate
heat of absorption

0.9

1.1

Heat of absorption (kJ s )

-1

Absorption rate (g s )

0.6

0.8

(b)

0.40
0.35

2.0

-1

Absorption rate (g s )

1.0

-1

0.7

Heat of absorption (kJ s )

1.6

absorption rate
heat of absorption

Heat of absorption (kJ s )

-1

Absorption rate (g s )

0.8

(c)

RSa= -1.6673
RSb= -1.784

(d)

Fig. 5. Inuence of ow resistance of nanouids on the absorption rate in different working conditions. (a): In reference condition; (b): in 0.2 MPa absorption pressure; (c): in
15% initial concentration of falling lm; (d): comparisons of AS and RS.

0.8

12

Model data
Raw experimental data
Regulated experimental data by Eq(17)

11
-1

Absorption rate (gs )

13

0.7

10
9
8
7
3

14

0.6
0.5
0.4
0.3

0.2
5
1

15

1 NH3 vessel 2 decompression valve 3 constant pressure controller; 4,11 container

of solution 5 inlet of cooling water 6, 10 constant flow controller; 7 falling film tube
8 visible absorbor body 9 solution distributor 12tubes for balancing pressure; 13 outlet
of cooling water 14 HP data acquisition instrument 15 computer
Fig. 6. Schematic diagram of the experimental system for NH3/H2O nanouid
falling lm absorption.

Figs. 810 show the comparison between the model data and
trimmed experimental data for the absorption rate at different initial ammonia concentration when the Al2O3, ZnFe2O4 and Fe2O3
nanouids are employed. It can be seen that, for each kind of
nanouid, there are great errors between the model and the raw

10

15

Initial ammonia concentration of water

Fig. 7. The comparisons between the model data, raw experimental data, and
regulated experimental data about absorption rate in different initial ammonia
concentration of water.

experimental data and the majority of them exceed 20%. However,

the errors between the model and the trimmed experimental data
obtained by Eq. (17) are much smaller, and most of them are
within 20%. Hence, it can be concluded that experimental data
trimming is important for obtaining the real absorption rate of
ammonia falling lm absorption outside vertical tube. The model
accuracy of absorption rate for each kind of uid is illustrated in
Fig. 11, which shows the relative errors for each kind of nanouid
at different initial ammonia concentration. When is considered the

248

L. Yang et al. / International Journal of Heat and Mass Transfer 79 (2014) 241250

0.9

1.2

-1

Absorption rate (gs )

-1

Absorption rate (gs )

0.8
0.7
0.6
0.5
0.4
Model data
Raw experimental data
Regulated experimental data by Eq(17)

0.3
0.2

10

1.0

0.8

0.6

0.4

15

Model data
Raw experimental data
Regulated experimental data by Eq(17)

Relative errors of absorption rate

0.8

0.6

0.4

10

15

10

15

Fig. 10. The comparisons between the model data, raw experimental data, and
regulated experimental data about absorption rate in different initial ammonia
concentration of Fe2O3 nanouids.

0.4

-1

Absorption rate (gs )

1.0

Initial ammonia concentration of Fe2O3 nanofluids

Initial ammonia concentration of Al2O3 nanofluids

Fig. 8. The comparisons between the model data, raw experimental data, and
regulated experimental data about absorption rate in different initial ammonia
concentration of Al2O3 nanouids.

Model data
Raw experimental data
Regulated experimental data by Eq(17)

Raw experimental data

0% initial ammonia content
5% initial ammonia content
10% initial ammonia content
15% initial ammonia content

Regulated experimental data by Eq(17)

0% initial ammonia content
5% initial ammonia content
10% initial ammonia content
15% initial ammonia content

0.2

0.0

-0.2

-0.4

Initial ammonia concentration of ZnFe2O4 nanofluids

Fig. 9. The comparisons between the model data, raw experimental data, and
regulated experimental data about absorption rate in different initial ammonia
concentration of ZnFe2O4 nanouids.

Water

Al2O3

ZnFe2O4

Fe2O 3

Fluid types
Fig. 11. The relative errors for each kind of nanoudis in different initial ammonia
concentration.

absorption at the interface of solution in containers and distributor, 87.5% of the relative errors are within 20% and the maximum
is 21.5%. However, if the absorption at the interface of solution in
containers and distributor are not taken into account, the relative
errors between model and raw experimental data are much greater
because 62.5% of the relative errors exceed 20% and the maximum
reaches 36.9%.
In our previous experimental study, to observe the conditions of
solution lm distribution in real-time through the transparent
shell of the absorber all through the experiment, the transparent
shell is made up by acrylic glass. To ensure the strength and security of the experimental unit, the experimental was not carried out
for higher ammonia concentration of uids which matched with
higher absorption pressure. It can be found by experiment or
model that the ammonia absorption rate decreases with the
increase of the ammonia concentration in initial uid. Therefore,
to obviously comparing the ammonia absorption performance
between nanouids and pure ammonia water, the experiments
were carried out with lower ammonia concentration uid. In this
paper, the results of higher ammonia concentration uid can also

be calculated. However, as a result of the value of model calculation cannot yet be veried by our experimental results, the results
of higher ammonia concentration uids are not listed in this paper.
The improvement of absorption rate by nanouids is mainly as
a result of drag reduction and the physical properties of nanouids.
In previous studies of some researchers [5,6,20,26,31,32], the
increase of thermal conductivity, the decrease of viscosity and drag
reduction are found in some kinds of nanouids. These variations
in the nanouids properties may be caused by the rod-shaped
micelles formed by surfactants [3337] and the smoothing effect
on the solid surface caused by the nanoparticles [26]. The experimental results showed that when the ammonia mass fraction of
initial nanouid increases, the absorption potential capacity
decreases, but the enhancing effect induced by the nanouid is
greater than the effect without nanoparticles [20].
Broadly speaking, the model presented herein includes some
physical factors that impact on the absorption performance of
ammonia falling lm absorption outside vertical tube when

249

L. Yang et al. / International Journal of Heat and Mass Transfer 79 (2014) 241250

2.0

Relative errors of heat of absorption

-1

Model data
Heat absorbed by falling film (experimental)
Heat absorbed by cooling water (experimental)

1.5

1.0

0.5

0.4

0.2

0.0
-21.22%

-0.2

-0.4

0.0

Water

Al2O3

ZnFe2O4

Fe2O3

Water

Fluid types

Al2O3

ZnFe2O4

Fe2O3

Fluid types

(a)

(b)

Fig. 12. The comparisons of the absorption heat (a) and relative errors (b) between the model and experimental data in 0% initial ammonia concentration nanouids.

nanouids are employed. There are other factors, such as stability

[2830,38], surface tension, specic heat, etc. that were not considered in this model. The mass transfer coefcient was also not considered in calculating the gross absorption. Consequently, this
model can still be improved by considering those parameters. It
is expected that this model can establish the foundation for further
numerical research on ammonia falling lm absorption outside
vertical tube with nanouids.
Fig. 12 shows the comparisons of the absorption heat (a) and
the relative errors (b) between the model and experimental data
at 0% initial ammonia concentration nanouids. It can be seen that
the heat of absorption is used for heating the nanouids and the
cooling water. The experimental data of absorption heat are not
trimmed like the data of absorption rate. The reason for this is that
the phenomenon of absorption at the interface of solution in containers and distributor, besides the heat released to the metal tube
and the environment, are factors that can cancel each other out. It
can be found from Fig. 12(b) that the model has a high accuracy
when calculating the absorption heat of the pure waterAl2O3 mixture, since the maximum error is within 5%. But it shows a relative
low accuracy when calculating the absorption heat of pure water
ZnFe2O3 and waterFe2O3 nanouids mixtures, because the maximum error is about 20%. Altogether, the model is considered to
be able to predict the absorption heat of ammonia falling lm
absorption outside vertical tube with nanouids and it is expected
that this model can be further improved by modifying it or enhancing the accuracy of experiments and measurements.
The theoretical model of this paper is mainly focused on the inuence of physical properties and ow state of nanouids on the
ammonia falling lm absorption performances. It is not specic to
a certain kind of nanoparticles. And the physical properties and ow
state of nanouids need to be obtained by experimental or mathematical method when the model is applied to a specic kind of nanouid. However, the nanouids for ammonia falling lm absorption
should be stable and highly mobile, otherwise the sedimentation
and adsorption of nanoparticles on the tube will affect the ammonia
absorption performance which is not considered in the model.
5. Conclusions
(1) A numerical model of ammonia falling lm absorption outside vertical tube with nanouid was established by considering the variation of falling lm thickness along falling, and
the changes in physical properties of ammonia water when
adding nanoparticles and dispersants.

(2) The numerical results show that when absorption pressure

decreases or when initial concentration of mixture increases,
the relative intensity of effect on absorption rate is weakened by the variation of thermal conductivity but enhanced
by the variation of mass transfer coefcients and ow resistance, while the variation of mixtures viscosity exhibits very
low effect.
(3) The numerical results are compared to similar experimental
data for the ammonia falling lm absorption with nanouids. The comparison results show that the heat and mass
transfer are mainly affected by the lm drag reduction and
its physical properties. The numerical model obtained can
be used for calculating the absorption rate of ammoniawater-nanouid mixtures with acceptable accuracy, since
87% of relative errors are lower than 20%.
Conict of interest
None declared.
Acknowledgments
The work of this paper is nancially supported by the Science
Foundation of China (51176029), the 12th Five Year National
Science and Technology support Key Project of China (Nos.
2011BAJ03B05 and 2011BAE14B06) and the Scientic Research
Foundation of Graduate School of Southeast University
(YBPY1205). The supports are gratefully acknowledged.
References
[1] Y.T. Kang, J. Kim, J.-K. Kim, C.K. Choi, Comparisons of mechanical, chemical and
nano technologies for absorption applications, in: Proceedings of International
Seminar on Thermally Powered Sorption Technology (Japan), 2003, pp. 6977.
[2] G.D. Xia, H.M. Jiang, R. Liu, Y.L. Zhai, Effects of surfactant on the stability and
thermal conductivity of Al2O3/de-ionized water nanouids, Int. J. Therm. Sci.
84 (2014) 118124.
[3] A. Ghadimi, I.H. Metselaar, The inuence of surfactant and ultrasonic
processing on improvement of stability, thermal conductivity and viscosity
of titania nanouid, Exp. Therm. Fluid Sci. 51 (2013) 19.
[4] B.Q. Xiao, Y. Yang, L.X. Chen, Developing a novel form of thermal conductivity
of nanouids with Brownian motion effect by means of fractal geometry,
Powder Technol. 239 (2013) 409414.
[5] L. Yang, K. Du, X.S. Zhang, Inuence factors on thermal conductivity of
ammonia-water nanouids, J. Cent. South Univ. 19 (2012) 16221628.
[6] L. Yang, K. Du, A thermal conductivity model for low concentrated nanouids
containing surfactants under various dispersion types, Int. J. Refrig. 35 (2012)
19781988.

250

L. Yang et al. / International Journal of Heat and Mass Transfer 79 (2014) 241250

[7] A.A.R. Darzi, M. Farhadi, K. Sedighi, Heat transfer and ow characteristics of

Al2O3water nanouid in a double tube heat exchanger, Int. Commun. Heat
Mass Transfer 47 (2013) 105112.
[8] B.Q. Xiao, B.M. Yu, Z.C. Wang, L.X. Chen, A fractal model for heat transfer of
nanouids by convection in a pool, Phys. Lett. A 373 (2009) 41784181.
[9] A. Ghadimi, R. Saidur, H.S.C. Metselaar, A review of nanouid stability
properties and characterization in stationary conditions, Int. J. Heat Mass
Transfer 54 (2011) 40514068.
[10] A.A.R. Darzi, M. Farhadi, K. Sedighi, S. Aallahyari, M.A. Delavar, Turbulent heat
transfer of Al2O3water nanouid inside helically corrugated tubes: numerical
study, Int. Commun. Heat Mass Transfer 41 (2013) 6875.
[11] B.Q. Xiao, B.M. Yu, A fractal model for critical heat ux in pool boiling, Int. J.
Therm. Sci. 46 (2007) 426433.
[12] G.L. Ding, H. Peng, W.T. Jiang, Y.F. Gao, The migration characteristics of
nanoparticles in the pool boiling process of nanorefrigerant and
nanorefrigerantoil mixture, Int. J. Refrig. 32 (2009) 114123.
[13] H. Peng, G.L. Ding, H.T. Hu, W.T. Jiang, Effect of nanoparticle size on nucleate
pool boiling heat transfer of refrigerant/oil mixture with nanoparticles, Int. J.
Heat Mass Transfer 54 (2011) 18391850.
[14] B.Q. Xiao, B.M. Yu, A fractal analysis of subcooled ow boiling heat transfer, Int.
J. Multiphase Flow 33 (2007) 11261139.
[15] H.T. Hu, H. Peng, G.L. Ding, Nucleate pool boiling heat transfer characteristics
of refrigerant/nanolubricant mixture with surfactant, Int. J. Refrig. 36 (2013)
10451055.
[16] H. Peng, G.L. Ding, H.T. Hu, Effect of surfactant additives on nucleate pool
boiling heat transfer of refrigerant-based nanouid, Exp. Therm. Fluid Sci. 35
(2011) 960970.
[17] J.-K. Kim, J.Y. Jung, Y.T. Kang, The effect of nano-particles on the bubble
absorption performance in a binary nanouid, Int. J. Refrig. 29 (2006) 2229.
[18] F.M. Su, Y.B. Deng, H.B. Ma, Numerical analysis of ammonia bubble absorption
in a binary nanouid, Chem. Eng. Commun. (in press), http://dx.doi.org/
10.1080/00986445.2013.850578.
[19] C.W. Pang, W.D. Wu, W. Sheng, H. Zhang, Y.T. Kang, Mass transfer
enhancement by binary nanouids (NH3/H2O + Ag nanoparticles) for bubble
absorption process, Int. J. Refrig. 35 (2012) 22402247.
[20] L. Yang, K. Du, X.F. Niu, B. Cheng, Y.F. Jiang, Experimental study on
enhancement of ammonia-water falling lm absorption by adding nanoparticles, Int. J. Refrig. 34 (2011) 640647.
[21] A. Szen, E. zbas, T. Menlik, M.T. akr, M. Gr, K. Boran, Improving the
thermal performance of diffusion absorption refrigeration system with
alumina nanouids: an experimental study, Int. J. Refrig. 44 (2014) 7380.
[22] J.K. Kim, J.K. Lee, Y.T. Kang, The effect of oil-droplet on bubble absorption
performance in binary nanoemulsions, Int. J. Refrig. 30 (2011) 17341740.

[23] Y.T. Kang, A. Akisawa, T. Kashiwagi, Analytical investigation of two different

absorption modes: falling lm and bubble types, Int. J. Refrig. 23 (2000) 430
443.
[24] X.F. Niu, K. Du, S.X. Du, Numerical analysis of falling lm absorption with
ammonia-water in magnetic eld, Appl. Therm. Eng. 27 (2007) 20592065.
[25] X.F. Niu, The research of the effect on ammonia-water absorption process of
ammonia-water absorption refrigeration by external magnetic eld (Doctorate
dissertations), School of Energy and Environment, Southeast University, 2008,
pp. 2744 (in Chinese).
[26] L. Liao, The ow and heat transfer characteristics of drag-reducing
nanoparticle suspension (Doctorate dissertations), School of Mechanical
Engineering, Shanghai Jiao Tong University, 2009, pp. 5055 (in Chinese).
[27] I. Eremisinoff, P. Bucgilas, Handbook of Heat and Mass Transfer, V. 2. Mass
Transfer and Reactor Design, Gulf Publishing Company, Texas, United States,
1986, pp. 238244.
[28] L. Yang, K. Du, S.Y. Bao, Y.L. Wu, Investigations of selection of nanouid applied
to the ammonia absorption refrigeration system, Int. J. Refrig. 35 (2012) 2248
2260.
[29] L. Yang, K. Du, X.S. Zhang, B. Cheng, Preparation and stability of Al2O3
nanoparticle suspension of ammonia-water solution, Appl. Therm. Eng. 31
(2011) 36433647.
[30] L. Yang, K. Du, X.F. Niu, Y.J. Li, Y. Zhang, An experimental and theoretical study
of the inuence of surfactant on the preparation and stability of ammoniawater nanouids, Int. J. Refrig. 34 (2011) 17411748.
[31] L. Yang, K. Du, Y.H. Ding, B. Cheng, Y.J. Li, Viscosity-prediction models of
ammonia water nanouids based on various dispersion types, Powder
Technol. 215216 (2012) 210218.
[32] Z.L. Li, J.M. Li, B.X. Wang, H.T. Hu, Inuence of SDBS on viscosity of copper
oxide nano-suspensions, J. Eng. Thermophys. (China) 24 (2003) 849851.
[33] K. Schmitt, P.O. Brunn, F. Durst, Scaling-up correlations for drag reducing
surfactants, Rheol. Acta 26 (1988) 249252.
[34] H. Rehage, H. Hoffman, Rheological properties of viscoelastic surfactant
systems, J. Phys. Chem. 92 (1988) 47124719.
[35] Y. Kawaguchi, H. Daisaka, A. Yabe, Turbulent characteristics in transition
region of dilute surfactant drag reducing ows, in: Proc. 11th Int. Sym,
Turbulent Shear Flows, Grenoble, vol. 9, 1997, pp. 4954.
[36] J.L. Zakin, J. Myska, Z. Chara, New limiting drag reduction and velocity prole
asymptotes for nonpolymeric additives systems, AIChE J. 42 (1996) 3544
3546.
[37] M. Fichman, G. Hetsroni, Electrokinetic aspects of turbulent drag reduction in
surfactant solutions, Phys. Fluids 16 (2004) 43464352.
[38] L. Yang, K. Du, An optimizing method for preparing natural refrigerant:
ammonia-water nanouids, Integr. Ferroelectr. 147 (2013) 2433.