You are on page 1of 7

Energy and Buildings 49 (2012) 209215

Contents lists available at SciVerse ScienceDirect


Energy and Buildings
j our nal homepage: www. el sevi er . com/ l ocat e/ enbui l d
Study on the key ejector structures of the waste heat-driven ejector air
conditioning system with R236fa as working uid
Bo Zhang

, Xutong Song, Jinsheng Lv, Jixue Zuo


School of Energy and Power Engineering, Dalian University of Technology, Dalian 116024, China
a r t i c l e i n f o
Article history:
Received 30 November 2011
Received in revised form
23 December 2011
Accepted 6 February 2012
Keywords:
Ejector
Refrigeration
Computational uid dynamics
R236fa
a b s t r a c t
The ejector performance is numerically and experimentally studied using R236fa as working uid to
analyze the inuence of the key geometries of ejector. The computational uid dynamics (CFD) was
employed to predict the ow process and performance of the ejector. A case study was provided in this
paper, in which the angles of nozzle outlet were 8

, 10

and 14

, the distance of nozzle exit to mixing


chamber was varied as 23 and 33 mm and the length of diffuser was varied as 88 and 103 mm. The CFD
results were veried with an experiment. It can be concluded that, the ejector with the nozzle outlet
angle of 8

, the distance of nozzle exit to mixing chamber of 23mm and the diffuser length of 88 mm has
the best performance.
2012 Elsevier B.V. All rights reserved.
1. Introduction
The ejector refrigeration systemwas rstly developed by Mau-
rice Leblanc [1] in 1910. The main advantage of this systemwas its
fewer moving parts (no compressor) and signicant durable. How-
ever, it usually suffers fromlowperformance. The ejector is the key
component of the ejector refrigeration system.
To improve the ejector performance, many researchers have
devoted into new working uids selection. Huang and Chang [2]
suggested R141b as refrigerant of ejector refrigeration systemand
derived a set of correlations for ejector design through a series
of experiments. Sun [3] studied the performance of an ejector
refrigeration cycle operating with various refrigerants including
R123, R134a, R142b and R500 and so on, among which R500
showed better performance and the entrainment ratio reached
0.23. Khalil et al. made a theoretical study on the ejector using
R134a as refrigerant and the simulated entrainment ratio reached
0.21 [4]. Fong et al. assessed the effect of three refrigerants,
R22, R134a and R410a on the ejector-assisted vapor compres-
sion chiller. It was found that the coefcient of performance was
increased for three refrigerants, in which the degree of enhance-
ment from R134a was the most signicant [5]. Yapici carried out
an experimental study on the performance of vapor ejector refrig-
eration system using R123 as refrigerant and a performance of

Corresponding author at: 2# Linggong Road, Ganjingzi District, Post Code:


116024, Dalian, Liaoning, China. Tel.: +86 411 84706537; fax: +86 411 84691725.
E-mail address: zhangbo@dlut.edu.cn (B. Zhang).
coefcient of 0.39 was obtained [6]. However, no satised refrig-
erant has been achieved according to the consideration on per-
formance, environment protection and economy feasibility for the
ejector refrigeration system, especially for big-capacity industrial
applications.
The structure of the ejector has a very important inuence on
the performance of the ejector. There have been some CFD stud-
ies on supersonic ejectors since the 1990s [7,8]. Riffat and Omer
simulated the ow process inside the ejector of the refrigeration
system, using methanol as refrigerant [9]. It was found that the
CFDresults agreed well with the experimental results and could be
used to predict other various conditions. Pianthong et al. simulated
the ow structure inside ejector and showed the CFD results and
experiment results were within 5% error [10].
In this paper, a new working uid, R236fa, is suggested for
the ejector refrigeration to improve the ejector performance and
system feasibility. The main advantages of R236fa are its mod-
erate pressures, relatively big specic volume that makes the
heat exchangers easy to build and the safety considerations,
such as inammability and nontoxic. Experiments and CFD stud-
ies on the ejector entrainment ratio are provided to show the
feasibility and possibility of R236fa as the refrigerant to the
high capacity ejector refrigeration system. The effects of the
ejector geometry parameters: nozzle outlet angle, the distance
of nozzle exit to mixing chamber and the diffuser length on
the ow characteristic and entrainment ratio are numerically
investigated based on the CFD simulation. The results are also ver-
ied by available experimental data [11] and a new experiment
set.
0378-7788/$ see front matter 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.enbuild.2012.02.009
210 B. Zhang et al. / Energy and Buildings 49 (2012) 209215
Nomenclature
d1 the distance of nozzle exit to mixing chamber (mm)
d2 the diffuser length (mm)
E total energy (J)
G turbulent generation term
k turbulent kinetic energy (J)
m
1
the mass owrate of primary ow
m
2
the mass owrate of secondary ow
P pressure (MPa)
R
m
entrainment ratio
S modulus of the mean rate-of-strain stress tensor
(1/s)
s the mixing entropy of primary ow and secondary
ow
s
1
the entropy of primary ow
s
2
the entropy of secondary ow
u velocity (m/s)
x axial distance (m)
Y
m
overall dissipation rate
stress tensor (N/m
2
)

eff
effective thermal conductivity (w/mk)
the nozzle outlet angle

ij
Kronecker delta
density (kg/m
3
)
turbulence dissipation rate (m
2
/s
3
)
dynamic viscosity (kg/ms)
kinematic viscosity (m
2
/s)
i, j, k space components
2. Ejector refrigeration system
2.1. Systemdescriptions
Fig. 1 shows the schematic diagram of the experimental
rigs. Compared to the traditional electricity driven compression
refrigeration cycle, it can be seen that the ejector replaces the
compressor. The refrigerant in the generator is heated to high tem-
peratureandpressurerefrigerant vapor, whichis calledtheprimary
ow. It undergoes expansion through the nozzle throat to super-
sonic ow. The expansion process causes a low pressure area and
Fig. 2. TS chart of an ejector cooling system.
the vapor of evaporator, called secondary ow, is extracted into the
ejector. The primary owand the secondary oware mixed in the
ejector and owinto the condenser. To complete the refrigeration
cycle, part of the liquid refrigerant is pumped back to the generator,
while the other is fedto the evaporator throughanexpansionvalve.
The performance of the ejector refrigeration cycle depends on the
efciency of the ejector by a huge margin. Such an experimental rig
was built to study ejector entrainment ratio and the performance
of the system.
2.2. Thermodynamics analysis
Fig. 2 is the TS chart of an ejector cooling system. The line 51
shows the evaporation of the refrigerant in the generator. The line
1a reects the process that the primary stream expands to the
evaporating pressure throughthe nozzle of the ejector. The line 62
represents the process that the refrigerant evaporates to saturated
vapor inthe evaporator. The dot b stands for the state of the mixture
of primary and secondary streams at the exit of the nozzle. The line
b3 shows the process that the mixing stream realize a pressure-
rise in the diffuser. The line 34 represents the process that the
mixing streamis cooled and condensed in the condenser. The line
45 shows the process that the refrigerant-liquid is returned to
the generator and receives a boost in pressure. The line 46 shows
the process that some condensate is returned to the evaporator via
Fig. 1. The schematic diagramof the test facility.
B. Zhang et al. / Energy and Buildings 49 (2012) 209215 211
Fig. 3. Conguration of the ejector.
Table 1
The structural parameters of ejector.
The structure of ejector Parameters
Nozzle angle (

) 8

10

14

The distance of nozzle exit to mixing


chamber inlet d1 (mm)
23 33
Diffuser length d2 (mm) 88 103
an expansion valve and completes the throttling depressurization
process.
The entrainment ratio (R
m
), is an important parameter to indi-
cate the performance of the ejector, which is usually dened as,
R
m
=
mass owrate of secondaryow
mass owrate of primaryow
(1)
Fig. 2 shows a comparison of properties between R236fa and
R141b. The entrainment ratio can be calculated formthe following
equation according to Fig. 2.
R
m
=
m
2
m
1
=
s s
1
s
2
s
(2)
The entrainment ratio of R141b can be calculated by the dis-
tance ratio between

and

. The entrainment ratio of


R236fa is the distance ratio between

b a

and

2 b

. The dis-
tance of

and

b a

are nearly equal, however the distance


of

is greater than

2 b

obviously. It canbe concludedthat


the performance of ejector with R236fa as working uid is better
than that of the R141b.
3. Experimental rigs
Fig. 3 shows the conguration of the ejector to be studied in
the paper, including ve parts: the nozzle, the mixing chamber,
the diffuser, the assembly sleeve and the stealing ring. Based on
the calculation method of Sokolov [12] and integrating uid prop-
erties, the structures of the ejector with R236fa as refrigerant are
designed. The key structure parameters and physical dimensions of
ejector were listed in Tables 1 and 2, on the basis of which the CFD
and experimental studies were carried out. In order to study the
inuence of ejector structures, different key structures parameters
were used, such as the nozzle outlet angle varied as 8

, 10

and 14

,
the distance of nozzle exit to mixing chamber as 23 and 33mmand
the length of diffuser as 88 and 103mm.
A 15kW(9kW+6kW) electric heaters is placed at the bottom
of the water tank as the heat source. Pressures are measured by
Table 2
Physical ejector dimensions.
The structure of ejector Size
Nozzle inlet diameter (mm) 18.00
Nozzle throat diameter (mm) 5.28
Nozzle exit diameter (mm) 8.77
Mixing section diameter (mm) 14.68
Diffuser exit diameter (mm) 34.00
The total length (mm) 192.60
the absolute pressure sensors, whose error are within 0.3%. The
precision of temperature sensors is (0.15+0.002 |t|)

C. The mass
ow rates of water are measured using turbine owmeters. The
precision of the turbine owmeters is 0.5% |m|. The precision of
the refrigerant owmeters, its type of DMF-1, is 0.2%.
4. CFD simulation
To simulate the ejector performance, the ejector was modeled
as a two-dimensional axisymmetric body according to Table 2. The
CFDcode was GAMBIT 2.3andFLUENT 6.3. The gridhada quadrilat-
eral structure. The grid initially comprised about 11,747 elements
and was later adapted to about 20,000 elements to conrm that
the results were grid independent. The solving method was cou-
ple implicit. The realizable k turbulence model was selected
[8,13,14]. Boundary conditions were set to pressure inlet and
outlet. The simulation conditions were designed as follows: gener-
ator temperature 85

C, evaporator temperature 10

C, condenser
temperature 2937

C and condenser pressure 0.380.47MPa. The


simulation results of velocity, pressure and temperature are shown
in Figs. 46.
The governing equations can be written in the compact Carte-
sian formas

t
+

x
i
(u
i
) = 0 (3)

t
(u
i
) +

x
j
(u
i
u
j
) =
P
x
i
+

ij
x
j
(4)

t
(E) +

x
i
[u
i
(E +P)] =

.
_

eff
T
x
i
_
+

.[u
j
(
ij
)] (5)
where

ij
=
eff
_
u
i
x
j
+
u
j
x
i
_

2
3

eff
u
k
x
k

ij
(6)
The realizable k turbulence model is derived from the exact
NavierStokes equations, using a mathematical technique called
Renormalization Group.

t
(k) +

x
i
(ku
i
) =

x
j
_
_
+

t

k
_
k
x
j
_
+G
k
Y
m
(7)

t
() +

x
i
(u
i
) =

x
j
_
_
+

t

_

x
j
_
+C
1
S

C
2

2
k +

+C
1

k
C
3
G
b
(8)
where
C
1
= max
_
0.43,

+5
_
=
sk

In above equations, C
2
and C
1
are constants.
k
and

are the
turbulent Prandtl numbers for k and , respectively. The model
constants are C
1
= 1.44, C
2
= 1.9,
k
= 1.0,

= 1.2. The eddy vis-


cosity is computed fromthe belowequation.

t
= C

k
2

(9)
5. Results and discussion
Some important structure, such as the ratio between the diam-
eter of mixing chamber and that of the nozzle, was often discussed.
However, other parameters, such as the nozzle exit angle, the dis-
tance of nozzle exit to mixing chamber and the diffuser length,
212 B. Zhang et al. / Energy and Buildings 49 (2012) 209215
Fig. 4. Velocity chart of ejector.
Fig. 5. Pressure chart of ejector.
Fig. 6. Temperature chart of ejector.
B. Zhang et al. / Energy and Buildings 49 (2012) 209215 213
0.00
0.05
0.10
0.15
0.20
0.25
0.38 0.39 0.4 0.41 0.42 0.43 0.44 0.45 0.46 0.47
E
n
t
r
a
i
n
m
e
n
t

r
a
t
i
o
Back pressure(MPa)
Entrainment ratio
Simulation
Entrainment ratio
Experiment
Fig. 7. Experiment and CFD simulation of R236fa.
which also play important roles in determining the ejector per-
formance, were missed. They are numerically and experimentally
studied in this chapter, when the new working uid, R236fa, is
employed.
5.1. The comparison between experiment and CFD simulation
Fig. 7 shows the relationship between entrainment ratio and
condenser pressure with refrigerant R236fa. From Fig. 7 it can be
seen that the average value of the entrainment ratio of the CFD
simulationis about 0.24. However, the average value of the entrain-
ment ratio of the experiment is about 0.20. The relative error is
between 10% and 18%. The CFD results can be a very good experi-
mental verication.
5.2. The relationship between entrainment ratio and nozzle angle
Fig. 8 shows three curves describing the relationship between
the entrainment ratio and nozzle angle. The three curves repre-
sent numerical results at different nozzle angles with refrigerant
R236fa. This situation remains the same by adjusting the distance
fromnozzle outlet to the inlet of mixing chamber. Figs. 911 show
the velocity vector chart of ejector at different nozzle angles in
order to investigate the reason the ejector structure affecting the
performance. We can see that vertex of entrainment region is rel-
atively smaller when the nozzle angle is 8

but a relatively larger


vertex exists in the suction chamber with the nozzle angle of 10

and 14

. It should be mentioned that for different refrigerants, the


nozzle angles may have different values considering the ow pat-
terns varies, which needs further analysis on the owprocess from
outlet of the nozzle to the inlet of the mixing chamber, as described
above.
0.00
0.05
0.10
0.15
0.20
0.25
0.38 0.39 0.40 0.41 0.42 0.43 0.44 0.45 0.46 0.47
E
n
t
r
a
i
n
m
e
n
t

r
a
t
i
o
Back pressure(MPa)
=10
=8
=14
Fig. 8. The angle of the primary ownozzle and the entrainment ratio.
Fig. 9. Velocity vector chart of ejector (=8

, d1=23mm, d2=88mm).
Fig. 10. Velocity vector chart of ejector (=10

, d1=23mm, d2=88mm).
Fig. 11. Velocity vector chart of ejector (=14

, d1=23mm, d2=88mm).
214 B. Zhang et al. / Energy and Buildings 49 (2012) 209215
0.00
0.05
0.10
0.15
0.20
0.25
0.38 0.39 0.40 0.41 0.42 0.43 0.44 0.45 0.46 0.47
E
n
t
r
a
i
n
m
e
n
t

r
a
t
i
o
Back pressure(MPa)
d1 =23mm
d1 =33mm
Fig. 12. The distance of nozzle outlet to mixing chamber and the entrainment ratio.
With the increase of the back pressure, the entrainment ratio
remains unchanged in a certain range. But the entrainment ratio
will decrease when the back pressure reaches a certain value. The
entrainment ratio can reach 0.24 at nozzle angle 8

. As a compar-
ison, under the same condition, the entrainment ratio can reach
0.19 at nozzle angle 14

and the entrainment ratio is 0.21 at nozzle


angle 10

. However, whenthe nozzle angle continues todecrease to


less than 8

, the ejector entrainment ratio begins to drop. Thus, the


nozzle angle of 8

is suggested when R236fa is used as the working


uid.
5.3. The relationship between entrainment ratio and the distance
of nozzle exit to mixing chamber
Fig. 12 shows the relationship between the entrainment ratio
and the distance of nozzle exit to mixing chamber. The distance of
nozzle exit to mixing chamber is adjusted within a certain range,
the performance of the ejector changes. By comparison of the
Figs. 9 and 13, we can see the vertex of entrainment region is rela-
tively smaller when the distance of nozzle exit to mixing chamber
is 23mm and relatively larger when the distance is 33mm. From
these results, it is found that the ratio of d1 to mixing chamber
diameter reaches around 1.50 for the best ejector performance,
which could be veried when other refrigerants were used [15].
Thus, it can be concluded that the designing ratio of the distance
of nozzle exit to the mixing chamber diameter coincide with other
working uids.
Fig. 13. Velocity vector chart of ejector (=8

, d1=33mm, d2=88mm).
0.00
0.05
0.10
0.15
0.20
0.25
0.38 0.39 0.40 0.41 0.42 0.43 0.44 0.45 0.46 0.47
E
n
t
r
a
i
n
m
e
n
t

r
a
t
i
o
Back pressure(MPa)
d2=88mm
d2=103mm
Fig. 14. The length of diffusion chamber and entrainment ratio.
5.4. The relationship between entrainment ratio and the length of
diffuser
Fig. 14 illustrates the effect of the length of diffuser on the
entrainment ratio. The entrainment ratio changes with different
back pressures. It can be seen that, the entrainment ratio can reach
0.24 when the length of diffuser is 88mm. With the increase of the
back pressure, the condenser duty increases, too. For a xed back
pressure, the entrainment ratio will decrease with the increase of
the length of the diffuser.
The relationship between diffuser length and mixing chamber
diameter was veried by Hewedy et al. [16] for the best ejector
performance. Fromthe results, it is found that the ratio of the dif-
fuser length to mixing chamber diameter is around 6.0 for the best
ejector performance with R236fa as working uid.
6. Conclusions
This paper employs CFDand experiment methods to investigate
the effect of the ejector geometry on the performance of ejector
withR236fa as refrigerant. The followingconclusions canbe drawn.
To improve the ejector performance, an appropriate nozzle
angle should be chosen. The nozzle angle of 8

is the best for the


ejector with R236fa as working uid. The ejector performance will
become worse when the nozzle angle is greater or less than 8

.
The distance of nozzle exit to mixing chamber has also an
important inuence on the performance of the ejector. A better
performance of the ejector with different refrigerants can be real-
ized when the ratio of d1 to mixing chamber diameter is around
1.50.
The diffuser lengthis important for the designof the ejector. The
ejector with R236fa as working uid can play better when the ratio
of diffuser length to mixing chamber diameter reaches around 6.
The CFDresults reveal that the structure of ejector has animpor-
tant inuence on the performance of the ejector. To get a better
performance, theejector shouldbedesignedproperly. It canbecon-
cluded fromthe paper with R236fa as working uids that a better
performance of the ejector can be obtained under the conditions:
the nozzle outlet angle 8

, the distance of nozzle exit to mixing


chamber 23mm and the diffuser length 88mm.
Acknowledgement
This work was supported by projects of the National Nature
Science Foundation of China (no. 50706005).
B. Zhang et al. / Energy and Buildings 49 (2012) 209215 215
References
[1] W.B. Gosney, Principleof refrigeration, CambridgeUniversityPress, Cambridge,
1982.
[2] B.J. Huang, J.M. Chang, Empirical correlation for ejector design, International
Journal of Refrigeration 22 (1999) 379388.
[3] Da-Wen Sun, Comparative study of the performance of an ejector refrigeration
cycle operating with various refrigerants, Energy Conversion Management 40
(8) (1999) 873884.
[4] A. Khalil, M. Fatouh, E. Elgendy, Ejector design and theoretical study of R134a
ejector refrigeration cycle, International Journal of Refrigeration 34 (2011)
16841698.
[5] K.F. Fong, C.K. Lee, T.T. Chow, Improvement of solar-electric compression
refrigeration system through ejector-assisted vapour compression chiller for
space conditioning in subtropical climate, Energy and Buildings 43 (2011)
33833390.
[6] R. Yapici, Experimental investigation of performance of vapor ejector refriger-
ation systemusing refrigerant R123, Energy Conversation and Management 49
(2008) 953961.
[7] S.D. Kim, I.S. Jeong, D.J. Song, A computational analysis of unstead tran-
sonic/supersonic ows over backward facing step in air jet nozzle, Journal of
Mechanical Science and Technology 1 (2) (2007) 336347.
[8] E. Rusly, L. Aye, W.W.S. Charters, A. Ooi, CFD analysis of ejector in a com-
bined ejector cooling system, International Journal of Refrigeration8 (7) (2005)
10921101.
[9] S. Riffat, S. Omer, CFD modeling and experimental investigation of an ejector
refrigeration systemusing methanol as the working uid, International Journal
of Energy Research 5 (2) (2001) 115128.
[10] K. Pianthong, W. Seehanam, M. Behnia, T. Sriveerakul, S. Aphornratana, Inves-
tigation and improvement of ejector refrigeration systemusing computational
uid dynamics technique, Energy Conversion and Management 48 (2007)
25562564.
[11] MyoungKukJi, TonyUtomo, YongHunLee, HyoMin Jeong, Han ShikChung, CFD
investigation on the owstructure inside thermo vapor compressor, Energy 35
(2010) 26942702.
[12] Huang Qiuyun Trans. Sokolov, Jingel, Ejector, second ed., Science Press, Beijing,
1977, pp. 4144.
[13] Y. Bartosiewicz, Z. Aidoun, Y. Mercadier, Numerical assessment of ejector oper-
ation for refrigeration applications based on CFD, Thermal Engineering 26
(2006) 604612.
[14] T. Sriveerakul, S. Aphornratana, K. Chunnanond, Performance prediction
of steam ejector using computational uid dynamics. Part 1. Validation
of the CFD results, International Journal of Thermal Science 46 (2007)
812822.
[15] B.J. Huang, J.M. Chang, C.P. Wang, V.A. Petrenko, A 1-D analysis of
ejector performance, International Journal of Refrigeration 22 (1999)
354364.
[16] N.I.I. Hewedy, Mofreh H. Hamed, F. Abou-Taleb, Sh. Tarek, A. Ghonim, Optimal
performance and geometry of supersonic ejector, Journal of Fluids Engineering
130 (2008) 110.

You might also like