You are on page 1of 14

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/245053257

Use of ejectors in a multi-evaporator refrigeration system for performance


enhancement

Article in International Journal of Refrigeration · September 2009


DOI: 10.1016/j.ijrefrig.2009.03.009

CITATIONS READS

69 1,443

4 authors, including:

Lakdar Kairouani Mouna Elakhdar


École Nationale d'Ingénieurs de Tunis University of Tunis El Manar
95 PUBLICATIONS 740 CITATIONS 33 PUBLICATIONS 335 CITATIONS

SEE PROFILE SEE PROFILE

Nahla Bouaziz
University of Tunis El Manar
41 PUBLICATIONS 197 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

heat transfer in micro duct View project

Ejector Refrigeration systems View project

All content following this page was uploaded by Lakdar Kairouani on 07 June 2018.

The user has requested enhancement of the downloaded file.


international journal of refrigeration 32 (2009) 1173–1185

available at www.sciencedirect.com

w w w . i i fi i r . o r g

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

Use of ejectors in a multi-evaporator refrigeration system for


performance enhancement

L. Kairouani*, M. Elakhdar, E. Nehdi, N. Bouaziz


Unité de Recherche Energétique et Environnement, Ecole Nationale d’Ingénieurs de Tunis, Tunis, Tunisia

article info abstract

Article history: In this study, an improved cooling cycle for a conventional multi-evaporators simple
Received 12 July 2008 compression system utilizing ejector for vapour precompression is analyzed. The ejector-
Received in revised form enhanced refrigeration cycle consists of multi-evaporators that operate at different pres-
12 February 2009 sure and temperature levels. A one-dimensional mathematical model of the ejector was
Accepted 18 March 2009 developed using the equations governing the flow and thermodynamics based on the
Published online 28 March 2009 constant-area ejector flow model. The model includes effects of friction at the constant-
area mixing chamber. The energy efficiency and the performance characteristics of the
Keywords: novel cycle are theoretically investigated. The comparison between the novel and
Refrigeration conventional system was made under the same operating conditions. Also, a comparison
Review of the system performances with environment friendly refrigerants (R290, R600a, R717,
Ejector system R134a, R152a, and R141b) is made. The theoretical results show that the COP of the novel
Improvement cycle is better than the conventional system.
Performance ª 2009 Elsevier Ltd and IIR. All rights reserved.
Modelling
Simulation
Ejector
Comparison
Refrigerant

Utilisation d’éjecteurs dans un système frigorifique à


plusieurs évaporateurs afin d’améliorer la performance
Mots clés : Système frigorifique ; Enquête ; Système à éjecteur ; Amélioration ; Performance ; Modélisation ; Simulation ; Éjecteur ;
Comparaison ; Frigorigène

1 Introduction vapour compression cycle, is 100 years old, inefficient, and


environmentally unsound. Since the 1980s, the refrigeration
Refrigeration is one of the leading uses of electric power in the industry has faced pressure to improve efficiency and reduce
world. The technology most often used in refrigeration, the emissions of the chlorofluorocarbon (CFC) compounds used in

* Corresponding author. Tel.: þ216 71874700; fax: þ216 71872729.


E-mail address: lakdar_kairouani@yahoo.fr (L. Kairouani).
0140-7007/$ – see front matter ª 2009 Elsevier Ltd and IIR. All rights reserved.
doi:10.1016/j.ijrefrig.2009.03.009
1174 international journal of refrigeration 32 (2009) 1173–1185

Nomenclature F ejector area ratio


q temperature ratio
a sound speed
U diffuser area ratio
A area of cross-section
D increment
COP coefficient of performance
Cp specific heat capacity at constant pressure, Subscripts and superscripts
J kg1 K1 c condenser
D diameter co conventional
F Friction factor comp compressor
h enthalpy, J kg1 e1 motive nozzle outlet
k ratio of specific heats ( ¼ Cp/Cv) e2 mixing tube inlet
L length e3 diffuser inlet
M dimensionless velocity e4 diffuser outlet
m mass flow rate, kg s1 ev1 evaporator 1
P pressure, Pa ev2 evaporator 2
Q heat load, W ev3 evaporator 3
R gas constant ej1 ejector1
r compression ratio ( ¼ P4/P00 ) ej2 ejector2
s entropy, J kg1 K1 H high
T temperature L low
U mass flow rate ratio (¼m00 /m0 ) m mixed
V velocity, m s1 ne new
W mechanical work (J/kg) r ratio
X distance between the primary nozzle exit and the s isentropic
mixing chamber inlet sat saturation
sup superheating
Greek letters
0 stagnation
h efficiency 0
primary
x driving pressure ratio 00
secondary
G expansion ratio
* nozzle throat

vapour compression cooling. Attempts to decrease CFC mechanical compressors. Since that time, development and
emissions by using alternate compounds have typically made refinement of ejector refrigeration system have been almost at
refrigeration devices less efficient. Owing to the high cost of a standstill as most efforts have been concentrated on
energy and the decrease of energy resources, ejector refrig- improving vapour compression refrigeration systems. The
eration systems have become a current issue and have been compression–ejection refrigeration system (Sokolov and
an interesting subject for researchers. Hearshgal, 1990a,b, 1993; Sokolov, 1991; Sun, 1997; Gokthun,
The ejector, Fig. 1, which is the heart of the ejector refrig- 2000) consists of a conventional compression and ejector sub-
eration system, was invented by Sir Charles Parsons around cycles with heat exchanger as an interface between them.
1901 for removing air from a steam engine’s condenser. In According to its configuration, pressure ratios across the
1910, an ejector was used by Maurice Leblanc in the first steam ejector and the compressor are maintained at low level. The
jet refrigeration system (Gosney, 1982). This system experi- heat load is transferred from an evaporator to a heat
enced a wave of popularity during the early 1930s for air exchanger and then compressed and rejected to the
conditioning large buildings (Stoecker, 1958). Ejector refriger- surrounding at a condenser. In other words, the ejector sub-
ation systems were later supplanted by systems using cycle was used as the heat rejection system. If a single
refrigerant is used, the heat exchanger is replaced by the
Primary nozzle Mixing Chamber Diffuser
mixing chamber and combined both heat and mass transfer
processes. Sokolov and Hershgal (1990c) explained the
Pe4 concepts and design procedures of the system, while the
Primary fluid d m’+m’’
D dD description of the constructed multi-ejector R114 machine
m’,P0’,T0’ Te4
and its experimental data were given in Sokolov and Hershgal
(1990b). In Sokolov (1991), the mathematical simulation (effect
X L Ld of operating condition) of a constructed machine and
Secondary fluid e1 e2 e3 e4 a machine with the utilization of solar energy were made
m’’,P0”,T0’’ respectively. However, R114 was found to be harmful to the
environment and be prohibited. Sun (1997) conducted the
Fig. 1 – Schematic of a constant-area ejector and geometric mathematical simulation of an environmental friendly solar
parameters. system. Steam and R134a were used as the refrigerant in an
international journal of refrigeration 32 (2009) 1173–1185 1175

ejector and a compression sub-cycle, respectively. The simu- novel MECS (NMECS) is proposed. This configuration increases
lated results showed that the COP of the system could be the suction pressure. In fact, in the diffuser, the kinetic energy
improved up to 50% compared to the conventional vapour of the mixture is converted into pressure energy. The specific
compression system. The analysis of maximum possible COP work of the compressor is reduced and then the COP of the
of a solar powered hybrid compression-jet refrigerator, in system is improved as compared to the conventional refrig-
terms of Carnot efficiency, was provided in Gokthun (2000). It eration system and to the indirect refrigeration system. A one-
is obvious that the booster and the vapour compression cycle dimensional mathematical model was developed using the
can provide higher thermal efficiency to the ejector refriger- equations governing the flow and thermodynamics based on
ation system. On the other hand, either installation of an the constant-area ejector flow model.
ejector (Tomasak and Radermacher, 1994) or its refrigeration The model presented in this paper is simulated with
system (Huang et al., 2001) can also improve the thermal FORTRAN. The refrigerants properties are evaluated by using
efficiency of the ordinary vapour compression refrigerator. REFPROP V7.1 (NIST, 2006). This model is applied to some
Multi-evaporator compression systems (MECS) (used in refrigerants known as ‘‘natural’’ (R290, R152a, R717 and R600a),
transport refrigeration applications), Fig. 2, are subjected to with some transitory fluids R141b which is a good working
demanding performance requirements because of the need to fluid for an ejector and R134a (having a high contribution to the
carry a large variety of cargos under wide variations in greenhouse effect but always used in refrigeration).
climatic conditions. Although direct expansion systems are
utilised almost exclusively in multi-refrigeration applications,
there is increasing interest in indirect secondary system 2 System description
designs (Jessica et al., 2005). In an indirect refrigeration
system, the direct expansion cycle cools a single-phase Fig. 3 presents a schematic of the ejector cycle. The ( p–h)
coolant, which constitutes the working fluid of a secondary charts for conventional and ejector cycles are shown in Figs. 4
loop and for transport refrigeration applications should and 5, respectively.
ideally be a non-toxic, non-flammable substance with good The refrigerant enters the compressor at low pressure PL at
thermophysical properties. The secondary loop then cools state (1) and is compressed isentropically to the high-side
each of the compartments directly by means of a cross flow air pressure PH at state (2s). The real compression process to the
heat exchanger. Apart from potential environmental benefits, high-side pressure PH with an isentropic efficiency hcomp ends
indirect systems may also address control and defrost at state (2). The fluid enters the condenser where it condenses
complexities associated with current direct expansion multi- to state (3) by rejecting heat to the surroundings. Then the
temperature systems. Previous work on supermarket indirect condensate is divided into three flows, states (4), (5) and (6).
systems shows that secondary loop refrigeration systems can One enters the evaporator 1 after a pressure reduction in the
deliver the same cooling capacity as a direct expansion expansion device 1, state (7), the other enters the evaporator 2
refrigeration system, but using in some cases up to 15% less after a pressure reduction in the expansion device 2, state (8)
energy (California Energy Commission, 2004; Horton and and the last enters the evaporator 3 after a pressure reduction
Groll, 2003). in the expansion device 3, state (9). The refrigerant coming
This present study takes a new approach to enhance the from expansion device 1 enters the evaporator 1 and draws
COP of the conventional (MECS). For this purpose, an ejector is heat from the surroundings. It is at saturated vapour state
introduced into the conventional MECS (CMECS), and the (10sat). Then it is at superheated state (10). The refrigerant
coming from expansion device 2 enters the evaporator 2 and
C draws heat from the surroundings. It is at saturated vapour
2 state (11sat). Then it is superheated state (11). The fluid
3
entering the evaporator 3 is vaporised from state (9) to (15),
superheated to state (15sat). The flow at state (10) enters the
4 5 6

COMP EJ1 EJ2


14 1 2
ED1 ED2 ED3 10
6 7 11 12 13 17
8 16
15

EV1 EV2 EV3


EV1 EV2 EV3 C

9 10 1 7 8 9
ED4 ED5 ED1 ED2 ED3

11 12 13 4 5 6 3

Fig. 2 – Direct expansion multi-temperature refrigeration Fig. 3 – Schematic of the novel multi-evaporators
system (CMECS). compression system.
1176 international journal of refrigeration 32 (2009) 1173–1185

diffuser where the pressure increases. The diameters and


P
lengths of various parts forming the nozzle, the diffuser
and the suction chamber, together with the fluid flow rate and
properties, define the ejector capacity and performance. The
3,4,5,6 entrainment ratio (U ) is the flow rate of the entrained vapour
2
divided by the flow rate of the motive vapour. As for the
9 expansion ratio (G), it is defined as the ratio of the motive
6 vapour pressure to the entrained vapour pressure. The
compression ratio (r) gives the pressure ratio of the
10 compressed vapour to the entrained vapour. The driving
7
pressure ratio (x) is defined as the ratio of the motive vapour
8 pressure to the back pressure.
13 1 12 11 Le Grives and Fabri (1969) defined three different operating
regimes for ejectors. In these definitions, the ejector operating
regimes are named based on the dependence of the entrain-
h
ment ratio on the back pressure at the exit of the ejector (or
Fig. 4 – Pressure–enthalpy chart of the conventional multi- the driving pressure ratio). These regimes are the supersonic
evaporators single compression cycle. regime (SR), the transition regime (TR) and the mixed regime
(MR). The performance analysis of an ejector consists of
determining the formation conditions of these regimes.
During the operation of the ejector in the SR, since the primary
ejector nozzle and expands to a mixture at state (12). The
static pressure at section e1 shown in Fig. 1, is higher than that
saturated secondary vapour enters the ejector at pressure Pev2
of the secondary vapour, the primary fluid expands against
corresponding to state (11). The two vapours mix in the ejector
the secondary fluid and causes the velocity of the secondary
with the final state of the mixture corresponding to state (13).
fluid to reach supersonic speed at the aerodynamic throat
The mixture then flows through the ejector diffuser where it
formed by it. As a consequence of this secondary stream
recovers to pressure Pej1o at state 14. The vapour leaving the
choking phenomenon, the secondary mass flow rate becomes
ejector 1 at state (14) flows into the primary nozzle of the
independent of the back pressure. The MR includes all the
ejector 2 drawing vapour into the ejector 2 from evaporator 3
cases for which the secondary flow is not choked. The
and expands to state (16). The two vapours mix to state (17)
secondary flow cannot reach sonic speed within the mixing
and leave the ejector 2 after a recovery of pressure in the
chamber, and therefore, its mass flow rate changes depending
diffuser part at state (1).
on the back pressure. In the TR, the secondary vapour reaches
supersonic speed at the point of confluence of the primary and
secondary vapours. It gives the optimum performance of the
3 Ejector analysis
ejector, (Nahdi et al., 1993). For this reason, we will analyze the
ejector in TR. The ejector modeling in transition regime
Fig. 1 represents the schematic of the ejector. The refrigerant
developed by Lu (1986) by using the one-dimensional
vapour at high pressure is supplied to the ejector primary
constant-area ejector model is modified so that it could be
inlet. This vapour expands in a convergent–divergent nozzle,
applied to the NMECS operating with several working
thereby creating a depression and drawing in low pressure
refrigerants.
vapour through the ejector secondary inlet. The two flows
The following assumptions are made for the analysis:
come into contact in the mixing chamber and entering the

1. The flow inside the ejector is steady and one-dimensional.


P 2. The working fluid is an ideal gas with constant properties
Cp and k.
3. For simplicity in deriving the 1-D model, the isentropic
3,4,5,6 2 relations are used before mixing;
4. The two fluids are completely mixed at the exit of the
mixing chamber.
10
7 5. X s 0.
14

11 With X s 0 the flow in convergent part is sonic which


8
1 13 12 implies that the aerodynamic throat is situated in the cylin-
9 drical part of the mixing chamber. So, we deduce that, for TR
15 17 16 with a distance X s 0, the aerodynamic throat is located at the
entry of the mixing chamber. Therefore:

h M00e2 ¼ 1 (1)

Fig. 5 – P–h diagram of the novel multi-evaporators To form the sonic throat of the secondary fluid at the section
compression cycle. e2, the motive flow must expand, which imposes P0e2 > P00e2 .
international journal of refrigeration 32 (2009) 1173–1185 1177

 
After the section e2, we can only have P0e2 > P00e2 , since the case ðk þ 1Þ
mV þ PA ¼ ma f1 ðMÞ (12)
P0e2 < P00e2 , is physically impossible. For the condition: P0e2 > P00e2 , 2k
the primary fluid is going to continue to expand, the sonic
By using the Eqs. (10) and (12) and by dividing the momentum
throat is situated then downstream the section e3 and the
equation by ððk þ 1Þ=2kÞa0 m0 , Eq. (8) can be expressed as:
regime becomes supersonic. Therefore, the TR is character-
ized by: ae3 me3     a00 m00  
0 0
f1 ðMe3 Þ þ xMe3 ¼ f1 M0e2 þ 0 0 f1 M00e2 (13)
a m a m

P0e2 ¼ P00e2 (2) where x is defined as: x ¼ ðk=k þ 1ÞðL=DÞF


The distance L (Fig. 1) in the TR is assumed to be 10 D
By applying the mass, momentum and energy balances to the
(Paliwoda, 1968).
control volume defined between section e2 and section e3
By using the isentropic relations, the momentum equation
(Fig. 1), we can write the following equations.
becomes:
Continuity equation:
      
1 þ Uq1=2 f1 ðMe3 Þ þ xMe3 ¼ f1 M0e2 þ Uq1=2 f1 M00e2 (14)
m0 þ m00 ¼ me3 (3)
Calculation of the stagnation pressure P0e3 in the section e3:
with U ¼ m00 =m0 , the Eq. (3) can be expressed as: The primary mass flow rate can be expressed as:

me3
¼1þU (4)  1=2  k=ðk1Þ  1=2
m0 1 k 2 kþ1
m0 ¼ P00 pffiffiffiffiffi0 A0 (15)
Energy equation: T0 R kþ1 2

The mass flow rate in the section e3:


m0 Cp T0 þ m0 Cp T0 ¼ me3 Cp T0e3 (5)
 1=2  k=ðk1Þ  1=2
where T0e3 is the stagnation temperature at the section e3. 1 k 2 kþ1
me3 ¼ P0e3 pffiffiffiffiffiffiffiffiAe3 (16)
By dividing the two members of the equation by m0 Cp T0 , the T0e3 R kþ1 2
energy equation can be expressed as follows:
where Ae3 is the fictitious throat in the mixing chamber and
    can be expressed as follows by using isentropic relations:
T0e3
1 þ Uq ¼ ð1 þ UÞ (6)
T0
Ae3 ¼ Ae3 f2 ðk; Me3 Þ (17)
where q is defined as:
where:
T00  1=ðk1Þ  1=ðk1Þ
q¼ 0 (7) A kþ1 k1 2
T f2 ðk; MÞ ¼ ¼ M 1 M (18)
A 2 kþ1
Momentum equation:
By using Eq. (15) attributed to the section e2 and Eq. (16), we
me3 Ve3 þ Pe3 Ae3 þ DPAe3 ¼ m0 Ve2
0
þ P0e2 A0e2 þ m00 Ve2
00
þ P00e2 A00e2 (8) obtain:
DPAe3 expresses the frictional losses inside the mixing sffiffiffiffiffiffiffiffi
P0e3 me3 T0e3 A0 1
chamber and is obtained from the following equation: ¼ 0 (19)
P00 me2 T00 Ae3 f2 ðk; Me3 Þ
 2   where
rVm L
DPAe3 ¼ F Ae3 (9)
2 D
T0e3 1 þ Uq
¼
where F is the friction factor. T00 1þU
2 2
With the assumptionrVm ¼ re3 Ve3 , Eq. (9) can be expressed
as: me3
¼1þU
m0e2
Ve3
DPAe3 ¼ FðL=DÞme3 (10) therefore:
2
In order to simplify the equations resolution, the dimen- P0e3 1 þ Uq1=2
sionless velocity M ðM ¼ V=a Þ and the function f1(M ) are 0
¼ (20)
P0 Ff2 ðk; Me3 Þ
introduced, (Abramovich, 1970; Lu, 1986):
where F ¼ Ae3 =A0
Calculation of the static pressure Pe4 in the section e4:
1 By using isentropic relations we can write:
f1 ðMÞ ¼ M þ (11)
M
Pe4 ¼ P0e4 f3 ðk; Me4 Þ (21)
where a is the sound speed at the nozzle throat and it is given
by: where
pffiffiffiffiffiffiffiffiffiffiffi  k=ðk1Þ
a ¼ kRT k1 2
f3 ðk; MÞ ¼ 1 M (22)
kþ1
and after a series of transformations of the expression
mV þ PA, we obtain: We can also write:
1178 international journal of refrigeration 32 (2009) 1173–1185

Input K, L, D, d, dD, dx, F,ηD

Used functions
f2 (k, M) f3 (M) f4 (k, M)

Initial data : fluid, T0', T0''


Evaluation of refrigerant
thermodynamic properties

Determination of thermodynamic
state P0' , P0''
Calculate, θ

Calculate geometric parameters:


φ , ϕ , L / D, x

M2'' = 1
- Calculate :
f1 M2'' , f2 k, M2'' , f3 k, M2''

'
- Initialisation de M M2'
2:

- Calculation of Functions : f1 M2' , f2 k, M2' , f3 k, M2'


ΓA Equation (36)
ΓB Equation (37)

M2' = M2' _ ΔM2' ΓA _ ΓB < ε1

OK
1
Calculate U θ 2
Equation (35)

Compute: M3, M4
- M3: Resolution of equation (32)
- M4: Resolution of equation (33)

Calculate : f2 (k, M4), f3 (k, M4)


- ξ Equation (34)
- P4 Equation (38)
1
(θ ) U
2

Results

Fig. 6 – Simulation flowchart for ejector performance analysis in transition regime.

By using a relation similar to (15) the mass flow rate at the


Pe4 ¼ P0e3 hD f3 ðk; Me4 Þ (23) exit of the diffuser can be expressed as:
where hD is the pressure coefficient in the diffuser and is
 1=2  k=ðk1Þ  1=2
expressed by: 1 k 2 kþ1
me4 ¼ P0e4 pffiffiffiffiffiffiffiffiAe4 f2 ðk; Me4 Þ (25)
T0e4 R kþ1 2
P04
hD ¼ (24)
P03 By combining Eqs. (16), (17) and (25) we obtain:
international journal of refrigeration 32 (2009) 1173–1185 1179

 1=2  
me4 P0e4 T0e3 Ae4 f2 ðk; Me4 Þ 1 1  
¼ (26) Uq1=2 ¼ F  0
 f2 k; M00e2 (37)
me3 P0e3 T0e4 Ae3 f2 ðk; Me3 Þ G f2 k; Me2

The heat loss in the diffuser is neglected, the energy  k=ðk1Þ


2 1
conservation in the diffuser imposes: G¼   (38)
kþ1 f3 k; M0e2
m3 Cp T03 ¼ m4 Cp T04 (27) In which:

P00
Therefore the stagnation temperatures T03 and T04 are G¼ (39)
P000
equal.
By using Eqs. (24), (27) in (25) we obtain:
P00
x¼ (40)
f2 ðk; Me3 Þ Pe4
f2 ðk; Me4 Þ ¼ (28)
UhD
Pe4 G
where U ¼ Ae4 =Ae3 r¼ or r ¼ (41)
P000 x
Substituting Eqs. (20) and (28) in Eq. (23), we can find
a relation between the exit parameters ðPe4 ; Me4 Þ and the inlet In this system of equations, we have nine parameters:
parameters ðP00 ; qÞ: UðqÞ1=2 , x, G, F, U, M0e2 , M00e2 , Me3, Me4 (by supposing that the
pressure coefficient in the diffuser hD and the friction factor F
Uq1=2 þ 1 (therefore x) are fixed). The most important parameters are
f2 ðk; Me4 Þ ¼ f3 ðk; Me4 Þx (29)
FU thermodynamic parameters UðqÞ1=2 , x, G, and geometric
Entrainment ratio U: parameters: F, U. To have the solution of the system, it is
Similar to Eq. (19), the entrainment ratio U can be necessary to fix four initial parameters. Thus, the system of
expressed as: Eqs. (34)–(38) translates the relation between five variables
sffiffiffiffiffi among the set of nine parameters in which four are fixed.
m00 P000 T00 A00e2   In our case, the four fixed variables are: M00e2 , G, F and U. The
U¼ ¼ f2 k; M00e2 (30)
m0 P00 T000 A0 five unknown parameters are: UðqÞ1=2 , x, M0e2 , Me3, Me4.
The aim of the ejector modeling is to find the back
Finally, by using F we obtain:
pressures of the two ejectors and the two entrainment
  ratios.
1 1  
Uq1=2 ¼ F  0
 f2 k; M00e2 (31)
G f2 k; Me2

On the other hand, by using the hypothesis expressed by 4 System performance


P0e2 ¼ P00e2 and similarly to Eq. (23), we have:
  The system performance is evaluated by the coefficient of
P0e2 ¼ P00 f3 k; M0e2
performance COP, which represents the ratio of the cooling
and capacity to the mechanical power added to the cycle. It is
  written as follows:
P00e2 ¼ P000 f3 k; M00e2
Qev1 þ Qev2 þ Qev3
so COP ¼ (42)
Wcomp
 
P00 f3 k; M00e2 The heat transfer rates in the evaporators 1, 2 and 3 paid to
00
¼  0
 (32)
P0 f3 k; Me2 the total mass ðm1 þ m2 þ m3 Þ are given by:
By combining Eqs. (31) and (32)
ðh10  h7 Þ
Qev1 ¼    (43)
1 þ Uej1 1 þ Uej2
 
f2 k; M00e2 UðqÞ1=2
 00
¼    (33)  
f3 k; Me2 1
F  f k;M f3 k; M0e2 Uej1
2ð e2 Þ
0 Qev2 ¼ ðh11  h8 Þ    (44)
1 þ Uej1 1 þ Uej2
System of equations:
The final system of equations used for the transition Uej2
Qev3 ¼ ðh15  h9 Þ  (45)
regime is given by: 1 þ Uej2
      
1 þ Uq1=2 f1 ðMe3 Þ þ xMe3 ¼ f1 M0e2 þ Uq1=2 f1 M00e2 (34) The ejectors entrainment ratio, Uejl and Uej2 are defined as the
ejector suction mass flow rate at secondary inlet divided by
f2 ðk; Me4 Þ ¼ f2 ðk; Me3 Þ=ðhD UÞ (35) the motive mass flow rate at primary inlet. Therefore, they
can be obtained as follows:

Uq1=2 þ 1 m2
f2 ðk; Me4 Þ ¼ f3 ðk; Me4 Þx (36) Uej1 ¼ (46)
FU m1
1180 international journal of refrigeration 32 (2009) 1173–1185

0.6
1 2 3 curve Φ ξopt
1 2 1.48
0.5
2 2.5 1.79
3 3 2.19
0.4

0.3

U
0.2

0.1

0
1 1.5 2 2.5
r

Fig. 9 – Influence of geometric ratio F on the characteristic


U(r) in the transition regime, R141b.

5 Computation methodology

A computer simulation model has been developed on the


basis of the one-dimensional ejector model. The model
includes operating parameters T00 , T000 and geometric parame-
ters: F, U. The ratio of cylindrical chamber length to diameter
is assumed to be 10. The coefficients F and hD are assumed to
be: F ¼ 0.06, hD ¼ 0.96, U ¼ 3 (Lu, 1986).
This computational procedure yields the output of critical
Fig. 7 – Computer program flow diagram. entrainment ratio, U, and the back pressure Pe4 . The analysis
procedure follows the flowchart shown in Fig. 6.
The ejector model in transition regime is then employed in
a computer program, whose flow diagram is shown in Fig. 7, to
m3 evaluate the NMECS performances. All the equations are
Uej2 ¼ (47)
m1 þ m2 solved by the knowledge of operating temperatures. The
In the compressor, by supposing an adiabatic trans- refrigerants’ thermodynamics properties are calculated by
formation between 1 and 2, the mechanical work is expressed using NIST database subroutines and the calculating program
as: is written with Fortran Language. Several environment
friendly refrigerants like R290, R152a, R717, R600a and some
ðh2  h1 Þ transitory refrigerants R141b and R134a are applied to the
Wcomp ¼ (48)
hcomp above computation models for the simulation of the novel
where hcomp the compressor efficiency is given by Brunin et al. system.
(1997) as:

PH 6 Results and discussion


hcomp ¼ 0:874  0:0135 (49)
PL
The following results are obtained when simulation conditions
are given as: condensing temperature TC ¼ 45  C, evaporating
U (r) the present study
0.5 2 3
U(r) Nehdi et al.
1 2 R290 R717 R600a
0.4 curve Φ ξth ξexp R250a R134a R141b
1 4 2.96 3
0.3 2 4.84 3.61 3.5
3 5.76 3.92 4
U

COPr

0.2 1.5

0.1

0.0
1 2 3 4 1
r 1.8 2 2.2 2.4 2.6
Φ
Fig. 8 – Comparison of simulated results for R11 with
experimental data. Influence of area ratio F on the Fig. 10 – Influence of the area ratio on the COPr for various
characteristic U(r) in the transition regime. pure fluids.
international journal of refrigeration 32 (2009) 1173–1185 1181

Uej1 Uej2 COPr Φ =2 0.8


Φ=2
0.4 2
0.75

Wr (Wne/Wco)
0.3 1.5 0.7

0.65

COPr
0.2 1
U

0.6

0.1 0.5 0.55

0.5
0 0 R290 R717 R600a R152a R134a R141b
R290 R717 R600a R152a R134a R141b
Fig. 13 – Comparison of mechanical power request by new
Fig. 11 – Entrainment ratios Uej1 and Uej2 and COPr for
and conventional cycle for various pure fluids.
various pure fluids.

what corresponds, therefore, to a low value of back pressure


temperature Tev1 ¼ 5  C, Tev2 ¼ 18  C, Tev3 ¼ 28  C, super- and to a higher value of x. The same results were found by
heating DTsup ¼ 5  C. The superheat values allow realizing Paliwoda (1968). His tests with R11 for F included between
dry expansions in the primary and the secondary nozzles 4 and 6 showed that the higher F gives an important
(i.e. that the temperature of the fluid at the end of its expansion entrainment ratio, but on the other hand a low value of F
is higher than its saturation temperature in the final state). increases the compression ratio.
However, the new configuration of the CMECS, NMECS
6.1 Results of the ejector simulation model uses the two ejectors to increase the aspiration pressure at the
compressor. Therefore, the two driving pressures xej1 and xej2
The entrainment ratio U is a function of the operating condi- must be low what corresponds to the minimum possible value
tions and the geometry of the ejector defined mainly by the of F.
parameter F (Nahdi et al., 1993; Lu, 1986; Paliwoda, 1968). In
order to validate the ejector simulation model, the design 6.2 Effect of fluid nature on the NMECS performance
conditions of the experimental system using R11 and reported
by Nahdi et al. (1993) were reproduced. Fig. 8 shows the Fig. 10 represents the variation of the COPr (the COP ratio of
evolution of the entrainment ratio U with the compression the new and the conventional cycle) with several values of
ratio r (curves U(r)) and the optimum driving pressure ratio xopt area ratio F. It can be seen that the performance of the novel
for different values of area ratio F for TR. The simulated system varies from refrigerant to refrigerant and the novel
results agree fairly well with the experimental data. cycle has a remarkable improvement in COP over the
For the operating temperatures selected in the present conventional cycle whatever the value of the area ratio F.
work, the effect of the area ratio F on the characteristic U(r)x Fig. 11 is generated to indicate the variation of entrainment
using R141b was examined and represented in Fig. 9. In TR ratios Uej1 and Uej2 with the COPr for several refrigerants and
(curves giving optimum performances), if F increases, the by tacking F ¼ 2.
optimum curve moves towards the increasing entrainment The improvement is due to the increase of the inlet pres-
ratio and the increasing driving pressure ratio and leads to in sure at the compressor. This is reinforced by Fig. 12, which
increase of the ejector performances. In fact, for a high value represents the different pressure ratios Pr (compressor aspi-
of F, the mixing chamber section is relatively greater and the ration pressure ratio of the new and the conventional cycle)
primary flow must expand to form the aerodynamic throat,

0.5 Uej1 Uej2 rej1 rej2 COPne Φ=2 4


2.5
0.4
3
2
COP, R

0.3
1.5 2
U

0.2
Pr

1
1
0.1
0.5
0 0
-5 0 5 10
0
R290 R717 R600a R152a R134a R141b Tev1

Fig. 12 – Comparison of pressure ratio Pr at the compressor Fig. 14 – The effect of the evaporating temperature Tev1 on
aspiration for new and conventional cycle for various pure the entrainment ratios Uej1 and Uej2, the compression
fluids. ratios rej1 and rej2 and the COPne, R141b.
1182 international journal of refrigeration 32 (2009) 1173–1185

Uej1 Uej2 rej1 rej2 COPne Φ=2 5


Φ=2
0.5 3
4
0.4
3

COPne
2

COP, r
0.3
2
U

0.2
1 1
0.1
0
0 0 30 35 40 45
-20 -15 -10 Tc
Tev2
Fig. 17 – The effect of the condensing temperature Tc on the
Fig. 15 – The effect of the evaporating temperature Tev2 on COPne, R141b.
the entrainment ratios Uej1 and Uej2, the compression
ratios rej1 and rej2 and the COPne, R141b.

Fig. 14 represents, for R141b, the influence of the evapo-


rator temperature Tev1 on the entrainment ratios Uej1 and Uej2,
obtained by using several fluids. The increase of pressure at the compression ratios rej1 and rej2 and the COPne. It is noted
the compressor aspiration leads to a decrease of the pressure that for a fixed condenser temperature, the compression
ratio across the compressor and thus, the compression work ratios and COPne increase and Uej1 and Uej2 decrease when Tev1
decreases and the improvement in COP increases, Fig. 13. increases. When compression ratios increase, the back pres-
Figs. 10 and 11 show that R141b gives the highest variation sures at the ejectors exit increase. Any increase in the back
of COP between conventional and new cycle. It appears that pressure increases the COPne, reduces driving pressure ratio
R141b gives the uppermost value of COPr. We can see that the and subsequently entrainment ratio. Hence, critical entrain-
COP in NMECS using R141b is increased by 80% for F ¼ 2 and by ment ratio decreases when compression ratio increases.
50% for F ¼ 2.7. Fig. 15 depicts the effect of Tev2 on the COPne, Uej1, Uej2, Rej1
The same result about R141b performances in an ejector- and Rej2. The entrainment ratio Uej1 is found to increase with
compression refrigeration cycle was obtained by Dorantès and increasing evaporating temperatures Tev2 since any increase
Lallemand (1995). Several studies as Huang and Chang (1999) in Tev2 decrease the compression ratio Rej1. In addition, the
indicate that R141b is a good working fluid for an ejector since secondary mass flow rate in the ejector1 increases with Tev2,
R141b has a positive-slop saturated vapour line in the ther- the mass flow rate ratio Uej1 increases. Uej2, Rej2 and COPne vary
modynamic T–s diagram. This will not produce condensation slightly with the evaporating temperature Tev2.
of the vapour during an isentropic expansion in the ejector, Fig. 16 shows the variation of Tev3 on the COPne, Uej1, Uej2,
and thus reduces loss. Rej1 and Rej2. Uej1 and Rej1 are not affected by the variation of
Tev3, what is proved in Fig. 16. Furthermore, Rej2 decreases
with increasing Tev3 and, hence Uej2 will increase. COPne
6.3 Effects of operating temperatures on the NMECS changes slightly with the evaporating temperature Tev3.
performance The effect of the condenser temperature on the COP of the
NMECS (COPne) is examined. As indicated in Fig. 17, the
The effect of operating temperatures on the performances is increase in the condenser temperature causes the COPne to
studied here for R141b because it gives better performances decrease. Since the same amount of vapour cannot be
and it has a relatively low GWP. compressed to a higher condenser pressure, COP will decrease.

Uej1 Uej2 rej1 rej2 COPne Φ=2


0.2 3 7 Conclusion

The novel refrigeration cycle and its performance were dis-


2 cussed theoretically in this paper. A one-dimensional math-
COP, r

ematical model was developed using the equations governing


U

0.1

the flow and thermodynamics based on the constant-area


1
ejector flow model. The model includes effects of friction at
the constant-area mixing chamber. From the results, it can be
concluded that the novel cycle, using an ejector to the
0 0
-30 -28 -25 recompression of the vapour, can increase the cycle COP. For
Tev3 the same operating temperatures of the ejector refrigeration
systems, R141b gives the most advantageous relative coeffi-
Fig. 16 – The effect of the evaporating temperature Tev3 on cient of performance.
the entrainment ratios Uej1 and Uej2, the compression The new system is even more advantageous than the
ratios rej1 and rej2 and the COPne, R141b. indirect refrigeration system since this last gives an
international journal of refrigeration 32 (2009) 1173–1185 1183

improvement of 15% (California Energy Commission, 2004;  


T kþ1 2
Horton and Groll, 2003) relative to the conventional cycle. An ¼ 1 M (A.6)
T0 k1
additional theoretical and experimental work will be neces-
(2) r =P
sary to investigate the ejector and the design of the multi-
evaporators single compression system and verify the results.
r 1 k k
¼ ¼ ¼ (A.7)
P RT kRT a2
Appendix A (3) r=r

r r r
¼  0
Transformation of the expression mV þ PA r r0 r

 1=ðk1Þ
We have m ¼ rVA so r P T0 k1 2
¼  ¼ 1 M
  r0 P0 T kþ1
P  r 
mV þ PA ¼ rAV2 þ PA ¼  P0 A  1 þ V2 (A.1) Therefore:
P0 P
 1=ðk1Þ  1=ðk1Þ
r kþ1 k1 2
r r  ðr=r Þ r r P ¼ 1 M (A.8)
¼ ¼   r 2 kþ1
P P  ðP=P Þ P r P
(4) P =P
(1) P=P0
 k=ðk1Þ
 1g  g1
g P kþ1 k1 2
T P0 g P ¼  M
¼ ¼ P 2 2
T0 P P0  k=ðk1Þ  k=ðk1Þ
kþ1 k1 2
¼ 1 M (A.9)
2 kþ1
V2
CpT0 ¼ CpT þ
2 By using Eqs. (A.5), (A.7), (A.8) and (A.9), (A.1) can be
  rewritten as
V2 T "
¼ CpT0 1  k=ðk1Þ
2 T0 k1 2  kM2
mV þ PA ¼ P0 A  1  M2 þ
kþ1 kþ1
gr  1=ðk1Þ #
Cp ¼ k1
g1 1  M2 0mV þ PA ¼ P0 A
kþ1
where r is given byr ¼ R=M, M is the molecular weight of the " 1=ðk1Þ "
k1 2  kM2
gas.  1  M2 
kþ1 kþ1
Therefore
 1 ## " 1=ðk1Þ
  k1 k1
V2 gr T þ 1  M2 ¼ P0 A 1   M2
¼ T0 1  (A.2) kþ1 kþ1
2 g1 T0   #
2  kM2  ðk þ 1Þ  M2
þ1 ¼ P0 A
k¼g kþ1
" 1=ðk1Þ #
  ðk1Þ=k  k1  
2krT0 p 1 M 2
 1þM 2
V2 ¼ 1 kþ1
k1 p0

The sound speed at the nozzle throat is given by:


mV þ PA ¼ P0 Af ðk; MÞ
2kr
a2 ¼ kRT ¼ T0 (A.3)
kþ1 where:
 1=ðk1Þ
With Eqs. (A.2) and (A.3) and by introducing the dimen-   k1
f ðk; MÞ ¼ 1 þ M2  1   M2 0mV þ PA
sionless velocity M ðM ¼ V=a Þ, we obtain: kþ1
A
¼ P0   A  f ðk; MÞ
  A
V2 k þ 1 T
M2 ¼ ¼ 1  (A.4) we have:
a2 k1 T0
m ¼ m
 ðk1Þ=k !
kþ1 P This gives:
M2 ¼ 1
k1 P0

then
A r
¼ M
 k=ðk1Þ A r
P kþ1 2
¼ 1 M (A.5)
P0 k1 And by using Eq. (8), we obtain the function f2 ðk; MÞ
1184 international journal of refrigeration 32 (2009) 1173–1185

 1=ðk1Þ  1=ðk1Þ sffiffiffiffiffi  1=2


A kþ1 k1 2 T0 kþ1
¼ M 1 M ¼ f2 ðk; MÞ (A.10) ¼
A 2 kþ1 T 2

Eq. (A.14) becomes:


f ðk; MÞ 1 þ M2  k=ðk1Þ  1=2 rffiffiffi
mV þ PA ¼ P0 A ¼ P0 A 1 2 kþ1 k
f2 ðk; MÞ Mððk þ 1Þ=2Þ
1=ðk1Þ m ¼ P0 pffiffiffiffiffiAf2 ðk; MÞ
T0 kþ1 2 R
 1=ðk1Þ  
2 1
¼ P0 A Mþ For the primary fluid, the flow rate expression is expressed
kþ1 M
as:
 1=ðk1Þ  1=2  k=ðk1Þ  1=2
1 2 1 k 2 kþ1
f1 ðMÞ ¼ M þ 0mV þ PA ¼ P0 A f1 ðMÞ (A.11) m0 ¼ P00 pffiffiffiffiffi0 A0 (A.15)
M kþ1 T0 R kþ1 2

And since V  a 0V=a /0 And with an expression similar to (A.15), we can calculate the
 1=ðk1Þ  1=ðk1Þ mass flow rate in the section 3,
r k1 2
¼ 1 ¼  1=2  k=ðk1Þ  1=2
r0 kþ1 kþ1 1 k 2 kþ1
me3 ¼ P0e3 pffiffiffiffiffiffiffiffiAe3 (A.16)
T0e3 R kþ1 2
Eq. (11) is expressed then as:
The fictitious throat A*e3 can be expressed as:
r
mV þ PA ¼ P0  A f2 ðMÞ
r0  1=ðk1Þ  1=ðk1Þ
Ae3 kþ1 k1 2
¼ Me3 1  Me3
Ae3 2 kþ1
P0 kRT0 a20 ððk þ 1Þ=2Þa2 k þ 1 2
¼ RT0 ¼ ¼ ¼ ¼ a
r0 k k k 2k   1=ðk1Þ  1=ðk1Þ
A kþ1 k1 2
f2 ðk; MÞ ¼ ¼ M 1 M
A 2 kþ1
kþ1 2
mV þ PA ¼ a r A f3 ðMÞ
2k   Then Ae3 ¼ Ae3 f2 ðk; Me3 Þ
By combining the Eq. (A.15) attributed to the section (e2)
m ¼ a r A ¼ arA and Eq. (A.16), we obtain:
sffiffiffiffiffiffiffiffiffi
Finally, we obtain: me3 P0e3 T00 Ae3
¼ f2 ðk; Me3 Þ (A.17)
m0e2 P00 T0e3 A0
kþ1
mV þ PA ¼ a mf1 ðMÞ (A.12)
2k
sffiffiffiffiffiffiffiffiffi 0
P0e3 me3 T0e3 A 1
Calculation of the stagnation pressure P0e3 in the section e3 ¼ 0 0
P00 me2 T0 Ae3 f2 ðk; Me3 Þ

The mass flow rate is given by:


pffiffiffi T0e3 1 þ Uq
P a k ¼
m ¼ A V r ¼ A V ¼ A pffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffir T00 1þU
RT kRT RT
me3
a ¼1þU
pffiffiffiffiffiffiffiffiffiffiffi ¼ 1 m0e2
kRT
And by introducing the parameter F ¼ Ae3 =A0 we obtain:
rffiffiffi rffiffiffi
k 1 Ar k P0e3 1 þ UðqÞ1=2
m ¼ A r  pffiffiffiffiffi0m ¼ pffiffiffiffiffi (A.13) ¼ (A.18)
R T T R P00 Ff2 ðk; Me3 Þ
By introducing ðA =AÞ and ðP =P0 Þ in the Eq. (A.13), we obtain:
pffiffiffiffiffirffiffiffi
A P 1 T0 k references
m¼A P0 pffiffiffiffiffi pffiffiffiffiffi
A P0 T0 T R

sffiffiffiffiffirffiffiffi
1 A P T0 k Abramovich, G.N., 1970. Applied gasdynamics. Gas Ejectors.
m ¼ P0 pffiffiffiffiffiA (A.14) Foreign Technology Division, Air Force Systems Command,
T0 A P0 T R
Ohio, USA. no. F33657-70-D-0607-P002.
Brunin, O., Feidt, M., Hivet, B., 1997. Comparison of the working
A domains of some compression heat pumps and
¼ f2 ðk; MÞ
A a compression-absorption heat pump. International Journal of
And by knowing Refrigeration 20, 308–318.
California Energy Commission, 2004. Final Report – Investigation
 k=ðk1Þ of Secondary Loop Supermarket Refrigeration Systems.
P 2
¼ Contract Number 500-98-039.
P0 kþ1
Dorantès, R., Lallemand, A., 1995. Influence de la nature des fluides,
and purs ou en mélange non-azéotropiques, sur les performances
international journal of refrigeration 32 (2009) 1173–1185 1185

d’une machine de climatisation à éjecto-compresseur. Nahdi, E., Champoussin, J.C., Hostache, G., Cheron, J., 1993.
International Journal of Refrigeration 18 (1), 21–30. Optimal geometric parameters of a cooling ejector-
Gokthun, S., 2000. Optimazation of irreversible solar assisted compressor. International Journal of Refrigeration 16 (1), 67–72.
ejector-vapour compression cascaded system. Energy Conver NIST Standard Reference Database 23, 2006. Version 7.1.
Mngmnt 41, 625–631. Paliwoda, P., 1968. A review paper on the experimental study on
Gosney, W.B., 1982. Principle of Refrigeration. Cambridge low-grade heat and solar energy operated halocarbon vapour-
University Press, Cambridge. jet refrigeration systems, Topical studies. IIR Bulletin 1003.
Horton, W.T., Groll, E.A., 2003. Secondary loop refrigeration in Sokolov, M., Hershgal, D., 1990a. Enhanced ejector refrigeration cycles
supermarket applications: a case study. International powered by low grade heat. Part 1. Systems characterization.
Conference of Refrigation Paper ICR0345, Washington. International Journal of Refrigeration 13, 351–356.
Huang, B.J., Petrenko, V.A., Chang, J.M., Lin, C.P., Hu, S.S., 2001. A Sokolov, M., Hershgal, D., 1990b. Enhanced ejector refrigeration
combined-cycle refrigeration system using ejector cooling cycles powered by low grade heat. Part 3. Experimental
cycle as the bottomcycle. International Journal of results. International Journal of Refrigeration 14, 24–31.
Refrigeration 24, 391–399. Sokolov, M., Hershgal, D., 1990c. Enhanced ejector refrigeration
Huang, B.J., Chang, J.M., 1999. Empirical correlation for ejector cycles powered by low grade heat. Part 2. Design procedures.
design. International Journal of Refrigeration 22, 379. International Journal of Refrigeration 13, 357–363.
Jessica A. Carroll, Barry Brophy, Donal P. Finn, David J. Timoney, Sokolov, M., 1991. Operational envelope and performance curves
2005. Development of a steady state mathematical model to for a compression-enhanced ejector refrigeration system.
simulate the performance of an indirect multi-temperature ASHRAE Transactions 17 (part 2), 394–402.
refrigeration system. International Conference on Latest Sokolov, M., Hearshgal, D., 1993. Solar-powered compression
Developments in Refrigerated Storage, Transportation and enhanced ejector air conditioning. Solar Energy 51 (3), 183–194.
Display of Food Product Amman-Jordan. Stoecker, W.F., 1958. Steam-jet Refrigeration. McGraw-Hill,
Le Grives, E., Fabri, J., 1969. Divers régimes de mélange de deux Boston, MA.
flux d’enthalpie d’arrêt différentes. Astronautica Acta 14, Sun, D.W., 1997. Solar powered combined ejector-vapour
203–213. compression cycle for air conditioning and refrigeration.
Lu, L.T., 1986. Etudes théorique et expérimentale de la production Energy Conversion Management 38 (5), 479–791.
de froid par machine tritherme a éjecteur de fluide frigorigène. Tomasak, M.L., Radermacher, R., 1994. Analysis of a domestic
Ph.D. thesis, Laboratoire d’Energétique et d’Automatique, de refrigeration cycle with an ejector. ASHRAE Transactions 101
l’INSA de Lyon, France. (45), 1431–1438.

View publication stats

You might also like