Energy Conversion and Management 51 (2010) 277287

Contents lists available at ScienceDirect

Energy Conversion and Management

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

Transient analysis of a variable speed rotary compressor

Youn Cheol Park *
Department of Mechanical Engineering, Cheju National University, Ara-Dong, Jeju-City, Jeju-Do 690-756, Republic of Korea

a r t i c l e i n f o a b s t r a c t

Article history: A transient simulation model of a rolling piston type rotary compressor is developed to predict the
Received 8 September 2007 dynamic characteristics of a variable speed compressor. The model is based on the principles of conser-
Received in revised form 10 February 2009 vation, real gas equations, kinematics of the crankshaft and roller, mass ow loss due to leakage, and heat
Accepted 22 September 2009
transfer. For the computer simulation of the compressor, the experimental data were obtained from
Available online 10 November 2009
motor performance tests at various operating frequencies.
Using the developed model, re-expansion loss, friction loss, mass ow loss and heat transfer loss is esti-
Keywords:
mated as a function of the crankshaft speed in a variable speed compressor. In addition, the compressor
Variable speed compressor
Rotary compressor
efciency and energy losses are predicted at various compressoroperating frequencies. Since the tran-
Transient sient state of the compressor strongly depends on the system, the developed model is combined with
Dynamic of roller a transient system simulation program to get transient variations of the compression process in the sys-
Dynamic of crankshaft tem.
Motor efciency, mechanical efciency, motor torque and volumetric efciency are calculated with
respect to variation of the driving frequency in a rotary compressor.
2009 Elsevier Ltd. All rights reserved.

1. Introduction Research of rotary compressors can be divided into categories of

The variable speed compressor is designed to change the motor
speed compared to a constant speed compressor. The efciency of
a constant speed motor is dependent on the load of the motor.
However, the efciency of variable speed motor is changed with
not only the load of the motor but also the driving frequency.
Therefore, the efciency of the compressor with a variable speed
motor in the compression and suction process will be continuously
changed with driving frequency and load.
The rolling piston type rotary compressor shown in Fig. 1 is
widely used for controlling the speed of the compressor due to
the simple mechanism. The compressor consisted with a variable
speed motor, crankshaft, roller, cylinder, and vane. Refrigerant en-
ters through a suction side accumulator and is then forced into the
suction chamber directly, where there is small amount of suction
gas heating. The drawn refrigerant is compressed by the rolling
motion of the roller against the cylinder caused by an eccentricity
between roller and cylinder center. The vane sustains the roller by
spring force and it divides the suction and compression volumes in
inside of the cylinder. The compressed refrigerant, which is dis-
charged from the cylinder, resides for a short time in the compres-
sor shell where it undergoes heat transfer and leaves the
compressor and enters the condenser through the discharge port
of the compressor.
* Tel.: +82 64 754 3626; fax: +82 64 756 3886.
E-mail address: ycpark@cheju.ac.kr
constant speed compressors and variable speed compressors. In or-
der to study losses in the constant speed rotary compressor, vibra-
tion of the compressor has been studied through analysis of forces
and velocity of moving parts using the kinematics of the roller,
vane, and crankshaft [1]. The study found that the major source
of the vibration is the variation of angular velocity of the crank-
shaft rather than the velocity variation of the roller. Refrigerant
losses through the axial clearance between the outer wall of the
roller and the inner wall of the cylinder, as well as the radial clear-
ance between the roller and cylinder surface have been studied in
series of four papers [25].
For variable speed rotary compressors, comparisons of the
performance of rotary and reciprocating compressors were con-
ducted through calorimetric experiments with variation of
operating frequency of the compressors [6]. The results showed
that the volumetric efciency of the rotary compressor in-
creases with the operating frequency. In contrary, for the recip-
rocation compressor, mechanical, compression and volumetric
efciency decrease with operating frequency. Another result of
the study shows that the capacity of the rotary compressor is
proportional to the operating frequency. For the reciprocating
compressor, however, capacity of the compressor linearly varies
with operating frequency only in the low frequency range and
there is no uctuation of the capacity in the high frequency
range. From the above results, it is possible to note that the ro-
tary compressor might be more suitable as a capacity regulation
compressor.

0196-8904/$ - see front matter 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.enconman.2009.09.023
278 Y.C. Park / Energy Conversion and Management 51 (2010) 277287

Nomenclature

a length of vane (m) Ps suction pressure (Pa)

A area (m2) Pu upstream pressure (Pa)
Ac cylinder inside area (m2) Qcan heat transfer from compressor case (kJ)
Ah cylinder head area (m2) Qcyc heat transfer from cylinder (kJ)
Ar roller outside area (m2) Qmot heat transfer from motor (kJ)
Asp suction port area (m2) R1, R2 reacting forces on vane (N)
Av vane side wall area (m2) Rc radius of cylinder (m)
b ratio of roller radius to cylinder radius (=Rr/Rc) Re radius of crank n (m)
d diameter of suction port (m) Rj radius of journal bearing (m)
e eccentricity (m) Rr radius of roller (m)
F vector summation of forces acting on roller (N) Rs radius of shaft (m)
Fd gas force acting on root of vane (N) Rv radius of vane tip (m)
Fh gas force acting on vane (N) t time (sec) or vane thickness (m)
Fi inertia force of vane (N) T temperature (K)
Fk spring force acting on vane (N) Tcyl cylinder temperature(K)
Fp pressure force (N) v specic volume (m3/kg)
Fvn normal component of friction force at vane tip (N) V volume (m3)
g gravity (m/s2) W relative humidity
h cylinder height (m) or heat transfer coefcient (W/m2 K) Wr relative humidity at room
hj height of journal bearing (m) W _ compressor power (kW)
i enthalpy (kJ/kg) x vane displacement (m)
Is inertia moment of the crankshaft (N m s2)
k adiabatic coefcient () Greek symbols
l length of suction port (m) a offset angle of eccentric center (rad)
lr length between center of cylinder and exerting point of b angle of suction port (rad)
total force (m) d clearance (m)
lvs length of vane inside vane slot (m) / angle from vane axis (rad)
m mass of uid (kg) ginv inverter efciency
m_ mass ow rate (kg/s) lt coefcient of friction at thrust bearing
m_ loss rate of mass ow loss (kg/s) lj coefcient of friction at journal bearing
m_ cano discharge mass ow rate from compressor (kg/s) lvn coefcient of friction at vane tip
Mf moment action on crankshaft by roller (N m) lvs coefcient of friction at vane side slot
Mj moment of journal bearing (N m) h rotational angle of crankshaft (rad)
Mk moment of friction force between crank n and inside of hf angle of total force (rad)
roller (N m) q density(kg/m3)
Mr moment of friction force between roller and cylinder TL load torque (N m)
head (N m) TM motor torque (N m)
Mt moment of thrust bearing (N m) x angular velocity (rad/s)
N rotational speed of motor (rpm) xe relative angular velocity of between crank pin and roller
P pressure (Pa) (rad/s)
Pd discharge pressure (Pa) xp angular velocity of roller (rad/s)
Pn discharge pressure (Pa) xs angular velocity of crankshaft (rad/s)

A loss analysis on rotary compressors conducted experimentally
with a pressurevolume diagram [7]. The results show that
mechanical losses at low frequency are the most important among
all losses in the compressor even though the absolute value of the
loss is smaller than those seen at high frequency operation. Gener-
ally air conditioners run, for the most part, at low frequency condi-
tions; in order to increase energy saving of air conditioning
systems, the mechanical losses have to be reduced not only at high
frequency operation but also at low frequency operation. Loss anal-
yses of variable speed rotary compressors also can be found in
some literature, including a paper about high-speed compressors
that run at up to 180 Hz frequency [8].
The effect of suction gas heating, mass ow losses, re-expansion
of compressed refrigerant, and over-charging of the refrigerant on
the compressor performance has been studied [9]. This paper sug-
gests that over-charging of the refrigerant increases compressor
capacity with a small degradation of the efciency. Also, how the
compressor performance is affected by refrigerant leakage is the
most important parameter at low frequency operation; while on
the other hand, the re-expansion of the compressed refrigerant is
important at high frequency operation.
In the past, most compressor dynamic analysis was conducted
with constant motor efciency or without consideration of motor
performance, even studies on variable speed compressors. The var-
iable speed motor is the heart of the compressor, which drives the
crankshaft and changes the speed of the compressor. In this study,
the motor performance with variation of the driving frequency was
experimentally determined. This study is concerned with develop-
ing a dynamic program to predict friction loss, mass ow loss, re-
expansion loss, and heat transfer loss as a function of the variation
of driving frequency in a variable speed compressor. Several com-
putational models for the compression process are based on the
polytropic process. In this study, the model is developed with con-
servation equations, a real gas equation and kinematics of the
crankshaft and roller.
The present study can be characterized as follows:
(1) The model in this study is a dynamic analysis model based
on the principles of the mass and energy conservation equa-
tions. The variations of temperature and pressure of the
refrigerant can be calculated with the rotation of the
crankshaft.
 Y.C. Park / Energy Conversion and Management 51 (2010) 277287 279

Hermetic Discharge tube

terminal

Case

Rotor
assembly

Stator
assembly

Top flange

Suction tube
Cylinder

Roller
Bottom flange
Fig. 1. Schematic view of rolling piston type rotary compressor.

(2) The refrigerant is considered as a real gas rather than an
ideal gas.
(3) All of the efciencies of the compressor are calculated based
on performance curves or measured values.
(4) With analysis of the dynamic equation for the moving part,
it is possible to calculate accurate friction loss although
there is angular velocity variation in the crankshaft and
roller.
(5) This model considered variation of the heat transfer coef-
cient and area with rotation of crankshaft.
(6) The variable motor efciency was calculated based on exper-
imental results. Therefore, the uncertainty of the motor ef-
ciency is reduced and it is possible to predict the motor
efciency with the variation of driving frequency.
The calculation for suction and compression process is com-
puted by increasing the angle of the crankshaft with small time
intervals. The refrigerant temperature, pressure, heat transfer,
and mass loss can be calculated by analyses of the compression
and suction process in the cylinder at the same time. In order to en-
sure a more accurate calculation of the losses in the compression
process, dynamic analysis of the compressor is necessary. In this
model, the dynamic equations of the mechanical part of the com-
pressor are used to predict accurate friction loss with calculated
velocity of the moving parts. The mechanical loss, indicated work,
volumetric efciency etc. also can be calculated at the end of a sin-
gle cycle.
A rolling piston type rotary compressor with 3.5 kW nominal
cooling capacity is selected as model compressor. The compressor
using HCFC-22 refrigerant has the dimensions is shown in Table 1.
2. Mathematical modeling
2.1. Thermodynamic governing equations

Table 1
One refrigerant volume in the compressor is being drawn into
Dimensions of model compressor.
the suction chamber while simultaneously another volume is com-
Components Length (cm) pressed in the compression chamber. The rate of change of mass in
Case diameter 13.26 the control volume is given by Eq. (1) and temperature variation
Case height 24.80 Eq. (2) which is obtained using the energy conservation equation
Cylinder height 2.80
and the real gas equation.
Cylinder diameter 5.4
Shaft eccentricity 0.48
Roller outer radius 4.68
dm X X
_ in 
m m _ out 1
Roller inner radius 3.3 dt
Vane height 2.78 P P n    o
Vane width 0.47 dT
dQ
dt
dmindtiin i
 dmoutdt iout i
 m @@iv T  v @@Pv T ddtv
Vane length 2.45 n  @P o 2
dt
v  v @T v
@i
Vane tip diameter 0.50 m @T
280 Y.C. Park / Energy Conversion and Management 51 (2010) 277287

The rate of change of pressure, which is a function of mass, tem- h 2 sin 2h
Vh R2c 1  b  h  1  b2
perature and specic volume is calculated from: 2 2
    
dP 1 dm 1 dT 1 dV 2 1 1
P  3 b sin  1 sin h
dt m dt T dt V dt b
h i1=2

2
The compressed refrigerant from the compression process is b  1  b  sin h  1  b  12 sin h
discharged to the compressor shell, which contains the compres- h i1=2

1
sion compartment and the motor. The refrigerant in the shell is  t  h  Rc 1  1  b cos h  1  b2 cos2 h 2b  1
heated by heat transfer from the motor and cylinder and cooled 2
by environmental air through the compressor case. The refrigerant 6
ow rate discharged from the compressor is obtained with the con-
where b denotes the ratio of the roller radius (Rr) to cylinder ra-
servation equations of mass as shown in
dius (Rc) like as Rr/Rc. The vane displacement with the crankshaft
ml dv v mv dv l v v angle is obtained by
m _ in m
_ cano m _ lv _
m 4
vv dt v v dt v l lv h i1=2

where subscript 00 l00 denotes liquid phase refrigerant, 00 v00 is vapor Xh Rc 1  1  b  cos h  1  b2 cos2 h 2b  1 7
phase, and m_ lv is rate of evaporating refrigerant from liquid to vapor
at saturation state i.e. Due to the fact that the vane tip is not square, the curvature of
 the vane is considered with
dQ cyl dQ mot dQ can div
_ lv m
m _ in iin  iv   mv
dt dt dt dt X r h Rv = cos a  Rv cos a 8

dil dP
ml V iv  il 5 where a is eccentric angle of the roller based on the contact point of
dt dt
the vane and roller as follows:
 
Rc  Rr sin h
2.2. Geometric parameters a sin1 9
Rr Rv
The governing equations should be solved considering volume The volume change rate is obtained by the differentiation of the
change rate for the transient model. While the crankshaft rotates above geometric equations.
with the motor, the volume of the cylinder is obtained by Eq. (6)
with consideration of the volume swept by the vane. Notations dVh R2 n 2
h  c 1  b  1  b2 cos 2h  b1  b  cos h
in the equation can be seen schematically in Fig. 2. dh 2
h i1=2
2
 1  b  12 sin h b1=b  1  cos h
h ih i1=2

2 2
 1  b  sin h  b 1  1=b  12 sin h

1 n h i
  t  h  Rc 1  b  sin h  1  b2 sin h  cos h
Spring h2 i1=2

Fd Fk 1  b2 cos2 h 2b  1 10
R2
Ideal displacement volume of the compressor for the volumetric
t efciency is obtained by the following equation:
Vane lvs Z 2p
a V s p  h  R2c  R2r  t  h  xhdh  Vhs Vhd  11
R1 0
f Fvn where V(hs) and V(hd) are residual volumes with locations at the
x suction and discharge port respectively.
Fh

Suction
chamber

Mc 2.3. Dynamic equation of crankshaft

vnFvn
Rr Because the angular velocity of the motor changes with motor
Compression torque and load, angular velocity of the crankshaft changes with
chamber
Mk the angle of crankshaft. If the motor torque and load are assumed
constant with time, the angular velocity of the crank shaft is ob-
Re Roller tained by [10]:

dx
Is TM  TL 12
dt

Cylinder Fp where Is, TM, and TL are the inertial moment of the crankshaft, motor
torque, and load, respectively.
The load acting on the compressor motor is the same as the to-
tal moment exerted on the crankshaft [2,8,11]. The load is obtained
F
by following:
Fig. 2. Forces and moments acting on compression compartment of the compressor
TL M f Mj Mt 13
and geometric information.
 Y.C. Park / Energy Conversion and Management 51 (2010) 277287 281

Moment of thrust bearing (Mt) in Eq. (13) is calculated with 2.5. Mass ow losses
x 
M t plt 2R4j  R4s
2d
The journal bearing, which consists of a main bearing and sub
bearings, sustains the crankshaft with spring force of the vane.
Friction force caused by the weight of the rotor and the crankshaft
is applied to the journal bearing as a moment [6]. The moment of
the journal bearing (Mj) can be obtained by
0:66lj hj R3s x
Mj
d
The forces acting on the crankshaft by the roller are composed
of (1) a radial direction force (Fvn) on center of the roller imposed
by vane, (2) a friction force (lvnFvn) caused by the aforementioned
force, (3) a force caused by a pressure difference between the com-
pression and suction chambers (Fp), and (4) a centrifugal force aris-
ing from the orbital motion of roller [1]. Among the forces, the
force by the pressure difference is the highest force acting on the
crankshaft. Moment acting on the crankshaft by the roller (Mf) is
calculated with
 
ah
M f e  F p  sin  e  F v n  sina h e  lv n  F v n
2
 cosa h
Fig. 2 shows the forces exerted on the crankshaft, roller and
vane. The force of the vane acting on the center of the roller (Fvn)
is calculated using Eq. (17) and the force caused by the pressure
difference between the suction and compression chamber is calcu-
lated with
14 Suction gas heating, oil circulation, re-expansion of residual vol-
ume, back ow, and gas leakage through the discharge valve affect
mass ow losses in the compressor. The re-expansion ow of resid-
ual gas is calculated by an orice equation with the assumptions of
one-dimensional isentropic process, stagnated upstream condi-
tions, and constant ow coefcient. The mass ow loss by suction
gas heating can be neglected in a low-pressure type rotary com-
pressor due to the fact that the refrigerant is being directly drawn
15 into the cylinder from the accumulator or evaporator. The effect
of oil is also neglected in this study. During the compression pro-
cess, the mass ow is lost through the contact surface between
the roller and cylinder, the vane and roller, and the roller and cylin-
der head.
The calculation of mass losses are based on the following
assumptions: (1) The mass losses come only from pressure differ-
ences, (2) The effect of Coutte ow due to the velocity difference of
the stationary part and the moving part is negligible, (3) The type
of path is assumed to be a convergingdiverging nozzle, and (4) the
refrigerant losses through small paths are critical ow. With above
assumption, the mass loss equation can be delivered as Eq. (25).
The ow area in the equation is calculated at each angle of rotation
of the crankshaft.
16
(  2=k  k1=k !)1=2
2k P P
_ loss
m Pu A  25
k  1RT u Pu Pu


lv s F d F k F i lv s F h lv s x t  lv s
Fvn   17
cos a lv n sin a  lv s 2lv s Rv sin a lv s sin a  lv n cos a  2x lv s t  lv s  2Rv 1  cos a
F p 2Rr  h  sinh a=2  Pd  Ps 18

2.4. Dynamic equation of roller 2.6. Heat transfer

The roller rotates about its center regardless of the rotation of In this study, the refrigerant discharged from the cylinder is
the crankshaft. The motion of the roller is determined only by the accumulated in the compressor case. The heat transfer to the
moment that is acting on the roller. The dynamic equation of refrigerant in the case is accompanied by the motor heat (Qmot),
the roller can be determined by the moment about the center heat transfer from the cylinder (Qcyl) and heat transfer from the
of the roller as in Eq. (19). This assumes that the friction between compressor case (Qcan). The motor heat and heat transfer to the
the roller and cylinder is negligible compared to other terms. compressor case easily calculated by using the motor efciency
and heat transfer equation. The heat transfer in the compres-
dx 2M k  2M r  lv n Rr F v n sion chamber comes out through the cylinder, roller, cylinder
19
dt qplr R4r  R4e head, suction line and vane. The heat transfer is calculated by
the following equation with a corresponding heat transfer area
The moment caused by friction force between the crank n and
[12].
the inner surface of the roller is obtained from Eq. (20). The mo-
ment caused by friction force between the roller and the cylinder Z b2p
head is described by h
Q cyl f h  dh 26
xs b
3
M k 2pl x
hRe e =d 20
M r pl 4 4
xr =d where h is heat transfer coefcient, xs is angular velocity of the
o Rr  Re 21
crankshaft, and f(h) is a function of heat transfer area which is
The relative angular velocity (xe) between the crank n and ro- shown in Fig. 3 and temperature difference as seen in
tor is obtained from [8]:
Z Ac h Z Ah h
xe lr cos hr  x  Re xp 22 f h T cyl /  ThdAc / T cyl /
h i1=2 0 0
lr R2e e2  2Re  e  cosp  hf h 23
 ThdAh / Ap hT cyl  Th Av hT cyl  Th
n 2 o1=2
cos hr 1  e  sinhf  h=lr 24 Asp T cyl  Th 27
282 Y.C. Park / Energy Conversion and Management 51 (2010) 277287

after one rotation of the crankshaft with slip variation of

the crankshaft.
Cylinder (14) Repeat procedure until the variation of the mass in suction
and discharge chambers at each crankshaft angle are con-
Roller verged within the criteria outlined for the study.

Rr Rc 3. Results and discussion

Or
3.1. Model validation
O
Ah Ar Compression To validate the developed simulation program, an experiment
chamber
was conducted with 1 RT nominal capacity air conditioning sys-
tem. Fig. 5 shows schematic diagram and Table 2 shows specica-
Suction
Vane tion of the system. The capacity of the system was measured air
Ac
chamber

enthalpy method by ANSH/ASHRAE Standard 116-1995 [13] in

Asp l
Av
t across the evaporator.
d sure side pressure transducer with 0.11% and 0.5% error were used
Suction port
respectively. All temperature in key location of the system was
Fig. 3. Heat transfer area in suction chamber.
2.7. Calculation procedure
Quasi stead state analysis is conducted with small time inter-
vals, under the assumption that inlet and outlet conditions are con-
stant during the time interval. Mass balance and angular velocity of
the crankshaft are selected as convergence criteria. The compres-
sion process is calculated by solving the physical conservative dif-
ferential equation with rotation angle of the crankshaft. A fourth
order RungeKutta method was used to solve the differential equa-
tion with variation of the crankshaft angle.
psychrometric chambers and the capacity were veried with
refrigerant side capacity by measuring mass ow rate and enthalpy
Coriolis type mass ow rate with 0.4% error and high/low pres-
to measure the mass ow rate of the refrigerant and pressure,
measured with T-type thermocouples. Power consumption was
measured by 0.25% accuracy watt transducer. The measured data
was achieved with data logger and stored directly at personal
computer.
The total error of the experiments was calculated with uncer-
tainty analysis. In the air enthalpy method, capacity of the system
was sum of latent and sensible heat transfer of the air through the
evaporator. Sensible capacity could be calculated by Eq. (28) and
measurement error for the sensible capacity could be achieved
by derivation of the Eq. (28) with respect to each measured param-
eter which effecting on capacity calculation as shown in
 1=2
Ps Pn
qs 1:414  C  An cpa cpw W n T 2  T 1 28
RT n 1 W n
The frictional losses are calculated based on an angular velocity 2 2  2  2 31=2
that is obtained from an analysis of the kinetic equation between @qs =@A
DA @qs =@W n
DW n @qs =@T 2 T 1
DT 2  T 1
the crankshaft and roller rather than the friction that is calculated Dq s 6 qs qs qs 7
6
4  2  2  2
7
5
based on a constant motor speed. qs @qs =@P s @qs =@P n @qs =@T n
qs
DP s qs
DP n qs
DT n
The calculation owchart is shown in Fig. 4 where the proce-
dure is as follows: 29
Similarly, the error by latent heat capacity measurement can be
(1) Volume of the cylinder and length of the vane with
calculated by Eq. (30) from derivation of latent heat capacity
geometry.
(2) Properties of suction refrigerant. 2  2  2  2 31=2
@ql =@A @ql =@W n @qs =@T n
(3) Rotational speed of the crankshaft. D A D W D T
Dq l 6 ql ql n qs n 7
(4) Heat transfer area and heat transfer coefcient. 6
4  2  2 
7
2 5 30
(5) Temperature, mass ow rate, pressure, heat transfer rate, ql @qs =@P s @qs =@P n @qs =@W r
q DP s q
DP n q
DW r
s s s
ow loss by the fourth order RungeKutta method.
 1=2
(6) Friction loss at crankshaft. Ps Pn
(7) Motor torque from motor performance curve with assumed ql 1:414  C  An  ifg W r  W n 31
RT n 1 W n
rotational speed of the crankshaft.
(8) Angular velocity of the crankshaft by Euler method with load Based on measured pressure, temperature and nozzle area, the
and motor torque exerted on the crankshaft based on the error of each parameter in Eqs. (29) and (30) were estimated as
calculated friction loss. shown in Table 3.
(9) Friction loss and angular velocity of the roller. Error of latent heat capacity was calculated as 0.0462 kW and
(10) Heat transfer rate, suction loss, discharge loss and indicated sensible heat capacity as 0.1663 kW. Total error of the capacity
work in the cylinder. measurement of the system could be calculated with
(11) Volume change rate, velocity of the vane and variation of the Dq 0:16632 0:4622 1=2 0:1726 kW 32
enthalpy with respect to rotation of the crankshaft.
(12) Continue above (4)(11) calculation until the crankshaft With the measured latent capacity (1.247 kW) and sensible
reached 360. capacity (2.082 kW), the uncertainty of the experiment for the sys-
(13) Continue above (4)(12) calculation until the angular veloc- tem capacity was 5.2% as shown in
ity at initial condition matched with the angular velocity
 Y.C. Park / Energy Conversion and Management 51 (2010) 277287
Fig. 4. Flow chart of the variable speed compressor model.
Dq 0:1726
0:052 5:2% 33
q 2:082 1:246
Fig. 6 shows an air conditioning system performance with var-
iation of driving frequency of the compressor. It shows that mean
error is 20.8% in the capacity and 10.7% in the power consumption.
The error is increased with increasing the compressor frequency.
The experimental error of the system is about 5.2% as Eq. (33).
Therefore, the most of error in capacity prediction of the system
comes from model development. In worst case, the error of the sys-
tem modeling is about 15.6% in this study. It is assumed that this is
due to the thermal mass of the system components. The simulation
program had over estimation in mass of the component.
3.2. Motor performance
The motor efciency and motor torque with variation of load are
determined at each driving frequency in the simulation program.
283
The motor performance data at specied frequency from the man-
ufacture was curve tted in this study at various frequencies.
Motor slip is the ratio of actual rotational frequency of the shaft
to the driving frequency applied to the motor. If the slip ratio is
small, the driving frequency is similar to the motor rotational fre-
quency. Fig. 7 shows rotational frequencyefciency diagram at
60 Hz. The efciency increased with slip ratio until slip ratio be-
came approximately 0.04 (3456 rpm). After the efciency reached
the maximum, the efciency decreased with slip ratio. If the com-
pressor is designed as the load match with this rotational fre-
quency, the compressor efciency should be maximized. The
motor torque is linearly proportional to slip.
Motor torque and efciency variation with frequency is shown
in Fig. 8. The data was achieved at a given load of the compressor at
25 kg cm motor torque. The motor torque imposed to the motor is
assumed that the load of the motor is the same as the motor tor-
que. If the driving frequency is changed from 30 Hz to 60 Hz, the
efciency is increased from 67% to 80%, approximately. Neverthe-
284 Y.C. Park / Energy Conversion and Management 51 (2010) 277287

Outdoor

Check Valve
Indoor
P T Mass Flow Meter

P T
T P
Capillary

Heat Exchanger
Tube

Axial Flow Fan

Cooling

Outdoor
Heating
Heat Exchanger

Cross Flow Fan

P Pressure Transducer
T Thermocouple
Indoor

Capillary
Tube

T P

Separator
Accumulator

Oil
Cooling
Heating
4-Way
P T Compressor Valve

Fig. 5. Schematic diagram of an experimental setup.

Table 2
Specication of test unit.

Indoor unit Outdoor unit

Rated performance Cooling capacity kcal/h 3550
Heating capacity kcal/h 5000
Heating capacity (low temperature) kcal/h 4000 (outdoor temperature 10 C)
Air ow rate Cooling m3/min 8.4
Heating m3/min 9.7
Noise dB 45 57
Electrical characteristic Power /-V-Hz 1-220-60
Power consumption Cooling W 1830
Heating W
Operating current Cooling A 8.5
Heating A 12
Specication Dimension WHD mm 1050  298  180 900  550  300
Net weight kg 14 51
Compressor Type Rotary type
Output W 1100
Motor Output W 30 80
Fan Type Cross ow Axial ow
Size /  L (mm) 95  860 402
Capillary tube /  L  EA 1.4  700  2,
1.7  300  1
Refrigerant charge g 1250

less, as the driving frequency increases over 60 Hz, there is no sig-
nicant increase in the efciency, with the maximum efciency
(81%) at a frequency of 90 Hz frequency.
The motor efciency is lower at low frequencies than it is at
high frequencies. In addition, the motor output is a function of tor-
que and rotational frequency as represented in
P 2  p  T  N=60
The motor output varies linearly with the driving frequency;
this is clear from the above equation. The linearity of the motor
output variation changes a little with the slip of the motor.
The power supplied to the three-phase induction motor in the
compressor is a single-phase AC power originally. The AC power
changed to DC power at a rectier and a PWM driver in converter
changes the DC power again to three-phase AC power that is then
supplied to the induction motor. There is a power loss during these
converting processes and can be captured by an inverter efciency.
The inverter efciency in Fig. 12 is expressed by Eq. (35) as func-
34
tion of the compressor power.
ginv 86:2237 5:5879W_  0:7233W_ 2 35
 Y.C. Park / Energy Conversion and Management 51 (2010) 277287 285

Table 3 1800 100

Estimated uncertainties for error analysis.
Output
1600 Efficiency
Items Estimated uncertainties Remarks
90
DA 1.8414e05 m2 k = 3 Std. deviations of the mean 1400

Motor efficiency (%)

DP s 6.9346 Pa k = 3 Std. deviations of the mean

Motor output (W)

DP n 27.7385 Pa k = 3 Std. deviations of the mean 1200 80
DP 27.7385 Pa k = 3 Std. deviations of the mean
DP w 43.829 Pa Calculated 1000
DTn = DTd 0.8 C Manufacturer spec. uncertainty
DWn = DWr 2.6982e04 Calculated 800 70
D(T2  T1) 0.8 C Manufacturer spec. uncertainty
600
60
400
5.0
Cooling capacity ; [experiment], [simulation] 200 50
4.5 20 40 60 80 100 120
Cooling capacity, Power input(kW)

Power input ; [experiment], [simulation]

4.0 Frequency (Hz)
3.5 Fig. 8. Variation of motor efciency and output power with driving frequency at a
constant torque (25 kg cm).
3.0

2.5

2.0 25
1.5
Condensing pressure
1.0
20
0.5

0.0
Pressure (bar)
20 30 40 50 60 70 80 90 100 110 120
with heat transfer
Frequency(Hz) 15
without heat transfer
Fig. 6. Verication of the developed model with experimental data.

10
80 40
m Evaporating pressure
Tm
5
79 35
0 100 200 300 400 500 600 700 800
Tm(motor torque, kg cm)
m (motor efficiency, %)

o
Crank angle ( )
78 30
Fig. 9. Variation of refrigerant pressure in the cylinder with crankshaft angle
considering heat transfer during one compression cycle (condition: evaporating
temperature at 7.2 C, condensing temperature at 54.4 C).
77 25

76 20 Table 4
Effect of the heat transfer on compressor performance.

Without heat transfer With heat transfer

75
0.03 0.04
s (slip ratio)
Fig. 7. Variation of motor efciency and torque with slip ratio at a rated frequency
(60 Hz).
3.3. Compression characteristics
The analyses are conducted by varying the crankshaft angle
through one complete revolution. In this analysis, the operating
condition for the compressor is 7.2 C (624.9 kPa) evaporating tem-
perature and 54.4 C (2145 kPa) condensing temperature. Fig. 9
shows the pressure variation with crankshaft angle. The suction
process is nished at 360. The compression process starts at
397.75 due to the location of the suction port at 37.75 from the
vane. The angle was increased to more than 300, the pressure be-
came slightly higher than the evaporating pressure due to the re-
expansion of the drawn refrigerant in the suction chamber.
Although the pressure drop during the suction process is small,
the suction losses are not small compared to the discharge losses
15
0.05 0.06 Input power (W) 2042 1913
Volumetric efciency (%) 94.3 74.1
Motor efciency (%) 60.8 63.9
in the compression process because of the prolonged suction pro-
cess. The dotted line shown in the Fig. 9 is the pressure variation
with crank angle without considering heat transfer during one
rotation. The effect of heat transfer is more dominant in the com-
pression processor than in the suction process. The change in per-
formance of the compressor caused by the heat transfer is shown
in Table 4. The heat transfer does not signicantly affect the perfor-
mance of the motor. The motor efciency changed by only 3%. The
indicated power increased by 6.7% compared to the case with heat
transfer. However, the volumetric efciency decreased about 20%
with the heat transfer present.
The load and motor torque variation with crank angle are
shown in Fig. 10, these are calculated by the dynamic equation of
the roller. The difference between the load torque and motor tor-
que makes an angular velocity variation as seen in Fig. 11. From
the angular velocity variation, the friction loss in the compressor
286 Y.C. Park / Energy Conversion and Management 51 (2010) 277287

16 losses are coupled, these should be calculated using an iteration

Motor torque(TM) method.
14
Load torque(TL) The variation of the compressor performance with driving fre-
12 quency was calculated as shown in Fig. 12. All efciencies shown
in Fig. 12 are normalized with respect to the efciency at a rated
10
condition. The results indicate that the volumetric efciency de-
Torque (N m)

8 creases slowly when the frequency deviates from the rated condi-
tion. The adiabatic efciency is almost constant for low
6
frequencies; however, it decreases due to suction and discharge
4 losses as the frequency increases. The mechanical efciency is in-
2
versely proportional to the angular velocity of the crankshaft; it de-
creases as the frequency of the compressor increases.
0

-2 3.4. Start-up characteristics

0 60 120 180 240 300 360
o
Crank angle ( )
Most studies on compressors focus on the compressor perfor-
Fig. 10. Variation of load torque and motor torque with crankshaft angle (condi- mance at steady state. Because the transient state of the compres-
tion: evaporating temperature at 7.2 C, condensing temperature at 54.4 C). sor strongly depends on the system, the developed transient model
for variable speed compressors in this study was combined with a
transient system simulation program. The developed model is ap-
0.11 350 plied to predict transient variation of compression processes dur-
ing start-up period with results being shown from start-up to
steady state. The calculation runs at the condition specied in ASH-
345 RAE standard 116-1995 [13] with 26.7 C indoor and 35 C outdoor
0.10
Angular velocity (rad/s)

temperature for air conditioning mode.

The refrigerant pressure, work done by refrigerant and mass of
Slip ratio (-)

the refrigerant in the cylinder with crankshaft angle are shown in

0.09 340
Figs. 1315. In these gures, the variation of the parameters with
time is dependent only on system performance. The pressure var-
iation in Fig. 13 shows that the discharge loss at the early stage of
335 start-up is signicantly large due to the discharge valve opening
0.08
early (450) and remaining open for a long period (700). However,
Angular velocity there is advantage in volumetric efciency due to lower tempera-
Slip ratio
330 tures in the compressor compartment. As time goes on, the dis-
0.07
0 60 120 180 240 300 360 charge valve opens at 570 angle of crankshaft, which is the
o steady state value.
Crank angle ( )
The work of the refrigerant during suction and discharge pro-
Fig. 11. Variation of slip ratio and angular velocity of crankshaft with crankshaft cesses is shown in Fig. 14 as an energy term. Positive means work
angle (condition: evaporating temperature at 7.2 C, condensing temperature at is done by the refrigerant and negative means the work is acting on
54.4 C). the refrigerant. Since the suction pressure is higher than steady
state in the early state of the compressor start-up, the work done
by the refrigerant is also higher.
1.3
The refrigerant mass during the suction and discharge process
Volumetric varies as shown in Fig. 15. When the compressor is started, the
Adiabatic
1.2 Mechanical mass of refrigerant in the cylinder is large due to a high evaporat-
Inverter
Relative efficiency

1.1 20
t=t+t
18
1.0 16
14 t=t+t
Pressure (bar)

0.9
12
10 t=t+t
0.8
20 30 40 50 60 70 80 90 100 110 120 8
Frequency(Hz)
6
Fig. 12. Variation of efciencies with driving frequency of the compressor as 4
normalized with efciency at the rated frequency (60 Hz).
2
0 100 200 300 400 500 600 700
o
can be calculated. The slip ratio shown in the Fig. 11 is inversely Crank angle ( )
proportional to the angular velocity of the crankshaft. The motor
Fig. 13. Variation of refrigerant pressure in the cylinder with crankshaft angle from
efciency can be calculated at a specied angular velocity using start-up of compressor (indoor air temperature at 26.7 C, ambient air temperature
the obtained slip ratio. Because the angular velocity and friction at 35 C).
 Y.C. Park / Energy Conversion and Management 51 (2010) 277287 287

10 (1) Motor output linearly increases with driving frequency.

However, efciency of the motor increases only up to the
t=t+t
5 rated frequency, while over the rated frequency the ef-
ciency remains approximately constant value. On the other
hand, indicated efciency decreases due to incremental
Energy (kJx 10 )

0
-3

changes of the suction and discharge losses with incremen-

tal changes of the compressor driving frequency.
-5 (2) The mechanical efciency is inversely proportional to the
angular velocity of the crankshaft, it decreases with the driv-
-10 t=t+t ing frequency because friction losses are directly propor-
tional to the rotational speed of the crankshaft.
-15 (3) If the load of the compressor changes at a specied driving
frequency, motor torque is linearly proportional to the slip
-20 between the rotor and the stator. When the imposed fre-
0 100 200 300 400 500 600 700 800 quency of the motor is 60 Hz, the motor efciency varies
o
Crank angle ( ) quadratically with respect to the slip ratio, with a maximum
efciency at a slip ratio of 0.04.
Fig. 14. Work done on () and done by (+) refrigerant with crankshaft angle from (4) If there is no heat transfer during the compression process,
start-up of compressor (indoor air temperature at 26.7 C, ambient air temperature
the volumetric efciency increases about 20%, however,
at 35 C).
the efciency change in the motor is only 3% compared to
the heat transfer case.
0.8 (5) When the variable speed compressor is installed in an air
conditioning system, the suction and discharge loss is large
and motor efciency is low at early stages of start-up. How-
0.6 ever, volumetric efciency is higher due to the lower tem-
Refrigerant mass (g)

perature of the compressor.

0.4 t=t+t
Acknowledgements

0.2 The presented research is supported by research grant of the

Cheju National University, the 2nd phase Bk21 Project, and the
Korea Science and Engineering Foundation (MEST R01-2008-000-
0.0 10993-0).

0 100 200 300 400 500 600 700 References

o
Crank angle ( )
Fig. 15. Variation of refrigerant mass in the cylinder with crankshaft angle from
start-up of compressor (indoor air temperature at 26.7 C, ambient air temperature
at 35 C).
ing pressure and a small compression ratio. A high evaporating
pressure creates a large pressure difference between the evaporat-
ing and suction chamber pressures, increasing the mass ow to the
suction chamber, which was calculated by the orice equation.
4. Conclusions
A dynamic simulation model for a variable speed compressor is
presented. Using this model, the compressor efciency and energy
losses can be predicted at various compressoroperating frequen-
cies. It is also possible to predict the variation of compressor char-
acteristics with time upon start-up of the system.
The model is based on the physical principles of conservation of
mass and energy, including the dynamic equation of the roller and
crankshaft, rather than effectivenessefciency based models. In
addition, the model considered other parameters that affect the
compressor performance such as leakage, friction, heat transfer,
and re-expansion of compressed refrigerant etc.
Through the simulation of the compressor characteristics, the
following results are presented in this study.
[1] Imaichi K, Fukushima M, Muramatsu S, Ishii N. Vibration analysis of rotary
compressor. In: Proceeding of the Purdue compressor technology conference,
vol. 55; 1982. p. 27582.
[2] Yanagisawa T, Mori M, Shimizu T, Ogi Y. Vibration of a rolling piston type
rotary compressor. Int J Refrig 1984;7(4):23744.
[3] Yanagisawa T, Shimizu T. Leakage losses with a rolling piston type rotary
compressor: I. Radial clearance on the rolling piston. Int J Refrig
1985;8(2):7584.
[4] Yanagisawa T, Shimizu T. Leakage losses with a rolling piston type rotary
compressor: II. Leakage losses through clearance on rolling piston faces. Int J
Refrig 1985;8(3):1528.
[5] Yanagisawa T, Shimizu T. Friction losses in rolling piston type rotary
compressor III. Int J Refrig 1985;8(3):15965.
[6] Itami T, Okoma K, Misawa K. An experimental study of frequency-controlled
compressor. In: Proceeding of the Purdue compressor technology conference,
vol. 44; 1982. p. 297304.
[7] Iwata H, Sakazume A, Yokoyama K. Loss analysis of rolling piston type speed
controlled rotary compressor. JSME 1985;51(465):173641.
[8] Mitsunaga T, Nonaka R, Matsumoto K. Rotary compressor for 180 Hz inverter
air conditioner. SANYO Tech Rev 1987;19(2):10212.
[9] Liu Z, Soedel W. Performance study of a variable speed compressor with
special attention to supercharging effect. In: Proceeding of the international
compressor engineering conference at Purdue, vol. 2; 1994. p. 499506.
[10] Matsuoka F, Nagamoto H. Dynamic response and electrical control for the air
conditioner. Trans Jpn Assoc Refrig 1988;5(1):4354.
[11] Sakaino K, Kawasaki K, Shirafuji Y, Ohinata O. The study of dual cylinder rotary
compressor. In: Proceeding of the Purdue compressor technology conference,
vol. 33; 1986. p. 292304.
[12] Yanagisawa T, Shimizu T, Dohi M, Nihashi S. Suction gas heating with rolling
piston type rotary compressor. JSME 1983;49(444):176271.
[13] ASHRAE Standard 116-1995. Methods of testing for rating seasonal efciency
of unitary air conditioners and heat pumps. Atlanta, GA, USA: American
Society of Heating, Refrigeration and Air-Conditioning Engineers, Inc.; 1995. p.
146.