Performance simulation of a two-evaporator refrigerator-freezer

charged with pure and mixed refrigerants

D. S. Jung and R. Radermacher
Department of Mechanical Engineering, University of Maryland, College Park, M D 20742,
USA
Received 5 J a n u a r y 1990; revised 17 July 1990

A computer simulation of two-evaporator refrigerators charged with pure and mixed refrigerants operating
on a Lorentz and Meutzner's cycle has been performed to determine possible substitutes for RI2 with an
improved energy efficiency. The results for pure fluids indicate that the coefficients of performance (COPs)
and volumetric capacities obtained with two-evaporator units are enhanced up to 6 and 15.5%, respectively,
compared to those with single evaporator units. This is due to some of the evaporation occurring at higher
temperatures in the two-evaporator units. For mixtures, a significant increase in COP of up to 18% is
observed by matching the large overall drop in temperature of the air streams, 23C, with that of refrigerant
mixtures in the evaporator. The evaporator area ratio is a very important parameter and depends largely on
the load distribution between the two compartments. The effect of the low temperature heat exchanger has
been studied by varying its size. As the size increases, the COP also increases with a decreased pressure ratio
across the compressor. An optimized two-evaporator refrigerator freezer unit charged with alternative,
ozone-safe refrigerant mixtures may increase the energy efficiency considerably, helping to alleviate the
environmental impact of refrigeration.

(Keywords: computer simulation; two-evaporator refrigerators; energy efficiency)

Simulation de la performance d'un r6frig6rateur-cong61ateur fi deux 6vaporateurs, charg6 avec du frigorig6ne pur
ou en m61ange
On a simulk par ordinateur des rbfrigbrateurs ?1 deux bvaporateurs, chargbs avec des frigorigbnes purs ou en
mblange, fonctionnant selon un cycle de Lorentz et Meutzner, dans le but de dbterminer des substituts possibles
du R12 avec une fficacitb bnergbtique accrue. Les rbsultats avec les fluides purs montrent que les coefficients de
perJbrmance et les puissances volumbtriques, obtenus avec des unitbs a deux bvaporateurs, augmentent de 6 et
15,5 % respectivement, par rapport h des unitbs gt bvaporateur unique, parce qu'une partie de l~vaporation se
produit h une tempbrature plus blevbe dans les unitbs h deux bvaporateurs. Pour les mblanges, on observe une
augmentation importante du COPjusqu'h 18 % due h une adbquation entre la chute globale et importante de
tempbrature de l'air en circulation, 23 C, et celle du m~lange de frigorigbnes dans l'bvaporateur. Le taux de la
surface bvaporante est un paramOtre trbs important et dbpend considbrablement de la distribution des charges
entre les deux compartiments. On a bgalement btudib l'effet de l'bchangeur de chaleur ~ basse tempbrature par
variation de sa dimension. Si la dimension augmente, le COP augmente bgalement avec un taux de compression
rbduit au compresseur. Un rbfrigbrateur-congblateur h deux bvaporateurs que l'on a optimisb et chargb avec un
mblange de substituts sans effet sur l'ozone, peut augmenter considbrablement l'effi'cacitb bnergbtique, ce qui
contribue h diminuer l'impact de la production de Jroid sur l'environnement.

(Mots cl6s: simulation par ordinateur; r&rig6rateurs/~ deux 6vaporateurs; efficacit6 6nerg6tique)

The agreement of the Montreal ProtocoP to regulate the

production and trade of ozone-depleting substances has
greatly influenced the air-conditioning and refrigeration
industries. At present, extensive research is being undertaken at many institutions to replace the fully halogenated chlorofluorocarbons (CFCs), R I I , R12, R I 3 B I ,
R113, R114 and R! 15 covered by the Montreal Protocol. One of the major uses of CFCs is in domestic refrigerator-freezers (here referred to as 'refrigerators')
which predominantly utilize R12 as the working fluid.
The work reported here is an extension of previous
studies z dealing with conventional refrigerators with a
single evaporator. In the previous work, only drop-in
replacements for R 12 were considered as the replacement
014~7007/91/050254~ 10
~C 1991 Bunerworth Heinemann Ltd and fIR

254

Int. J. Refrig. 1991 Vo114 September

would be even more costly if it were accompanied by any

major system modifications. The results indicated that
there are a number of mixtures which can be used as
drop-in replacements. These fluids, however, are not
expected to yield a significant increase in the coefficient
of performance (COP) due to a small drop in temperature of the air stream in the evaporator side. In other
words, one of the favourable characteristics of mixtures,
a gliding temperature effect, cannot be fully utilized in
conventional refrigerators with drop-in replacements
due to an inherent restriction in the air stream tempera-

tures.
It is expected that the greenhouse effect will be a more
serious issue than ozone layer depletion in the next few

Simulation of a two-evaporator refrigerator: D. S. Jung and R. Radermacher

Nomenclature

A
Cp
DSH
DT
H
LMTD
m
P
Q

REo
U
S
T
TS
V
VC
VAo
W
X

Heat transfer area (m 2)

Specific heat (kJ kg-~ C-L)
Degree of superheat (C)
Specified log mean temperature difference
Specific enthalpy (kJ kg-~)
Log mean temperature difference
Mass flow-rate (kg s-~)
Pressure (kPa)
Rate of energy (W)
Array containing residuals
Overall heat transfer coefficient (W m 2K-I)
Specific entropy (kJ kg-lC 1)
Temperature (K or C)
Heat transfer fluid temperature (K or C)
Specific volume (m 3 kg-1)
Volumetric capacity (kg m 3)
Array containing unknowns
Power (W)
Overall composition

Greek

~c

Isentropic compressor efficiency

decades 3. One of the most effective means of helping to

solve the greenhouse problem is to increase the energy
efficiency of energy conversion devices, including refrigerators. Considering that virtually every household in
the developed world is equipped with a refrigerator, an
increase in the efficiency of domestic refrigerators would
mean considerable energy savings. Consequently, any
study involving the replacement of RI2 with ozone-safe
fluids should also focus on increasing the energy
efficiency.
The current design of the conventional one-evaporator
refrigerator, however, has its own shortcomings. American standards for domestic refrigerators typically require
the maintenance of 5 and - 15C in the food compartment and freezer section, respectively. In other words,
two distinct evaporator temperature levels exist. From
the thermodynamic requirement that the smaller the
temperature lift (the temperature difference between the
condenser and evaporator), the higher the COP, the COP
for cooling the food compartment is higher than that for
the freezer due to the smaller temperature lift.
The current configuration supplies cooling at the
freezer condition of - 15C, even if the food compartment requires cooling at 5C. An air damper is used to
maintain the food compartment temperature at 5C. This
damper splits the mass flow of air into two streams
according to need. The cooling load in the food compartment is typically 40-60% of the total cooling load. Thus,
for this 40-60% of the cooling load, the refrigeration
cycle operates very inefficiently, without taking advantage of the smaller temperature lift necessary in this
compartment.
One possible solution to the above problem is to
provide two independent refrigeration cycles for the
freezer and food compartments. This design, with two
separate refrigeration circuits, would considerably
increase the initial cost of the refrigerator. Another possible means of increasing the energy efficiency is to utilize
refrigerant mixtures as working fluids in refrigerators.

Chx

Heat exchanger correction factor

Percentage increase in COP

Subscripts
air
Air stream
amb
Ambient
c, cond Condenser
e, evap Evaporator
he, hte High temperature evaporator
hthx
High temperature heat exchanger
le, lte
Low temperature evaporator
lthx
Low temperature heat exchanger
1
Liquid
max
Maximum
o
Overall
12S
Compared to R12 in a single-evaporator
unit
12T
Compared to RI2 in a two-evaporator unit
ref
Refrigerant
sub
Subcooled
suph
Superheated
tp
Two-phase
v
Vapour
vc
At the same volumetric capacity of R12

During the evaporation of non-azeotropic refrigerant

mixtures (NARMs) at constant pressure, the saturation
temperature changes due to the volatility difference of
the pure components, which is termed a gliding temperature effect. By choosing a mixture with a large gliding
temperature difference (GTD), two temperature level
evaporation may be obtained in a single refrigeration
cycle, resulting in an increase in COP with only a moderate increase in the initial cost.
Figures 1 and 2 illustrate a schematic and a temperature-entropy diagram, respectively, for a two-evaporator
refrigerator charged with mixtures originally proposed
by Lorenz and Meutzner 4. Both the refrigerant and air
stream temperatures are illustrated with heat flows for
reference. Compared to the conventional unit, the only
difference is the addition of a high temperature evaporator (HTE) and a low temperature heat exchanger
(LTHX).
At state 7, the two-phase refrigerant enters the low
temperature evaporator (LTE). Evaporation occurs due
to heat exchange with the freezer air stream, with the exit
condition still in the two-phase region (state 10).
Through the LTHX, further evaporation is achieved
(from state 10 to state 11) while the liquid is being subcooled further internally (from state 9 to 12). Further
heat is supplied by the food compartment air stream in
the HTE to complete the evaporation and usually superheated vapour leaves the evaporator (state 1). During
evaporation through the LTE, LTHX and HTE, the
refrigerant temperature rises for mixtures (gliding temperature effect), while it remains constant for pure components without consideration of the pressure drop.
On examining the temperature-entropy diagram
(Figure 2), it can be seen that it is possible to match the
overall GTD of the refrigerant to the overall temperature
drop of the two air streams in both compartments
(roughly AT= TS~ - T S 7 = 20-300C). This solves the previous problem in one evaporator unit of having a smaller
temperature drop in the air stream (AT= 5 7C) and
Rev. Int. Froid 1991 Vo114 Septembre

255

Simulation of a two-evaporator refrigerator. D. S. Jung and R. Radermacher

( Sat. liquid ) ( Sat, vapor )
51
14
w

6X

TS6

enser

[3

J~AAAAAA/%---L---~--TS3
r"''"r

TS4

TS$

High-Temperature
Heat Exchanger

Com wessor
High-Temperature
Evaporator

-TS,
-

"1" ( SatS.

~ll--~ ] - - -

TS8

Vapor'l~

Low.Temperature

Heat Exchanger
~TSI0
I

Q LTE ~ - - - ~

TS7

Low-Temperature

Evaporator
Figure 1

Schematic diagram of a two-evaporator refrigerator

Figure 1 Schema d'un r~'frig~rateur Zt deux ~vaporateurs

Refdgeram
Air

T~

2 ~

Entropy(KJ/ KgK)
Figure 2 Temperature~ntropy diagram for a two-evaporator refrigerator
Figure 2 Diagramme temperature entropie pour un r~['rig~rateur

deux kvaporateurs

not allowing the maximum potential of the mixtures to

be exploited. When the temperatures are well matched,
the mixtures with the larger G T D yield a higher COP due
to a larger reduction in the compressor work. Thus, the
key point in Lorenz and Meutzner's 4 cycle is that the
favourable characteristics of mixtures, i.e. the G T D in
evaporation, can be fully utilized with a larger temperature drop available in the air streams.
Lorenz and Meutzner 4 experimentally observed a
power saving of 20% compared to that of R I 2 when an

256

Int. J. Refrig. 1991 Vo114 September

R 2 2 - R I 1 mixture at 0.5 mass fraction of R22 was used in

a two-evaporator refrigerator. They also experienced a
favourable control in the cycle due to the good matching
of evaporation temperatures against the air temperatures
in both compartments. Based on their experimental findings, certain criteria were established to help choose the
best mixtures to yield a considerable gain in COP. The
mixtures should have an actual G T D of roughly 20 to
25C in the evaporator to give a good match with the air
stream temperatures. Owing to pressure drop, however,
the theoretical G T D in the evaporator would be near to
30C. This means that the boiling point difference
between two pure components at the evaporator pressure is close to 60C. Another desirable characteristic is
that the pure components should have high volumetric
capacities.
Stoecker 5 used an R I ~ R 1 1 4 mixture in a pilot refrigerator unit and failed to obtain any increase in COP,
which is in sharp contrast to the results of Lorenz and
Meutzner 4. An Oak Ridge National Laboratory report 6
revealed that when an R22-R114 mixture was substituted for R12 in an actual two-evaporator refrigerator
unit, no significant change in performance was observed
at all compositions. In recent discussion, Kruse ? presented experimental results obtained with an R 2 2 - R 142b
mixture in a two-evaporator unit. Increases in the COP
of 10 and 1.5% in the pull-down and steady-state tests,
respectively, were observed.
This strongly indicates that although the use of
N A R M s is a thermodynamically sound idea, the success
strongly depends on the operating conditions, applications and theoretical COPs of the chosen refrigerant
pairs. As discussed by Lorenz and Meutzner 4 and Jung
and Radermacher 2, a detailed thermodynamic analysis
may be needed to select the proper mixtures before an
expensive experimental investigation is carried out. This
indicates the necessity for a proper model of two-evaporator refrigerators operating on Lorenz and Meutzner's 4
cycle.
Two reports on the simulation of two-evaporator
refrigerators are available. Stoecker and Walukas s
obtained results with R I 2 - R ! 14 mixtures. In the simulation, two thirds of the refrigeration load was arbitrarily
assumed to be performed at the LTE, with the remainder
performed at the HTE. The results showed a power
saving of 12% at 0.5 mass fraction of R114. The selection
of R12 R114, however, is not the best possible choice
according to the criteria set by Lorenz and Meutzne#
because the G T D is too small with an R I 2 - R 1 1 4 mixture.
Kruse 9 reported the results of the simulation of a twoevaporator refrigerator with R22 R l I 4 and R I 3 B 1 R114 mixtures. A similar approach to that of Stoecker
and Walukas s was taken in the simulation. The model
employed the Redlich K w o n ~ S o a v e (RKS) equation of
state to compute the properties needed. The same load
distribution as for Stoecker and Walukas 8 was assumed
in the analysis. For R22 R114 mixtures, the theoretically
calculated improvement in COP was 18 20% at 0.4 mass
fraction of R22. For the R I 3 B I - R 1 1 4 mixtures, a 20%
increase in COP was obtained at 0.7 mass fraction of
RI3BI.
The above results indicate that various features of
Lorenz and Meutzner's cycle4 were not fully examined
and hence a comprehensive model of two-evaporator

Simulation of a two-evaporator refrigerator. D. S. Jung and R. Radermacher

refrigerators is necessary to study the effects of load

distribution, L T H X size and fluid combination. The
Montreal Protocol necessitates the development of new
ozone-safe refrigerants such as R123, R124, R125, R134,
R134a, Rl41b, R142b and R143a. It would therefore be
of interest to investigate the fluid mixtures containing
these new refrigerants. The objectives of this work were
to describe the simulation method, to examine various
fluids of interest and to suggest any system modifications
necessary.
Simulation

In our previous work 2, a detailed description of singleevaporator refrigerators was given and hence only the
differences are discussed here. The only changes in the
single- and two-evaporator models are an additional evaporator and the L T H X (Figure 1). The total evaporator
area remains the same as that in a single evaporator
model to make a fair comparison. The area, however, is
distributed in both compartments according to the heat
transfer requirement. Initially, the size of the L T H X is
designed to be 7% of the total available evaporator area
with the same U as UE throughout the analysis. Later, the
effect of changing the size of the L T H X while the total
evaporator area is held constant is investigated. Except
for the H T E and LTHX, all the other components
remained unchanged. Further details are given in Jung
and Radermacher 2.
To compare the performance of various fluids on a fair
basis, it is reasonable to require that the net refrigeration
effect, Qe~p = Qh~ + Qh,, is constant regardless of the
working fluids. This is accomplished by specifying the air
stream temperatures entering and leaving the evaporators, TSm and TS7 in the LTE and TS. and TSI. in the
HTE, with fixed mass flow-rates, m~,~and rht,~, generated
by fans in the food compartment and freezer. As the
compressor power, W~, varies with the working fluids, so
does the coefficient of performance. (Note that fan power
is not considered in the COP.)
As W~ differs with various fluids, the heat discharged
through the condenser, Q~o,d, which is the sum of Qevap
and We, also depends on the fluids. The air stream temperature entering the condenser, TS6, is assumed to be at
ambient. Because the volume flow-rate of the air stream
on the condenser side, me, is fixed by another fan, the
temperature of the air stream on the condenser side
including TS3, TS4 and TS5 in Figure 1 must change to
satisfy the energy balance.
A proper formulation of heat transfer in the evaporator and condenser is necessary to simulate the overall
system performance. The evaporators and condenser are
specified by the product of an overall heat transfer coefficient and an area (UA). The overall heat transfer coefficient, U, was taken from actual measurements 6. The
overall heat transfer coefficients for the condenser and
the L T H X were assumed to be equal to UE.
With given UA values, the heat transfer in the H T E
and condenser is treated as:

LMTDht~ = (1 -f~uph~) LMTDtph +.f~.ph,LMTD~uph~

(1)

LMTD,: =~ubLMTD,;ub,:+ (1 --f~bc--fsupc)LMTDtp~ +

+.~.p~ LMTD~,p,:

(2)

Q~wp =

Ue(AjteLMTDtt~ + A ht~LMTDht~) %,

(3)

Q,:o,a= U~ At LMTD~ %,

(4)

wheref~uohc is the fraction of H T E heat in the superheated

vapour portion of HTE, f~ub~,,fsup~are the fractions of the
total condenser heat in the subcooled and superheated
portions of the condenser, LMTDtphe, LMTDsuphe are the
log mean temperature differences in the two-phase and
superheated portions of the HTE, LMTDs,b,:, LMTDtp,:,
LMTD,,p,: are the log mean temperature differences in
the subcooled, two-phase and superheated portion of the
condenser and %, is the heat exchanger correction factor
to account for the effect of having cross or parallel flow
heat exchange instead of pure counter-current heat
exchange in the heat exchangers.
Pressure drops in the condenser and evaporator are
prorated according to the amount of heat transfer in
subcooled, two-phase and superheated regions as (P2 and
P3 are assumed initially):
Pl = P2
P4 = P3 - A Pcfsupc

P5 = P3 - AP~ (1
P6 = / 9 =

P7 =

PI2 =

fsubc)

P3 - A P e

PI + APe

Ps = PI + APhe fsuphe
Plo = PII =

P T - APle

(5)

where APh0 and API~ are total pressure drops across the
H T E and LTE calculated as APhc = APe (Qhte/Qev,p) and
APse = APe-APhe. The total pressure drops in the evaporator and condenser, APe and APe, were imposed and
no pressure drop is assumed in the low and high temperature heat exchangers.
The unknowns resulting from the simulation of the
two-evaporator refrigerator were calculated by the Newton-Raphson (NR) method as described previously 2.
The successive substitution (SS) method, which was
shown to be as good as the NR method for a single
evaporator model, was employed at first. The addition of
another evaporator and LTHX, however, made the
simulation more complicated and the successive substitution method failed.
A proper model should calculate the refrigerant mass
flow-rate, mrcr, rather than specifying it to reflect that mr~r
varies as a working fluid is changed in the actual experiments. One of the main reasons for the failure with the
SS method is the calculation of mr~r. Were mrer specified,
the simulation would be much simpler and the SS
method would yield results. This, however, would not
yield meaningful results for proper comparison. This
problem was easily handled with the NR method as it
treats the simulation in a global manner with rnrer as an
explicit variable.
The N R method requires initial guesses for all
unknown variables so that the minimum number of
properties, as dictated by the Gibbs phase rule without a
chemical reaction, is known to fix the state. Table 1 lists
the known and unknown variables at each state point.
The total number of independent variables in the simulation is 10, as listed in Table 2. As the states at each point

Rev. Int. Froid 1991 Vo114 Septembre

257

Simulation of a two-evaporator refrigerator: D. S. Jung and R. Radermacher

Table 1 Known and unknown variables at each state point
Tableau 1 Variables connues et inconnues 2t chaque point d'~tat

State point K n o w n (fixed)

1

T&

T2 : Tamb

3
4
5
6
7
8
9
10
11
12

TS6 = T~m~
T~= T & + 0 . 1
TS7
TSm
TS,,

K n o w n (assumed)
P~. T~ = Ts + DSH, V,
P:, V2
P> $3 = S> ~
P4, X,
Ps, X,
P~, V6

P> H; = H~
P~, X,,
P~,T~.V~
Hm : H7 Jr- Ql,,./m~r. Pm
H,~ = H, -Qhtj;nrcr. P,,
Ti2, P,:, V,._

Unknown
H~
H>
H3.
H4,
Hs,
H~

$2
T . T&
14, TS4
T~. T &

T~. XQ7
H~. T~. T &
H.
Tto
T.
H~:

are known (either fixed or assumed), all unknowns such

as enthalpy can be evaluated by the Carhahan StarlingDeSantis equation of state t. With the N R method, typical run times on a 80386/20 M H z PC with a 80387/20
M H z maths-coprocessor were 35 and 105 s for pure and
mixed refrigerants, respectively.
Results and discussion

The steady-state performance simulation of the two-evaporator refrigerator cycle (TERCLE) was carried out
with 14 pure refrigerants and 17 mixtures. The specified
conditions for T E R C L E are as follows: superheat, 5C;
pressure drop in evaporator, 10 kPa; pressure drop in
condenser, 15 kPa; compressor efficiency, 0.55; heat
exchanger correction factor, 0.8; specified total cooling
load, 185 W; volume flow-rate in condenser, 100 CFM;
ambient temperature, 32C; UA in evaporator and condenser, 20 and 10 W C- ~, respectively; and total evaporator and condenser area: Ato,e= 0.2857, Atotc= 0.1428 m 2.
For mixtures, the overall composition (mass fraction
of the first component, always the more volatile component) varied from 0.0 to 1.0 in an interval of 0. I.
In conventional single-evaporator refrigerators, all the
cooling load is performed in the freezer, whereas in twoevaporator units the cooling load is distributed between
the two compartments according to requirements. The
first issue arising from this consideration is how to distribute the total cooling load to both evaporators. Stoecker
and Walukas ~ and Kruse 9 assumed that one third of the
total cooling load was performed in the H T E (Qh~c) while
the remainder was in the LTE (Q~c). The A S H R A E
Equipment H a n d b o o k on domestic refrigerators l~, however, indicates that roughly one half of the total cooling
is required in the HTE. Private discussions with the
engineers from major refrigerator manufacturing companies revealed that the fraction of Qhte varies from 0.4 to
0.6 depending on the refrigerator model.
As no definite value of the fraction of Qhte seems to be
known (actually it is a variable), two arbitrary cases are
considered in this study: Case 1, Qkc = 67%, Qhte= 33%;
and Case 2, Q,e = 50%, Qh~e= 50%. For both cases the
temperatures of the air stream entering and leaving the
evaporators are fixed to be T & =5, TSI~ = 0 , T&o = - 12
and T & = - 18C, which is necessary to maintain the
required temperature levels in both compartments. To
consider the changes in the cooling load in both compartments for both cases, however, the volume flow-rates

258

Int. J. Refrig. 1991 Vo114 September

of the air stream in both compartments were changed

accordingly.
Overall, two-evaporator refrigerators are assumed to
perform the same refrigeration duty (185 W) as that of a
single-evaporator unit with the same evaporator and
condenser area. To satisfy the energy balance between
the food and freezer compartments, the total evaporator
area has to be divided according to the need in both
evaporators. This was also carried out by T E R C L E . The
calculated area ratio, AR, is a percentage of the high
temperature evaporator area to the total area. The
importance of this parameter will be discussed later. The
present results would be much more meaningful if they
were compared to a certain baseline. For this purpose,
COP~2~ = 1.345 and VC12~= 769 (kJ m -3) of a single-evaporator refrigerator charged with R I2 with all other
parameters unchanged were used as baseline data 2.
Figures 3 and 4 illustrate the C O P , dplzx, which is a
percentage increase in COP compared to that of R12
with two evaporators, volumetric capacity and pressure
ratio of 14 pure fluids for Case 1. Only a few fluids such
as R22, R32, R142b and R152a have a comparable or
slightly higher COP than R12, which is similar to the
results obtained with a single-evaporator model. As the
volumetric capacity of all pure fluids are different from
that of RI 2, no single fluid seems to be a direct substitute
for RI2.
By having a portion of the evaporation at a higher
temperature in the food compartment, the COP and VC
of pure fluids are increased by up to 6 and 15.5%, respectively. This is caused directly by the elevation of the
evaporator pressure, resulting in a smaller pressure ratio
and a decreased vapour-specific volume entering the
compressor. For Cases 1 and 2, COP of R l 2 with a twoevaporator model, the C O P r a is 3.5 and 6% larger than
COP~2s while the VC of R I 2 with a two-evaporator
model is 8 and 15.5% higher than that with a singleevaporator model. This is in good agreement with the
thermodynamics that the smaller the temperature lift, the
larger the COP. From the above observation, it can be
said that the COP would increase in proportion to the
fraction of Qh~. This implies that if more effective insulation is considered to reduce the heat leakage, the freezer
section should be given priority.
Tables 3 and 4 summarize the results obtained with 17
mixtures for both cases. The best performance was
achieved with R22-R123 mixtures for both load distributions. The maximum increase in COP compared to
that of R 12 with a single evaporator, qb.... was 13.53 and
18.36% with R22 R123 mixtures. This increase was
achieved by two major contributions: evaporation at a
higher temperature in the food compartment and well
matched temperatures in the heat exchangers (mixture
effect). The fraction of the increase due to the mixture
effect alone is 0.75-0.8. Thus, the actual increase in COP
due to the favourable thermodynamics of the mixture
compared to that of R12 with a single evaporator is 10.86
(13.53 x 0.8) and 13.77 (18.36 x 0.75)% for Cases 1 and
2, respectively.
Unless major changes in the compressor are considered, the volumetric capacity of the mixture should be
very close to that of R I2 with a single evaporator. The
COPs obtained at the compositions in which the VC of a
mixture is the same as the VC of RI2, COPvc, are always
less than the maximum COP, as illustrated in Tables 3

Simulation of a two-evaporator refrigerator: D. S. Jung and R. Radermacher

Table 7~ Variablesand residualequationsfor the Newton Raphsonmethod
Tableau 2 Variables et bquations restantes pour la mdthode Newton Raphson

Variables

Residuals

Description

VA( l )=in,o~

RE(l)= - Qh,~+ UAh,~LM TDh, eh~

RE(2) = - Q.~ood+ U~A~LMTDeh,
RE(3) = - mrf(H,, - H~0)+ (UA L M TD) ....
RE(4) = - (//6 - H,~)+ (H2 - H,)
RE(5) = - H,, - H~o)+ (H~- H~.)
RE(6) = - (H3 - n6) xf~ob~+ (H5 - n6)
RE(7) = - (H3 -//6.) J~.r~+ (H3 -/44)
RE(8) = -Q.o.d+m~Cp.ir x ( T S 3 - TS~,)
RE(9) = - (H,- H~,) xf.uph~+ (H,- H~)
RE(10) = - Q~,.+ U~A~,~LMTD~,. ct,~

Heat transfer rate equation in HTE using Eq. (1)

Heat transfer rate equation in the condenser, Eq. (4)
Heat transfer rate equation in LTHX
Energy balance in HTHX
Energy balance in LTHX
Definition off, u~. Eq. (2)
Definition o f ] ~ , Eq. (2)
Air stream energy balance in the condenser
Definition ofJ~o~h,Eq. (1)
Heat transfer rate equation in LTE

VA(2)= ,2
VA(3)= P3
VA(4)= T9
VA(5)= TS3
VA(6)= TS~
VA(7)= TS~
VA(8)= TS~
VA(9)= T~
VA(IO)= A~,

Note that Lhereis no relationship between the variables and residual equations
On note qu 71 n'existe aueune relation entre les variables et les kquations restantes
1.5

25OO

TERCLE MODEL

TERCLE MODEL
2O0O

~1500

~10~

5011.

1.(3
3

6O
5O

40

-3

~-6
9

00

-12

0o0o
.
RI3B1

Rll

15
RI1 R 1 2 R I 3 B 1 R 2 2 R32 R l I 4 R I 2 3

R124

R134
RI41B
R143A
R134A
R142B
R152A

Figure 3 COP and percentage increase in COP for various pure refrigerants obtained with the two-evaporator refrigerator model
Figure 3 C O P et augmentation en pourcentage du C O P pour plusieurs

RI2

R32

R22

Rl14

.
.
.
.
.
R123
R134
RI41B
R143A
R124
R134A
R142B
RI52A

Figure 4 Volumetriccapacity (VC) and pressure ratio (PR) for various pure refrigerants obtained with the two-evaporator refrigerator
model
Figure 4 Puissance volum~trique ( V C ) et taux de compression ( P R )

ffigorig~nes purs, obtenus avec le module de r~frigkrateur ~ deux kvaporaleurs

pour plusieurs frigorig~nes purs, obtenus avee le module du r~frig~rateur h

deux ~;vaporateurs

a n d 4. The m a x i m u m percentage increases in C O P v c , dpv~,

are 11.52 a n d 16.95%, again observed with R 2 2 - R I 2 3
mixtures. O n l y two mixtures ( R 2 2 - R 1 2 3 a n d R32
R142b) seem to be the right substitutes for R12, with the
same compressor yielding m o r e t h a n I0 a n d 16%
increases in the C O P for Cases I a n d 2, respectively.
It w o u l d be interesting to use the T E R C L E model to
e x a m i n e the validity of the criteria set by Lorenz a n d
M e u t z n e r 4 for the selection of a mixture to be used in
t w o - e v a p o r a t o r refrigerators. The difference in the pure
c o m p o n e n t s ' boiling p o i n t temperatures should be larger
than 60C. Usually, the e v a p o r a t o r side pressure would
be close to atmospheric pressure. Hence, the n o r m a l
boiling points (NBPs) of pure fluids are used to approximately calculate the difference.
F i g u r e 5 shows the difference in N B P s o f various pure
refrigerants. The shaded squares in F i g u r e 5 c o r r e s p o n d
to the mixtures which have been s h o w n to yield more
t h a n a 10% increase in C O P as listed in T a b l e 4 for Case
2. The difference in N B P s for the shaded mixtures
actually varies from 27 to 74.4C a n d n o correlation

between the C O P a n d the difference in N B P s is apparent.

The best p e r f o r m a n c e was observed with R 2 2 - R 123 mixtures with a difference in N B P of 68.1C. Even t h o u g h
R23 seems to be a good candidate, yielding a wide range
of the differences in NBPs with m a n y other pure fluids,
its critical t e m p e r a t u r e is close to a m b i e n t t e m p e r a t u r e
a n d hence only a low c o n c e n t r a t i o n of R23 m a y be
practically used in a v a p o u r cycle.
As the R 2 2 - R I 2 3 mixture shows the best performance, more specific discussions will be m a d e with this
mixture. It was observed 2 that the m a x i m u m C O P occurs
when the G T D in the e v a p o r a t o r best matches the temperature d r o p of the air stream a n d c o n d e n s e r superheat
is m i n i m u m . F i g u r e 6 illustrates the COP, overall gliding
temperature difference in both e v a p o r a t o r s (T8 - TT)
a n d the fraction of c o n d e n s e r superheat as a function of
c o m p o s i t i o n for R 2 2 - R I 2 3 for b o t h cases. It is interesting to observe that the m a x i m u m C O P indeed occurs
when the G T D is close to 23C, the overall t e m p e r a t u r e
drop of the air stream in the evaporator, with a m i n i m u m

Rev. Int. Froid 1991 V o 1 1 4 Septembre

259

Simulation of a two-evaporator refrigerator." D. S. Jung and R. Radermacher

Table 3 Simulation results for various mixtures for Case 1. Qk, = 67%, Qht~= 33%
Tableau 3 Rksultats de simulation pour plusieurs mklanges pour le eas 1. Q~,,, = 67%, Q~,,. = 33%
Mixture

COP,,,,,

R22 R 152a
R22 R142b
R22 R124
R22 R123
R32 R152a
R32 R142b
R32-- R 124
R125 R152a
R125 R142b
R125 R124
R 134a R 123
R 143a R 142b
R143a R124
R143a R123
R 152a R 123

1.426
1.474
1.443
1.527
1.464
1.512
1.482
1.415
1.455
1.402
1.465
1.460
1.412
1.428
1.494

qbo,.,

6.00
9.60
7.29
13.53
8.85
12.41
10.18
5.20
8.18
4.24
8.92
8.55
4.98
6.17
11.07

VC . . . . .

~(~,~

CO P,~

1007
822
934
1039
1290
1349
1459
902
652
742
507
700
800
1156
487

0.6
0.5
0.5
0.8
0.4
0.5
0.4
0.3
0.3
0.3
0.7
0.3
0.3
0.8
0.6

1.41
1.48
1.43
1.50
1.41
1.49
1.44
1.41
1.45
1.4
1.42
1.46
1.41
1.4
1.40

dp~,
4.61
9.51
6.47
11.52
4.83
10.78
7.43
4.83
7.80
4.10
5.57
8.32
4.83
4.08
4.24

X,~
0.10
0.45
0.30
0.65
0.03
0.17
0.10
0.10
0.40
0.33
0.95
0.35
0.25
0.65
1.00

Percentage increase in COP, ~, is based on the COP of RI2 obtained with a conventional single-evaporator refrigerator (COP,2s = 1.345, VC,:s = 769
kJ m % X0,,, in Tables 1 and 2 is the overall composition at which the maximum COP occurs while X~ is the overall composition at which the
volumetric capacity of the mixture is the same as that of RI2 with a single evaporator
L'augmentation en pourcentage du COP, dp, est fondke sur le C O P du R12 obtenu avec un rkJHg~rateur classique h ~vaporateur unique (COP;.,, = 1,345,
VC;:~ = 769 k J m ~). X,,,,, dans les tableaux 1 et 2 est la composition globale ?t laquelle on obtient le C O P maximal, alors que X~, est la composition
globale ?t laquelle la puissance volumktrique du mklange est la m~me que celle du R I 2 avee un seul kvaporateur

Table 4 Simulation results for various mixtures for Case 2. Q~,~= 50%, Qh,~= 50%
Tableau 4 Rbsultats de simulation pour plusieurs mblanges pour le cas 2. Qt,,. = 50%, Qh,,, = 50%
Mixture

COPma~

R22 R 152a
R22-R142b
R22- R124
R22 R123
R23 R124
R23 R152a
R32 RI52a
R32 R142b
R32 R124
R 125 R 152a
R125 R142b
R 125 R 124
R134a R123
Rl43a R142b
R143a R124
R143a R123
R152a R123

1.466
1.525
1.49
1.592
1.446
1.514
1.516
1.585
1.551
1.454
1.502
1.447
1.536
1.51
1.461
1.483
1.561

~b~,

VC ....

X ....

COP~

~Pv~

X~

9.00
13.40
10.78
18.36
7.50
12.60
12.71
17.84
15.31
8.10
11.67
7.58
14.20
12.27
8.62
10.26
16.05

1075
894
1010
945
845
1321
1555
1333
1574
965
708
806
480
762
997
1264
549

0.6
0.5
0.5
0.7
0.1
0.2
0.5
0.4
0.4
0.3
0.3
0.3
0.6
0.3
0.4
0.8
0.6

1.437
1.520
1.465
1.573

6.84
13.00
8.92
16.95

0.00
0.35
0.25
0.60

1.437
1.560
1.490
1.437
1.502
1.444
1.469
1.510
1.452
1.455
1.483

6.84
15.98
10.78
6.84
11.67
7.36
9.21
12.27
7.95
8.17
10.26

0.00
0.15
0.09
0.00
0.35
0.25
0.90
0.30
0.20
0.55
0.92

" The critical temperature of R23 is close to ambient temperature. Hence only mixtures at low concentrations of R23 are allowed to be substituted
,, La temp~;rature critique du R23 est voisine de la temperature ambiante. Aussi, seuls les m~;langes ~.['aibles concentrations de R23 peuvent constituer des
substituts

condenser superheat, thus confirming the previous finding.

Figure 7 illustrates the fraction of high temperature
evaporator area, AR, for an R 2 2 - R I 2 3 mixture under
two compartment loading conditions. It is seen that the
AR is a strong function of composition. This is an
important factor, especially in designing evaporators. An
area fraction which is good for R I2 may not necessarily
be valid for other fluids. Design engineers and experimentalists should be aware of this in advance in order
not to make an unfair comparison of the results obtained
with various fluids using the same evaporators.
It is interesting to note that the maximum increase in
COP occurs near the compositions at which the area
fraction, AR, is close to the fraction o f Qhtc. For instance,
for Case i the fraction of Qhte is 0.33 and the maximum
COP occurs at 0.7-0.8 mass fraction R22 where the

260

Int. J. Refrig. 1991 Vol 14 September

fraction of the high temperature evaporator area is 0.3. A

similar trend has been observed for Case 2 and with
other fluid mixtures.
The AR is a strong function of working fluid, overall
composition and the fraction of Qhte. The most critical
parameter in determining AR is the fraction of Qhtc. A
computer model may be needed to calculate the amount
of heat leaks to both compartments, fan power and
defrost heater power. Based on this calculation, the
amount o f cooling needed in both compartments can be
better estimated. This, of course, would depend on the
configuration (top mounted or side by side) and insulation material of a particular refrigerator.
There was a concern that the LTHX may not be
needed ~z. To resolve this issue, a further study of the
effect of the LTHX size was conducted. The U A of the
LTHX was fixed at 7% of the total U A of both evapora-

Simulation of a two-evaporator refrigerator: D. S. Jung and R. Radermacher

1.6 I

47.51 *C )

~~---o

R22/R123

-~

1.4 ~
~ 1.2-

0' 1.0Increasing

NBP ( * C )

QLTE= 67%

"~ 0.8"

COP

QHTE = 33%

COND. S U P E R H E A T

~0.60.4-

641.7 43.1 40.4 36.9 16.8 3.3 "1241-13.2

[ ]
66.8 49.1 146.4 43.0 22,9 i 9,3

6.0

48.4 28.3 14,8 11.5 5.4 RI52a -24.6

73.6 55.9 ~ ,

49.8 29.7 16.1 12.8 6.8

30"

i.26,0

171.0 50.9

oC3

37 ?.,( .....
14.9 I1,1 R22

t)5.9 78.3 75.6 72.1 52.0 ~

20"

1.4 ~134~

77.3 59.7 56.9 53.5 33.4 19.9 16.6 10.5 5.1

94.8 77.1

Ozone depleting
substances

RI34 -19.2

72,2 54.5 ~ /

e~

~ o

.40.9

35.2 29.2 23.8 22.4 18.6 7.5

01,,80

105

87.6 84.8 81.4 61.3 47.8 44,5 38.4 33.O

755

3,5

1.2

1.6

2,............ ,313"1251.3:.5.;
Y
31.6 27.9 16.8 10.4

.2

105 85.1 71.6 68.3 62.2 56.8 55.5 51.7 40.6 34.3 33.11

,3 . . . . .

,0,

3 .....

20,

.48.4

129 111 108

.5 . . . . .

10'

34.0 28.0 22.6 21.2 17.,I 6.3 ]143; -47.2

....

109

0.2
40'

56.2 52.4 ....

6.0

1.4 "~

311 .37.7
4__
-81.:

34.9 33.7 t0.~ I 24.5 II ~1.7 R23

/-

1.2"
~:Y 1.0"

COP

"~ 0.8

Figure 5 Difference in normal boiling points for various combinations

COND. SUPERHEAT

QLTE = 50%
QHTE = 50%

0.6

(updated from the table by Bare ~2)

Figure 5 D ~ r e n c e dam les points d'bbullition normale pour plusieurs
combinaisons de m~langes (m&e ~ jour avec le tableau de Bare 12

04-

tors (20 W C -~) for all cases investigated. With this

assumption, the temperature glide of mixtures across the
L T H X was up to 4C observed with R22-R12Y
Figure 8 shows the COP and pressure ratio across the
compressor as a function of L T H X size for the 50%
R22-50% R123 mixture As the size of the L T H X
increases, the COP also increases while the pressure ratio
across the compressor decreases. In this analysis, the size
of the L T H X was increased until the temperature of the
liquid at the end of the LTHX, T~2,equalled the temperature of the two-phase mixture entering the LTHX, Ti0
(pinch point). This corresponds to the maximum heat
exchange between the two streams occurring, with the
L T H X size roughly 20% of the total evaporator UA.
Beyond 20%, however, the performance was not affected, as shown by the broken lines in Figure 8.
Figure 9 shows the refrigerant and air stream temperatures in both evaporators with different L T H X sizes
With an increase in the size of the LTHX, the refrigerant
temperatures were adjusted to better match the air
stream temperatures, resulting in a reduction of compressor work. The quality of the two-phase mixture
entering the LTE decreases, while the refrigerant temperature entering LTE, TT, is almost constant, as dictated by
the air stream temperature, TST.
For a given mixture at X=Xo, a temperature and
composition diagram (Figure 10) shows that with a
decrease in quality and a constant temperature after the
expansion valve, the pressure in the evaporator has to be
increased Consequently, the pressure ratio across the
compressor is reduced, as shown in Figure 8.
Some remarks have to be made about the use of mixtures with extremely large vapour pressure difference
such as R32-R123, R32-R141b and R22-RI41b (note
that the difference in NBPs shown in Figure 5 may not
necessarily be the same at the condenser pressure). One
problem was encountered with these mixtures on the
condenser side (pinch point problem). As the gliding
temperature in the condenser was too high with these

o20-

~_.~,_._o.--~ ~ ----~--'~

0.2

lO[...,
o"
~1o -

~
0.0

0.1

0.2

GTDE

QLTE = 5 0 ~ %
QHTE = 50%

0.3

0.4

0.5

0.6

0.7

0.8

09

1.0

COMPOSITION (MASS FRACTION R22)

Figure 6 COP, gliding temperature difference in evaporator (GTDE)

and condenser superheat (QC~ur) for R22 R123 mixture as a function
of composition under two different conditions
Figure 6 COP, diff&ence de tempbrature de glissement dans I'bvaporateur ( GTDE) et surchauffe au condenseur ( QC,.p) pour le mblange R22R123 en Jonction de sa composition pour deux conditions diffbrentes de
Jbnctionnement

mixtures the refrigerant leaving the condenser (point 6 in

Figure 2) is no longer subcooled but in the two-phase
region. What happens next is that the refrigerant temperature at point 5 (saturated liquid) is slightly below the air
temperature to satisfy the energy balance, which represents a physically impossible solution.
This finding indicates that in order to use these mixtures in actual refrigerators, either the condenser size or
the mass flow-rate of air in the condenser should be
reduced. By doing this, the condenser pressure is raised
so that subcooled refrigerant is present at the outlet.
This, of course, would degrade the system performance
as the compressor work might be increased with a higher
pressure ratio. Thus, care should be taken in the use of
mixtures with a large G T D exceeding 70C in the difference in the NBPs shown in Figure 5.
Finally, the shortcomings of the current simulation
will be discussed. In the heat transfer formulation, both
air stream and refrigerant temperatures are assumed to
be linear (A in Figure 9). In actual heat exchangers with
an assumed linear air temperature distribution, the refrigerant temperature may vary irregularly (B,C,D in
Figure 9) because of the non-linear slope, dT/dH, in the
Rev. Int. Froid 1991 Vol 14 Septembre

261

Simulation of a two-evaporator refrigerator: D. S. Jung and R. Radermacher

70

10
R 2 2 / R 123
QLTE = 67%, QHTE = 33%
PRI2= 1173
AR12 = 13.93

60
5o

5 0 % R 2 2 / 5 0 % R 123
LTHX size: 0% of evaporator UA
----o-~

PR
AR

o /

o~ 0
-10

~: 30
<
,.
<
r~
Z
<
O

t~

20
-30
10
-40
10

0
7o
QLTE = 50%, QHTE= 50%
PRI2= 11.17
ARI2 = 221

-----O-+

LTHX size: 21% of evaporator UA

PR
AR

G" 0

so
~-10

,0
3o

-20
-30
"~.-----o___._.o____o___._. o

~'/;L-+~i~

~ - - - ~ " ~ "
L'~

Temperatureglide
.

A : Ideal
B,C, D : Actual

~X
D

-40
0.0

0.1

02

()3

O4

O5

0.6

0.7

()8

I)9

20

10

40

60

80

100

12o

QUALITY (%)

COMPOSITION (MASS FRACTION R22)

Figure 7

Evaporator area distribution (AR) and pressure ratio (PR)

for R22-RI23 mixture under two different conditions
F i g u r e 7 Distribution de l'bvaporation ( A R ) et taux de compression

(PR) pour le mblange R22-RI23 dans deux conditions d_i.ffbrentes de

.lonctionnement

Figure 9

R e f r i g e r a n t a n d a i r s t r e a m t e m p e r a t u r e s in t h e f r e e z e r a n d
f o o d c o m p a r t m e n t e v a p o r a t o r s w i t h d i f f e r e n t sizes o f l o w t e m p e r a t u r e
heat exchanger
Figure 9
Tempbratures du.[~igorigbne et du courant d'air aux ~vapora-

teurs du congblateur et du compartiment pour aliments, avec diffbrentes

dimensions de I'dchangeur de chaleur ~t basse tempdrature

1.6050%R22/50%R123

1.55
r...)
1.50

Pinch Point

1.45

Pinch Point

7
14
21
UA OF LTHX (% OF UA OF EVAPORATOR)

28

Figure 8

COP and pressure ratio (PR) across the compressor for 50%
R 2 2 5 0 % R 1 2 3 m i x t u r e a s a f u n c t i o n o f size o f a l o w t e m p e r a t u r e h e a t
exchanger
F i g u r e 8 C O P et taux de compression (PR) au compresseur pour un

m~lange de 50 % de R22 et de 50 % de R123 en fonction de la dimension

de I'~changeur de chaleur ~t basse temperature

two-phase region. This can be handled by dividing the

heat exchangers into many segments and calculating the
average LMTDs, which would require a much longer
execution time.
The overall heat transfer coefficients of all heat
exchangers are assumed to be the same. This, of course,
is not true in actual systems. The overall heat transfer
coefficient (U) also varies from the subcooled liquid to
superheated vapour regions. To obtain more accurate
results, heat transfer data in all regions of the heat

262

Int. J. Refrig. 1991 Vo114 September

1.0
COMPOSITION

Figure 10

R i s e in e v a p o r a t o r p r e s s u r e w i t h i n c r e a s e d l o w t e m p e r a t u r e
h e a t e x c h a n g e r size
F i g u r e 10 Elbvation de la pression dans l'~vaporateur avec augmen-

tation de la dimension de l'bchangeur de chaleur ?t basse tempbrature

exchangers may be necessary. The compressor was

modelled in the simplest way by specifying the isentropic
efficiency. A more elaborate compressor model would
yield more realistic results.
Overall, despite the above shortcomings, the present
model would still yield a fair relative comparison and
may serve at least as a preliminary screening tool in
designing highly efficient refrigerators.

Simulation of a two-evaporator refrigerator: D. S. Jung and R. Radermacher

Finally, the use of N A R M s in two-evaporator refrigerators may serve as an alternative to help solve the
present environmental problem by employing ozonesafe fluids on a highly energy efficient cycle. The
success depends largely on the proper selection of
fluids, the accurate determination of the load distribution, evaporator area ratio and L T H X size.

Conclusions
Based on the computer simulation of steady-state twoevaporator refrigerators operating on Lorenz and
Meutzner's 4 cycle, the following conclusions can be
drawn.
1

By adding another evaporator and a low temperature

heat exchanger to a conventional single evaporator
refrigerator, the simulation becomes complicated,
such that the successive substitution method is not
successful. For the simulation of more complicated
systems, the Newton-Raphson method solving
simultaneously a set of non-linear equations is suggested due to its global systematic treatment of all
variables.
2 The results obtained with pure fluids indicate that
having a portion of evaporation at higher temperatures helps to increase the COP and VC up to 6 and
15.5%, respectively, as compared to that of a single
evaporator unit. The COP would increase with a
larger fraction of the high temperature cooling load,
Qht~. Owing to the mismatch of volumetric capacity,
no single fluid seems to replace RI2 directly unless
major modifications are made in the compressor.
3 A significant increase in COP of up to 15 18% is
achieved with R22-RI23 and R32 R142b mixtures,
demonstrating the capability of Lorenz and
Meutzner's 4 cycle. This was achieved mainly due to
the good matching of the large overall temperature
drop in the air streams in the two evaporators (23C)
against the overall refrigerant side temperature glide.
The maximum COP seems to exist at the concentration where the G T D in the evaporator is close to
the overall temperature drop of air and the condenser
superheat is at a minimum.
4 The area ratio between the high and low temperature
evaporators is a strong function of the load distribution, working fluids and composition. To design
highly efficient refrigerators, a proper cabinet heat
leak model may be needed to determine the load
distribution accurately. Based on this calculation, a
suitable mixture may be selected and the current
simulation model would yield the required area ratio
to fulfil the cooling duty in the most efficient way.
5 The effect of the low temperature heat exchanger is
examined by varying its size. With an increased size
of LTHX, the COP also increases while the pressure
ratio across the compressor decreases. The maximum
size of L T H X is limited to roughly 20% of total
evaporator UA. Beyond this, no change in performance would be expected.

Acknowledgements
The support for this work by Pennwalt Corporation, the
US Environmental Protection Agency and the University of Maryland, College of Engineering, is gratefully
acknowledged. The authors thank M. McLinden and (3.
Morrison at the US National Institute of Standards and
Technology for helpful discussions regarding the thermodynamic properties of refrigerant mixtures.

References
1
2
3
4

6
7
8
9
10

I1
12

Montreal Protocol on Substances that Deplete the Ozone

Layer, Final Act, United Nations Environment Programme,
Nairobi (1987)
,lung, D. S., Radermaeher, R. Performance simulation of singleevaporator domestic refrigerators charged with pure and mixed
refrigerants lnt J Re/?ig (1991) 14 223 232
Hell'man, ,l. Division of Global Change, US Environmental
Protection Agency, Washington, DC, personal communication
(August 1989)
Lorenz, A., Meutzner, K. On application of non-azeotropic two
component refrigerants in domestic refrigerators and home
freezers X1V hlternational Congress o[" Re/Figeration, Moscow,
1975, llR, Paris, (1975)
Stoecker, W. F. Improving the energy effectiveness of domestic
refrigerators by the application of refrigerant mixtures ORNL
Report. ORNL/Sub 78/55463/1 (1978)
A.D. Little, Inc. An evaluation of a two-evaporator refrigerator freezer using nonazeotropic refrigerant mixtures ORNL
Report, ORNL/SuI~82/47952/I (1982)
Kruse, H. ASHRAE Meeting, Chicago, IL, personal communication (January 1989)
Stoecker, W. F., Walukas, D. J. Conserving energy in domestic
refrigerators through the use of refrigerant mixtures lnt J Refrig
(1981)4201 208
Kruse, H. The advantages of non-azeotropic refrigerant mixtures for heat pump application lnt J ReJrig (1981) 4 119-125
Morrison, G., MeLinden, M. O. Application of a hard sphere
equation of state to refrigerants and refrigerant mixtures NBS
Technical Note 1226, National Bureau of Standards, Gaithersburg, MD (1986)
ASHRAE Handbook, Equipment Volume, ASHRAE, Atlanta,
GA (1988)
Bare, J. Air Toxics Research Division, US Environmental Protection Agency, Washington, DC, personal communication
(August 1989)

Rev. Int. Froid 1991 Vo114 Septembre

263