Energy Conversion and Management 42 (2001) 1607±1621

www.elsevier.com/locate/enconman

Performance analysis on a multi-type inverter air conditioner

Youn Cheol Park *, Young Chul Kim, Man-Ki Min
Department of Mechanical Engineering, Korea University, Anam-Dong, Sungbuk-Gu, Seoul 136-701, South Korea
Received 31 May 2000; accepted 27 October 2000

Abstract
An analysis was conducted for a multi-type inverter air conditioner with a linear electronic valve as the
expansion device and a variable speed compressor. The system performance was analyzed with variations
of operating frequency of the compressor, cooling load imposed on the system and cooling load fraction
(i.e. load ratio) between rooms in which is installed an evaporator. The optimum opening of the electric
expansion valve (EEV) was calculated when the compressor operating frequency was speci®ed at a given
cooling load. As compressor operating frequency increased with cooling load increment, the EEV should
have adjusted to get wide openings to get an optimum COP of the system. While total cooling load of the
system was constant, the cooling load fraction changed due to the cooling load di€erences between each
room in the multi-type air conditioning system with a number of evaporators and EEVs. The operating
frequency of the compressor was increased with increment of the load ratio, and consequently, the power
consumption of the compressor increased. The increment of the load ratio, which means increasing the load
di€erence between each room, causes a reduction of the system performance (COP), although the total
cooling capacity was constant. Ó 2001 Elsevier Science Ltd. All rights reserved.

Keywords: Multi-type air conditioner; Variable speed compressor; Electric expansion valve; Load ratio

1. Introduction

Air conditioners are the necessities of life at home and in public areas due to the large demand
for comfort in the thermal environment of living space in modern society. The conception of air
conditioning has now developed from one air conditioner for one house to independent air
conditioning for separate rooms. A multi-type air conditioner is the system that could distribute

*
Corresponding author. Tel.: +82-31-200-6244; fax: +82-31-200-6252.
E-mail address: younpark@samsung.co.kr (Y.C. Park).

0196-8904/01/$ - see front matter Ó 2001 Elsevier Science Ltd. All rights reserved.
PII: S 0 1 9 6 - 8 9 0 4 ( 0 0 ) 0 0 1 4 7 - 3
1608 Y.C. Park et al. / Energy Conversion and Management 42 (2001) 1607±1621

Nomenclature

A area (m)
CD coecient of discharge for ori®ce equation
D diameter (m)
Fsh superheat correction factor
Fv volumetric eciency correction factor
g gravity (m/s2 )
h enthalpy (kJ/kg)
m_ mass ¯ow rate (kg/s, lb m/h)
P pressure (Pa)
T temperature (K)
v speci®c volume (m3 /kg)
W_ compressor power (kW)
Greeks
q density (kg/m3 )
c speci®c weight (kgf/m3 )
Subscripts
act actual
c condenser
e evaporator
map map data
i inlet
isen isentropic
r refrigerant
sat saturation
o outlet

cooling capacity to di€erent spaces. The system consists of a number of indoor units and only one
outdoor unit. The electric expansion valves (EEVs) control the capacity of the system with varying
refrigerant mass ¯ow rates passing through the evaporators according to the cooling loads of each
evaporator. Fig. 1 shows a schematic diagram of the multi-type air conditioner with two evap-
orators.
The load of the system is widely changed due to the system having a number of evaporator with
only one outdoor unit. The performances of the compressor and the expansion device are a€ected
by those load variations. A variable speed compressor was used in order to cover the load
variation. The system with the variable speed compressor can control the cooling capacity by
changing its operating frequency. However, there should be limits of control when the system uses
constant area expansion devices, such as a capillary tube or a short tube ori®ce. Variable area
expansion devices (like an EEV) are needed in multi-type inverter air conditioners to control the
mass ¯ow rate of the refrigerant concerning the variable speed compressor.
 Y.C. Park et al. / Energy Conversion and Management 42 (2001) 1607±1621 1609

Fig. 1. Multi-type inverter air conditioner with two evaporators and EEVs.

The research trend of the multi-type air conditioner may be divided as follows: a control logic
of the variable speed compressor; an optimum distribution of the refrigerant to the evaporator
with cooling loads; a method of seasonal energy eciency ratio calculation; a control method of
the EEV etc. As one of the researches on the multi-type inverter air conditioner, Fujita et al. [1]
studied the capacity and refrigerant ¯ow rate control in the multi-type air conditioner with EEV.
He improved the comfort, reliability and working eciency of the system by use of control
technology with the EEV and a twin rotary compressor. Okuzawa [2] optimized the superheat
temperature in a multi-type system by use of PI control logic and improved the response to a
variation of load. He used a horizontally installed twin rotary compressor and four to six
evaporators. For a study on the refrigerant ¯ow rate control, Krakow et al. [3] calculated the e€ect
of refrigerant ¯ow rate and capacity control method on the performance of a heat pump and
shows 5% energy saving with those methods. He changed the opening of an ori®ce tube and the
refrigerant charge as a ¯ow rate control method and used a variable speed compressor as a ca-
pacity control method.
A simulation model for a single unit air conditioning system with a variable speed compressor
and an EEV was developed in this study. Performance analyses were conducted with variation of
operating frequency of the compressor and openings of the EEV. In addition, analysis on the
multi-type system that has a number of evaporators and EEVs was performed to determine the
1610 Y.C. Park et al. / Energy Conversion and Management 42 (2001) 1607±1621

optimum operating frequency of the compressor and openings of the EEV. The object of this
study is to study system performance variations with cooling load in an inverter air conditioner
with EEV and with the load ratio of each evaporator of the multi-type inverter air conditioning
system.

2. Modeling

The refrigerant ¯ow rate and pressure drop in each evaporator is di€erent in the multi-type air
conditioner even if the con®guration of the system is similar to that of a single unit system. The
exit pressure of each evaporator is assumed equal with di€erent inlet pressure because the exit is
united into one stream and goes to the compressor inlet. In addition, the valve openings are
assumed as if the refrigerant ¯ow rates in each evaporator ®t the cooling capacity of the evap-
orator in this study.
The variables of the study are operating frequency of the variable speed compressor, opening of
the EEV, cooling load of each room and inlet air temperature of the condenser and evaporators.
Initial values are the saturation temperatures of the evaporator and condenser and the superheat
after the evaporator. The ¯ow chart of the developed model is shown in Fig. 2.
The model is dependent on the saturation temperature of the condenser inlet, evaporator exit
temperature and valve opening of the EEV. The initial temperature of the condenser inlet and
evaporator outlet calculates the mass ¯ow rate of the compressor, and it is compared with the
mass ¯ow rate through the EEV, which needs to match the input capacity of the evaporator from
which the EEV openings and mass ¯ow rates are computed at a speci®ed evaporator exit tem-
perature. Finally, the optimum COP of the system is calculated by changing the evaporator outlet
temperature.

2.1. Compressor model

The compressor map data used in the compressor model was given by a manufacturer. The map
data was determined in a compressor calorimeter with variation of the saturation temperatures
of the condenser and evaporator and speci®ed superheat and sub-cooling. The compressor used
in this study is a rolling piston type rotary compressor with 36.55 cm3 piston displacement.
Compressor power and refrigerant mass ¯ow rate are calculated by Eq. (1). The equation is a
second-order function of the saturation temperatures and its coecients are shown in Table 1.
Interpolation and extrapolation was used for the operating frequencies that were not given by the
equation.
m_ map or W_ map ˆ f …C1 Tc2 ‡ C2 Tc ‡ C3 Te2 ‡ C4 Te ‡ C5 Tc Te ‡ C6 † …1†
Dabiri and Rice [4] de®ned a correlation for the superheat correction with a volumetric e-
ciency correction factor as Eqs. (2) and (3) because the map data has been taken at speci®ed
superheat, which did not consider the superheat e€ect on the ¯ow rate.
m_ act ˆ ‰1 ‡ Fv …vmap =vact 1†Šm_ map …2†

W_ act ˆ …m_ act =m_ map †…Dhisen;act =Dhisen;map †W_ map …3†
 Y.C. Park et al. / Energy Conversion and Management 42 (2001) 1607±1621 1611

Fig. 2. Flow chart of the multi-type air conditioner model.

Table 1
Coecients of the variable speed compressor performance equation (Eq. (1))
Operating fre- C1 C2 C3 C4 C5 C6
quencies (Hz)
Power (kW) 45 1.2851E 5 1.2676E 2 7.4858E 6 2.5590E 2 2.2152E 4 2.4971E 1
60 1.7733E 5 1.7325E 2 1.0163E 5 3.4840E 2 3.0158E 4 3.3613E 1
75 5.3555E 5 2.9191E 2 4.0193E 6 4.9882E 2 4.2177E 4 1.2674E 1
90 2.9479E 5 2.8978E 2 1.7093E 5 5.8427E 2 5.0576E 4 5.6835E 1
Mass ¯ow 45 5.1989E 6 1.1967E 1 2.9741E 2 3.6733E‡0 9.4398E 3 9.7148E‡1
rate (kg/s) 60 3.7356E 6 1.6511E 1 4.0854E 2 5.0472E‡0 1.2971E 2 1.3349E‡2
75 3.8329E 7 2.1439E 1 5.2631E 2 6.5002E‡0 1.6701E 2 1.7210E‡2
90 4.2674E 7 2.6268E 1 6.4561E 2 7.9745E‡0 2.0489E 2 2.1109E‡2
1612 Y.C. Park et al. / Energy Conversion and Management 42 (2001) 1607±1621

Low temperature refrigerant goes through the suction tube to the compressor shell in which it is
heated by the compression and motor. Dabiri and Rice assumed that the enthalpy increment by
heating was constant during the suction process. However, the suction gas temperature varies with
the evaporating and condensing temperatures. Eq. (4) was used to consider the e€ect of the
saturation temperature variation on the suction gas heating.
Dhin;suc ˆ Fsh …W_ map =m_ map † …4†

2.2. Heat exchanger model

The condenser and evaporator of the system are ®n tube type heat exchangers. They consist of
equivalent parallel refrigerant circuits. The mass ¯ow rate that is calculated from the compressor
model divided by the number of tube circuits with the assumption that each circuit has the same
conditions. Di€erent heat transfer correlations were applied to appropriate refrigerant conditions,
such as superheat, sub-cooling and the two phase state. The performance of the heat exchanger
was calculated with the e€ectiveness-NTU method. The speci®cations of the heat exchanger are
shown in Table 2.

2.3. Expansion device model

There are two types of expansion devices used in the vapor compression refrigeration system.
One is with a constant ¯ow area and the other is a variable area expansion device. In this study,
both expansion devices are considered for veri®cation of the developed model. A short tube ori®ce
was used as the constant area expansion device, and the EEV was used as the variable area ex-
pansion valve. The refrigerant ¯ow rate correlation for the ori®ce tube is de®ned by Mei [5] as Eq.
(5) as a function of sub-cooling and pressure di€erence through the ori®ce tube.
(
1=2
pD2 …0:68683 0:019337 DPe;i ‡ 0:006 DT †…2q gDPe;i †1=2 =4; DT 6 40°F
m_ ˆ …5†
pD2 …0:9175 0:00325 DT †…2q g…P Psat †1=2 †=4; DT > 40°F
The EEV changes the refrigerant ¯ow rate by activating a stepping motor as shown in Fig. 3.
The opening con®guration of the EEV was di€erent from manufacturer to manufacturer. How-

Table 2
Dimensions of the heat exchangers
Parameters Condenser Evaporator
2
Frontal area (m ) 0.84 0.36
Number of tubes 3 4
Number of circuits 3 3
Row pitch (mm) 27.432 21
Step pitch (mm) 31.75 25.4
Fin pitch (FPI) 13 14
Fin thickness (mm) 0.1524 0.127
Outer diameter, Do (mm) 10.033 10.06
Inner diameter, Di (mm) 9.42 9.42
 Y.C. Park et al. / Energy Conversion and Management 42 (2001) 1607±1621 1613

Fig. 3. Schematic of the EEV.

ever, the correlation for the ori®ce tube also could be used with every EEV. Eq. (6) is the ¯ow rate
correlations for the EEV which is taken from the performance data of the EEV (Japanese As-
sociation of Refrigeration [6]).
1=2
m_ ˆ 5470  CD  A  …Pi Po † =3600 …6†

3. Results and discussion

The environment of the system was ®xed at a speci®ed temperature and humidity as in
ASHRAE Standard 116 [7]. The supply air temperature to the evaporator was 27°C with 50%
relative humidity and 35°C with 40% relative humidity to the condenser. There are a number of
studies on multi-type air conditioners, however, performance data for comparison are not
available yet, such as system performance variation with operating frequency including system
speci®cations. Major subjects of the published studies were the refrigerant distribution or per-
formance of the system components. Because of insucient available data, an indirect method
was used to verify the developed model. The performance of the model was compared with the
well known ORNL program [8] in the case of variation of the operating frequency of the com-
pressor. The maximum COP deviation was only 3.3% within 30±90 Hz operating frequency of the
compressor compared to calculated results with the ORNL program.

3.1. Constant cooling load of an inverter air conditioner with electric expansion valve

The performance of an inverter air conditioning system with EEV was analyzed in this section
prior to the multi-type system analysis. Even though there are many studies on the performance of
1614 Y.C. Park et al. / Energy Conversion and Management 42 (2001) 1607±1621

Fig. 4. COP variation with EEV opening area at a ®xed frequency and cooling load.

an air conditioner with variable speed compressor [9±11], the studies of the e€ect of the EEV on
the inverter air conditioner are limited in the open literature.
The system performances were analyzed with variation of the evaporating temperature at an 8.7
kW cooling load of the system and with the assumption that the systemÕs cooling capacity mat-
ched the cooling load. Fig. 4 shows the COP variation with opening area of the EEV at a 60 Hz
compressor frequency and 8.7 kW cooling load condition. The trend of the COP variation shows
a parabolic type with opening area of the EEV, and there is an optimum opening with respect to
COP. The refrigerant mass ¯ow rate, evaporating temperature and superheat increased with the
EEV opening area. However, with the compressor operating frequency constant, the condensing
temperature goes down, and the sub-cooling decreases with variation of EEV opening area. The
frequency of the compressor was changed ®rstly to match the system cooling capacity to the
speci®ed cooling load. By changing the frequency, the mass ¯ow rate of the system varies, and
the opening of the EEV should also have to be changed. Consequently, it is possible to get a
maximum COP by variation of the openings of the EEV at a given compressor speed.
There are limitations of the compressor operating frequency variation for the inverter air
conditioning system to cover the speci®ed cooling load to the system. When the compressor runs
at low frequency, the mass ¯ow rate discharged from the compressor is small. As a result, the
systemÕs cooling capacity is lower than the speci®ed cooling load of the system, even after in-
creasing the opening area of the EEV. On the contrary, if the mass ¯ow rate is increased by high
frequency operation, the systemÕs cooling capacity is higher than the speci®ed cooling load, even
after reducing the opening area of the EEV. Because the mass ¯ow rates are strongly dependent
on the compressor frequency, it can be noted that there are limits to the control range in com-
pressor frequency of the inverter air conditioner at a speci®ed cooling load.
Fig. 5 shows the variation of COP with evaporating temperature and operating frequencies in
the case of the system covering the given 8 kW cooling load. The operating frequency range of the
compressor in this case is from 53 to 60 Hz. The system could not provide the 8 kW cooling
capacity outside this range even when changing the opening area of the EEV due to the reasons
mentioned before. The evaporating temperature in Fig. 5 changes by variation of the opening of
the EEV. The high COP at lower frequency is mostly due to small compressor power consumption
compared to that at the higher frequencies. At 53 Hz, the variable range of the EEV opening is
 Y.C. Park et al. / Energy Conversion and Management 42 (2001) 1607±1621 1615

Fig. 5. COP variation with evaporating temperature and operating frequency at an 8 kW cooling load.

smaller than in the 57 and 60 Hz cases, and the COP was higher than in the other cases, as shown
in Fig. 5. When the evaporating temperature was increased at constant operating frequency, the
mass ¯ow rate and compressor power consumption increased also. This trend was also found in
the compressor calorimeter test which was given by the manufacturer. The COP of the system
decreases with the evaporating temperature as a result of the increment of the EEV opening in the
case of constant operating frequency.
Fig. 6 shows the COP and the compressor power consumption variation with compressor op-
erating frequency when the imposed cooling load on the system was constant at 8 kW. This graph
was determined by connecting the optimum COP points on Fig. 5 within the compressor operating
frequency range. The compressor power consumption increased and COP decreased with increase
of the compressor frequency. When the compressor operating frequency increases, the evaporating
temperature, decreases, and the superheat of the refrigerant before the compressor inlet increases,
which reduces the refrigerant mass ¯ow rate. The increase of the mass ¯ow rate by the compressor
operating frequency is diminished due to the reduction of the mass ¯ow rate by the suction gas
heating. The e€ect of suction gas heating on the refrigerant mass ¯ow rate, which determines the
speci®c volume before the compressor, is much higher than the e€ect of the compressor operating

Fig. 6. COP and compressor power consumption variation with compressor operating frequency at an 8 kW cooling
load.
1616 Y.C. Park et al. / Energy Conversion and Management 42 (2001) 1607±1621

frequency on the mass ¯ow rate. When the compressor frequency is increased about 7%, from 53 to
57 Hz, the speci®c volume before the compressor increases about 15%.
From the above results, it is recommended that the operating frequency selection in the inverter
air conditioner should be the lowest frequency within the variable range of the compressor op-
erating frequency, in which the system capacity matches the imposed cooling load. Although, the
system shows higher COP at the lowest frequency within this range, the variable range of the EEV
opening is very narrow in that case, as shown in Fig. 5. This operating frequency can be noted as
the critical operating frequency to get a maximum COP of the system, but it is not easy to control
the EEV openings correctly at this operating frequency with those small clearances.

3.2. Variation of cooling load of an inverter air conditioner with electric expansion valve

The cooling capacity is controlled by the compressor operating frequency and the openings of
the EEV in the inverter air conditioner with EEV. If the EEV openings are changed with fre-
quency, as with a constant area expansion valve, the system performance varies as shown in Fig.
7. This graph shows the COP variation with cooling load and compressor frequency. The opening
of the EEV for the ®gure is set to an optimum opening at rated compressor frequency (60 Hz).
The COP variation at 40 Hz does not show an optimum point compared to the 60 and 80 Hz cases
because the opening of the EEV was wide enough to allow the ¯ow of refrigerant that was dis-
charged from the compressor. On the contrary, the EEV opening at the 80 Hz case is too small to
accept the mass ¯ow rate coming out of the compressor. The restriction of the refrigerant from the
compressor by the EEV makes di€erences between the condensing and evaporating temperatures
and a€ects the system performance variation. The maximum capacity the system could provide, is
6, 8.7 and 10.5 kW at 40, 60 and 80 Hz, respectively. If the system needs more cooling capacity,
the system has to increase the compressor operating frequency. However, it could not get a higher
COP than before.
When the EEV opening changes, it is possible to get more capacity than in the ®xed opening
area cases. An optimum (maximum COP) EEV opening was calculated for a wide range of
cooling load of the system, as shown in Fig. 8 with compressor operating frequency. This graph

Fig. 7. COP variation with cooling load and compressor operating frequency at ®xed EEV opening.
 Y.C. Park et al. / Energy Conversion and Management 42 (2001) 1607±1621 1617

Fig. 8. COP variation with cooling load and compressor operating frequency at varied EEV opening.

Fig. 9. COP, power consumption and capacity variation of the system with compressor operating frequency.

can be well used by manufacturers as design parameters for control of the EEV and compressor
frequency with cooling load variation of the system.
Fig. 9 shows the compressor power, COP and cooling capacity of the system with the com-
pressor operating frequencies determined as from the Fig. 8. When the compressor frequency
increases, the compressor power increases with second-order variation of the frequency, and the
COP of the system goes down, as was the case in the general inverter air conditioner. Fig. 9 was
calculated at 26.7°C indoor and 35°C outdoor temperatures with 50% and 40% relative humidity,
respectively. Therefore, the performance variation with compressor frequency in the inverter air
conditioner with EEV is the same as in the general inverter air conditioner that has a thermal
expansion valve or ori®ce tube. This is only the case of the system operating at optimal opening of
the EEV at a given operating frequency of the compressor. In other words, if the COP, com-
pressor power and cooling capacity variation in the system using the EEV does not follow itÕs
variation in the general inverter air conditioner that uses a constant area expansion device, the
system is not optimized with the opening of the EEV at a given frequency. Therefore, this is a
good estimation parameter for control of the EEV to determine whether the system was well
tuned or not.
1618 Y.C. Park et al. / Energy Conversion and Management 42 (2001) 1607±1621

3.3. System characteristics of the multi-type air conditioner

Analysis is performed for the multi-type air conditioner with two evaporators and EEVs. A
multi-type air conditioner has a number of expansion devices and evaporators with one com-
pressor and condenser compared to a single unit system. Generally, the cooling load of the system
is determined by the dimensions of the cooling space and environment conditions. The cooling
loads of the evaporators are di€erent from room to room. The performance is analyzed on a
multi-type air conditioner with two evaporators with a variation of the load ratio. The load ratio
is de®ned as the fraction of the cooling load of a room to the total cooling load of all rooms. The
total cooling load of the system is ®xed at 6 kW in this study.
Fig. 10 shows the compressor power consumption and COP variation with load ratio. The
power consumption increases with an increase of the load di€erence between each room (i.e. load
ratio), and the COP decreases. The reason for the power consumption increase is due to the
compressor operating frequency increase. The operating frequency increases with load ratio as
Fig. 11 even when the total cooling load is constant. Because of the load di€erence between the
evaporators, one evaporator has to increase and the other has to reduce its mass ¯ow rate. The
system changes the EEV openings to control the refrigerant distribution to the evaporators.

Fig. 10. COP and compressor power consumption variation with load ratio.

Fig. 11. Operating frequency variation with load ratio.

 Y.C. Park et al. / Energy Conversion and Management 42 (2001) 1607±1621 1619

Fig. 12. EEV opening area variation with load ratio at each evaporator.

Fig. 13. Evaporating temperature variation with load ratio at 6 kW total cooling load.

However, the total refrigerant mass ¯ow rate of the system increases due to the increase of the
compressor operating frequency.
Fig. 12 shows the variation of the EEV opening area with load ratio. The mass ¯ow rate to an
evaporator that needs more cooling capacity increases. The evaporating temperature variation
with load ratio is shown in Fig. 13. The evaporating temperature di€erence between each room
increases with load ratio. When the evaporating temperature is reduced, the cooling load de-
creases which makes the COP of the system decrease, as in Fig. 10.
The operating frequency of the compressor at a 50% load ratio was 37 Hz in this study. If the
load ratio were changed to 100%, it means only one evaporator was used, and the operating
frequency goes up to 48 Hz. Correspondingly, the system COP and EEV opening areas are
changed from 4.5 to 3.7 and from 0.00521 to 0.01 cm2 , respectively. When the load ratio is
changed from 50% to 100%, the compressor frequency changes only 30% (from 37 to 48 Hz), but
the EEV opening of an evaporator changes about 92% (from 0.00521 to 0.01 cm2 ). It is possible to
note that the major control parameter is the EEV opening in a multi-type air conditioner rather
than the compressor operating frequency when the load ratio is changed.
The system performance variation with load ratio changes with di€erent total cooling load of
the system. Fig. 14 shows the compressor frequency variation with load ratio in such a case. When
the total cooling load was 8 kW, the compressor operates at high frequency compared to 6 kW. In
1620 Y.C. Park et al. / Energy Conversion and Management 42 (2001) 1607±1621

Fig. 14. Compressor operating frequency variation with load ratio at a di€erent cooling load in the system.

addition, the COP reduction was much higher in the 8 kW case because the variation of the
compressor operating frequency with load ratio was much higher than for 6 kW.

4. Conclusions

A system simulation program has been developed to analyze the performance of the multi-type
air conditioner with two evaporators and EEVs. A parametric study was conducted with variation
of the compressor operating frequency, the loads of the conditioned rooms, the opening area of
the EEV etc.
In an inverter air conditioner with EEV, there are limitations of the operating frequency of the
variable speed compressor to cover the given cooling load. The operating frequency selection in
the inverter air conditioner should be the lowest frequency to get the highest COP within the
variable range of the compressor operating frequency when the system capacity matches the
imposed cooling load on the system.
When the cooling load of the system changes, the COP variation at constant operating fre-
quency is of parabolic shape with cooling load by changing the EEV openings. If the cooling load
increases, the system has to increase the operating frequency because of the limitation of the
cooling capacity of the system at a ®xed operating frequency. Consequently, the maximum COP
at an operating frequency decreases with cooling load like as shown in Fig. 8 even after adjusting
the EEV opening.
For the multi-type air conditioner with two evaporators and EEVs, the COP of the system
decreases with an increase of the load ratio due to increasing the compressor operating frequency
as the load di€erence between the evaporators increases. The multi-type air conditioner mainly
controls the total capacity by compressor operating frequency. Although, if the cooling loads are
di€erent from room to room, the major control parameter of the system is the EEV opening for
changing the refrigerant mass ¯ow rate through the evaporators.

References

[1] Fujita Y, Kubo T, Suma S. Multi-air conditioner with two indoor units. Refrigeration 1992;67:171±6.
[2] Okuzawa Y. One system family air conditioning: Multi-system air conditioner for 4±6 rooms. Refrigeration
1992;67:149±56.
 Y.C. Park et al. / Energy Conversion and Management 42 (2001) 1607±1621 1621

[3] Krakow KI, Lin S, Matsuki K. A study of the primary e€ects of various means of refrigerant ¯ow control and
capacity control on the seasonal performance of a heat pump. ASHRAE Trans 1987;93:511±24.
[4] Dabiri AE, Rice CK. A compressor simulation method with corrections for the level of suction gas superheat.
ASHRAE Trans 1981;87:771±82.
[5] Mei VC. Short tube refrigerant ¯ow restrictors. ASHRAE Trans 1982;88:157±68.
[6] Refrigeration and air conditioning handbook (application volume). vol. 4. Japanese Association of Refrigeration
1981. p. 611±3.
[7] ASHRAE Standard ANSI/ASHRAE 116. Method of testing for seasonal eciency of unitary air conditioners and
heat pumps. ASHRAE, New York, 1995.
[8] Fischer SK, Rice CK. The Oak Ridge heat pump models: I. A steady state computer design model for air-to-air
heat pumps. ORNL/CON-80/R1, 1983.
[9] Mitsunaga T, Nonaka R, Matsumoto K. Rotary compressor for 180 Hz inverter air conditioner. SANYO Tech
Rev 1987;19:102±12.
[10] Miller WA. Laboratory eciency comparisons of modulating heat pump component using adjustable speed drives.
ASHRAE Trans 1988;94:874±91.
[11] Hayashi M, Nariai S. Development of scroll inverter air conditioners. Refrigeration 1992;67:51±7.