IOP Conference Series: Materials Science and Engineering

PAPER • OPEN ACCESS You may also like

- Dynamic analysis of the reciprocating
A thermal model for analysis of hermetic compressor package on the offshore
platform
reciprocating compressors under the on-off cycling Y Zhao, W Wang, Q Zhou et al.

- Stretchers and compressors for ultra-high

operating condition power laser systems
I.V. Yakovlev

To cite this article: S K Lohn et al 2015 IOP Conf. Ser.: Mater. Sci. Eng. 90 012068 - Stiffness control of magnetorheological
gels for adaptive tunable vibration
absorber
Hyun Kee Kim, Hye Shin Kim and Young-
Keun Kim

View the article online for updates and enhancements.

This content was downloaded from IP address 161.9.127.36 on 21/01/2024 at 10:35

9th International Conference on Compressors and their Systems IOP Publishing
IOP Conf. Series: Materials Science and Engineering 90 (2015) 012068 doi:10.1088/1757-899X/90/1/012068

A thermal model for analysis of hermetic reciprocating

compressors under the on-off cycling operating condition

S K Lohn, M C Diniz and C J Deschamps1

POLO Research Laboratories
Federal University of Santa Catarina, 88040900, Florianopolis, SC, Brazil

E-mail: [email protected]

Abstract. The on-off cycling operating condition of compressors is very common in low
capacity refrigeration systems, being characterized by alternate periods in which the
compressor is either operating (on) or idle (off). Thermal interactions between the compressor
components affect its performance during the operating period and establish the initial
condition for the compressor start up from idle condition. This paper presents a numerical
model to predict the temperature field of hermetic reciprocating compressors under on-off
cycling conditions. The model adopts a lumped formulation for control volumes formed in the
fluid solution domain and the finite volume method to solve heat conduction in the solid
components. Some required heat transfer coefficients were experimentally adjusted.
Predictions for temperature were compared to measurements and good agreement was
observed, especially for the thermal transient during the period in which the compressor is off.

1. Introduction
Reciprocating compressors are widely applied in low capacity refrigeration applications, commonly
associated with an on-off cycling operating condition characterized by alternate periods in which the
compressor is either operating (on) or idle (off).
Due to economic and environmental issues, the refrigeration industry is oriented towards
performance improvement, thus requiring the analysis of aspects related to the transient operation of
the system. The current literature related to transient system conditions is focused on the heat
exchangers and expansion device [1]. Therefore, it is necessary to improve the knowledge on the
compressor performance during the on-off cycling condition in order to increase the system efficiency.
Thermal interactions between internal components of hermetic reciprocating compressors strongly
influence their thermodynamic efficiency, as well their reliability. On the other hand, thermal
interactions during the off period define the initial condition for the compressor start-up.
Heat transfer inside hermetic compressors has been studied both experimentally [2] and
numerically [3]. Despite the good insight provided by experimentation, the use of numerical
approaches presents significant benefits, such as versatility and fast results. Heat transfer in hermetic
compressors is modelled employing different degrees of complexity, and a wide variety of steady state
models is available in the literature [4]. Simple semi-empirical models, such as the one developed by
Todescat et al. [5], provide very fast predictions of the temperature field. Other models, such as the
totally differential model of Raja et al. [6] are characterized by high computational cost, although
1
To whom any correspondence should be addressed.

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
9th International Conference on Compressors and their Systems IOP Publishing
IOP Conf. Series: Materials Science and Engineering 90 (2015) 012068 doi:10.1088/1757-899X/90/1/012068

allowing deeper understanding of the phenomena. As a compromise, Ribas Jr. [7] developed a hybrid
model that combines a differential approach for the solid domains and a lumped formulation for the
gas chambers.
This paper presents a numerical model to predict the temperature field of a hermetic reciprocating
compressor under on-off cycling conditions, based on the hybrid approach [7]. Predictions for
temperature in different regions inside the compressor under the transient condition were validated
through comparisons with measurements.

2. Compression cycle model

The thermal model requires input data related to the compression cycle. The modelling of the
compression cycle employed in this work is based on four groups of equations [5], following a
transient lumped formulation. The time scale of the compression cycle (milliseconds) is considerably
smaller than that of the transient temperature field of solid components (minutes). Hence, the results
obtained through the compression cycle model are integrated over a representative cycle to be used as
input data in the thermal model.
The first group of equations calculates the instantaneous volume of gas inside the cylinder with
respect to the compressor geometrical parameters. The second group of equations are the conservation
equations (mass and energy) applied to the gas inside the cylinder. With the instantaneous cylinder
volume and the mass of gas inside the cylinder, obtained through equation (1), it is possible to
calculate the gas density.

mG
m
 suc  m
 suc,b  m
 dis  m
 dis,b  m
 leak . (1)
t
The manipulation of the lumped energy balance originates the following equation to predict the gas
temperature during the compression cycle:

 
 m h   m h
dTG 1 dmG
  H G AwTw  hG 
dt mG cv ,G  dt in out (2)
,
TG  p dG pG dm 
  H G Aw  G  G G 
mG cv ,G  TG v
dt TG v dt 

where HG is the convective heat transfer coefficient estimated with the correlation of Annand [8]. Two
temperatures are very important in equation (2): the refrigerant temperature in the suction chamber
(Tsc) and the temperature of the cylinder wall (Tw). These temperatures can either be specified from
experimental data or calculated with a thermal model, as in this work. With the temperature estimated
through equation (2) and the density obtained via equation (1), the instantaneous pressure is calculated
using an equation of state. The equation of state and all other thermodynamic properties are obtained
using the Refprop library [9].
The third group of equations represents the valve dynamics. A one-degree of freedom mass-spring
model is used, with effective force areas being used to characterize the force acting on the valve [10].
Finally, the fourth group of equations predicts the mass flow rates through the valves and the gas
leakage through the piston-cylinder gap. Theoretical estimates of mass flow rate in valves are obtained
from the isentropic flow through a convergent nozzle and then corrected by using the concept of
effective flow area to account for viscous effects. Leakage through the piston-cylinder gap was
modelled as a fully developed laminar Couette-Poiseuille flow [11].
Predictions from the compression cycle model that are used as input data in the thermal model are
given in Table 1. The amount of heat generated due to friction in the piston-cylinder gap, Q̇pc, is
calculated with the Couette-Poiseuille velocity profile, whereas the electrical efficiency, ηele, is
obtained experimentally.

2
9th International Conference on Compressors and their Systems IOP Publishing
IOP Conf. Series: Materials Science and Engineering 90 (2015) 012068 doi:10.1088/1757-899X/90/1/012068

Table 1. Input data from compression cycle model to thermal model.

Variable Description Calculation

m Mass flow rate  dis dt
f m


Heat transfer rate inside Q pc  AW f H G (TW  TG )dt
QW cylinder
W ind Indicated power f p G dG

Q bea Mechanical losses Obtained from experimental data

Q mot Electrical losses (Wind  Q bea  Q pc )(1/ele  1)

3. Thermal model
Figure 1 presents a schematic view of the compressor simulated in this work, identifying the solution
domains of each model employed. The hybrid thermal model solves the three-dimensional heat
conduction in solid components with a differential formulation and the control volumes with gas
through a lumped modelling approach.

Figure 1. Domains each sub models in the thermal model.

3.1. Solid components

The differential model is solved, with suitable boundary conditions, through a commercial code [12]
that uses the finite volume method. On the external surface of the compressor shell, convective heat
transfer is established with the ambient. For the internal solid components in contact with refrigerant
fluid (motor, cylinder, crankcase, etc.) a convective boundary condition was also specified:

H ch Tsol  Tch   k
dT
, (3)
dnˆ

where Tsol is the temperature of a generic solid surface, Tch is a temperature of a generic gas control
volume and Hch is the convective heat transfer coefficient. There are also regions inside the

3
9th International Conference on Compressors and their Systems IOP Publishing
IOP Conf. Series: Materials Science and Engineering 90 (2015) 012068 doi:10.1088/1757-899X/90/1/012068

compressor where internal heat generation must be specified. For instance, heat generation due to
electrical motor inefficiencies is specified in the finite volumes that discretize the electrical motor, i.e.:

Q
Q ' ' 'mot  mot . (4)
mot

Heat generation due to mechanical losses is characterized as a source term in a virtually defined
thin wall between the shaft and the bearing [12]. Therefore, similarly to the heat generation in the
motor:

Q
Q ' ' 'bea  bea . (5)
sur

Finally, the total heat transfer at the cylinder wall is also used as boundary condition in the
differential model:

QW dT . (6)
 k
AW dnˆ

3.2. Gas control volumes

The solution domain of gas inside the compressor is divided into several control volumes (internal
environment, suction muffler, discharge chamber, discharge muffler and discharge line) and the
temperature in each one of them is calculated using lumped energy balances. The transient term on the
energy balances was only employed in internal environment control volume and the time step depends
on the condition (on/off). Therefore, the implementation of the models for the on and off periods of
the compressor will be addressed separately.

3.2.1. Off period. During the period in which the compressor is off, the compression cycle is not
simulated. Furthermore, it is not necessary to simulate all the gas control volumes of the compressor
immediately after the shutdown, as the higher heat capacity of the solid components forces thermal
equilibrium between the solid walls and the adjacent gas in almost all the gas control volumes. The
exception is the internal environment (ie) and, for this reason, only this control volume was simulated.
Immediately after the compressor is turned off, pressure in all control volumes defined for the gas
domain is considered equal to the equalization pressure. The energy balance for the compressor
internal environment is given by:

H ch Tsol  Tie  t  mcP (Tiet  t  Tiet ) . (7)

During the off period, heat transfer is mainly caused by natural convection, with the convective
heat transfer coefficients being estimated using a modified version of the Catton equation for a
rectangular cavity [13]:

 Pr  
0.28

H ch  C   k  Tsol  Tie  , (8)
0.28

 0.2  Pr   
where C is an experimentally calibrated coefficient required to adjust the Catton equation to the
complex geometries of the compressor components. Four heat transfer coefficients are necessary to

4
9th International Conference on Compressors and their Systems IOP Publishing
IOP Conf. Series: Materials Science and Engineering 90 (2015) 012068 doi:10.1088/1757-899X/90/1/012068

represent the convective heat transfer in the compressor internal environment, and a different constant
C is required for each one. A new temperature for the internal environment chamber is calculated
using equation (7).

3.2.2. On period. The simulation of the compression cycle is carried out when the compressor is
switched on. The transient term in the energy balances on the control volumes is only important in the
equation for the internal environment, as the thermal capacities of the remaining volumes are
negligible. This energy balance remains in the same form shown in equation (7). The heat transfer
coefficients for the internal environment are calculated using a modified version of the correlation for
turbulent heat transfer over flat plates:

  Pr
H ch  C m oil
0.8 0.33

k 0.8 , (9)

where the mass flow rate, ṁoil, is calculated using the procedure presented by Drost [14]. The other
control volumes for the gas (suction and discharge paths) are simulated using steady state energy
balances given by

H ch Tsol  Tch   m
 hout  hin  , (10)

where the mass flow rate, m , is calculated in the compression cycle model. In this case, the
convective heat transfer coefficients were estimated using modified equations for turbulent flow inside
ducts, given by:

  Pr
H ch  C m 0.8 0.33

k 0.8 , (11)

where C is a constant that is experimentally adjusted, since it is affected by the compressor geometry.
In total, there are eight heat transfer coefficients to be determined for the on period.

3.3. Coupled solution

The thermal model is simulated in a coupled manner with the compression cycle model explained in
Section 2. The coupled solution is represented in the flowchart of Figure 2, where the inputs and
results of each sub model can be observed. The inputs in dark grey are given by experimental data,
while the ones in light grey are updated by the different models. Every time the compression cycle
model is called, five cycles are simulated in order to achieve a fully established cyclic condition for
the input conditions. After receiving data from the compression cycle model, the thermal model runs a
certain number of time steps while maintaining constant the input data received from the compression
cycle model. During these iterations, the differential model and the lumped models for the control
volumes exchange information through heat transfer between components. After the specified number
of iterations, the thermal model provides the temperatures to be used as input in the simulation of the
compression cycle. The procedure is repeated along the time defined for the on period.

4. Results
As mentioned in the previous section, it is necessary to adjust the constants C for the heat transfer
coefficients of the thermal model based experimental data. To this extent, a R134a compressor with a
displacement volume of 7.55 cm3 was tested in a hot cycle test bench, being submitted to the operating
condition characterized by Tevap = -23.3ºC (pevap = 0.115 MPa) and Tcond = 40.5ºC (pcond = 1.03 MPa).
Both the gas inlet temperature and the external environment temperature were set to 32ºC. The
compressor was instrumented with thermocouples in locations in the gas and solid: (a) gas: suction
line, suction muffler inlet, suction muffler outlet (suction chamber), discharge chamber, discharge
muffler, discharge line, internal environment; (b) solid: cylinder, shell and motor.

5
9th International Conference on Compressors and their Systems IOP Publishing
IOP Conf. Series: Materials Science and Engineering 90 (2015) 012068 doi:10.1088/1757-899X/90/1/012068

Start of simulation (t=0)

Specify starting values for temperature

field and guesses for input data

Tevap, Tcond Define instantaneous

Tamb operating conditions

Tch, Qw, Solve finite volume model and

Q'''bea, Q'''mot update Qch and Tcyl

Solve chamber models and

m, Qch, Tcyl
update gas temperatures (Tch) t = t + Δt

Yes Is the N a
multiple of 10 or has Tw
varied more than 1ºC?

Solve compression No
cycle model and
update m, Wind, Tch, Tcyl
Qw, Qbea and Qmot

No
t > t F?

Yes

End of simulation (t=t F)

Figure 2. Flowchart of coupled simulation.

The temperature distribution obtained after the compressor reached thermal equilibrium was used
to adjust the constants C in the heat transfer correlations for the on period. For the suction region, the
heat transfer coefficients were within 30 and 60 W/m2K during the on period. In the discharge region,
the values ranged from 100 to 250 W/m2K. For the off period, an experimental temperature
distribution obtained 25 minutes after the compressor was turned off was adopted to adjust the
constants C. For the off period the heat transfer coefficients are lower since heat transfer occurs by
natural convection.
The mesh used to discretize the solid components was generated for the numerical model after
introducing simplifications in the original geometry. In regions of the fluid flow where heat flux and
exchange of momentum are intense, special care was taken to ensure that the gradients were correctly
represented. The mesh generated had a total of 1.3 million volumes to guarantee the truncation errors
were negligible. Due to the very different heat capacities of the gas and solid parts, the time step in the
solution of the problem is limited by the thermal transient of the gas inside the chambers. The values
used for the time step were Δt = 7s for the off period and Δt = 2s for the on period.

6
9th International Conference on Compressors and their Systems IOP Publishing
IOP Conf. Series: Materials Science and Engineering 90 (2015) 012068 doi:10.1088/1757-899X/90/1/012068

For the off period, the simulation model was validated by comparing numerical and experimental
results for temperature distribution of the compressor, during the 210 minutes after the compressor
was shut down from a thermally stabilized condition. Figures 3 and 4 show very good agreement
between numerical results and experimental data. During the off period, there is no gas compression,
heat generation in the motor and bearings. Hence, the compressor average temperature is reduced due
to the heat transfer to the external environment. As can be observed, the decay of temperature is
similar for the solid and gas regions. This indicates that the temperature distribution during the off
period is governed mainly by the heat capacity of the compressor solid parts.

Figure 3. Numerical vs. experimental results Figure 4. Numerical vs. experimental

of Tw and Tmot for the off period. results of Tsc and Tie for the off period.

Figures 5 and 6 show the comparison between numerical and experimental results for a period after
the compressor is switched on. The agreement is still good, despite the presence of some deviations in
the first 90 minutes of operation. In fact, better agreement is observed as the thermal equilibrium
condition as reached (t > 180 min), because the constants C for the heat transfer coefficients were
adjusted using the temperature distribution for this condition.

Figure 5. Numerical vs. experimental results Figure 6. Numerical vs. experimental results
for Tw and Tmot for the on period. for Tie and Tsc for the on period.

7
9th International Conference on Compressors and their Systems IOP Publishing
IOP Conf. Series: Materials Science and Engineering 90 (2015) 012068 doi:10.1088/1757-899X/90/1/012068

After being validated, the model can be used to evaluate the compressor temperature distribution
for cycling operating conditions. One test was defined by maintaining the compressor on for 30
minutes and then off during the next 30 minutes. Results for temperature variation at the cylinder wall
and electrical motor are shown in Figures 7 to 10. Comparing Figures 7 and 8, it can be seen that the
variation of the cylinder temperature is greater than that of the motor during the on-off cycling. Figure
9 shows an interesting behaviour for the suction chamber temperature. While during the on condition
Tsc is lower than the Tie (Figure 10), immediately after the compressor is turned off the temperature of
the suction chamber rises due to heat transfer with discharge chamber and cylinder.
Figure 11 illustrates the temperature field in the solid components during one cycle of the on-off
transient operating condition. Figure 11 (a) shows the temperature of the solid parts immediately after
the compressor is turned on, i.e., after 30 minutes of off period. Figures 11 (a) to 11 (d) show the
increase in compressor temperature that occurs during the on period. This aspect is more pronounced
for solid components near the discharge system. The highest temperatures can be seen in Figure 11
(d), which is the moment when the compressor is switched off. The cooling down process after the
compressor is switched off is illustrated by Figures 11 (e) and (f).

Figure 7. Cycling results for Tw. Figure 8. Cycling results for Tmot.

Tsc

Figure 9. Cycling results for Tsc. Figure 10. Cycling results for Tie.

8
9th International Conference on Compressors and their Systems IOP Publishing
IOP Conf. Series: Materials Science and Engineering 90 (2015) 012068 doi:10.1088/1757-899X/90/1/012068

(a) – 300 min (b) – 310 min (c) - 320 min

(d) – 330 min (e) – 340 min (f) – 350 min

Figure 11. Transient temperature distribution of the solid parts.

5. Conclusions
This paper presented a thermal model developed to predict the temperature distribution of a small
reciprocating compressor under on-off cycling condition. The simulation approach is based on a
coupled solution procedure of three models: (i) a transient lumped model for the compression cycle,
(ii) lumped energy balances for the remaining gas inside the compressor and (iii) finite volume model
to solve heat conduction in the solid components. Some correlations of heat transfer coefficients were
adjusted by using experimental data of temperature distribution. The model was validated through
comparisons with temperature measurements in a period after the compressor was shut down (off
period) and in a period after the compressor was switched on (on period). Good agreement was
verified in both conditions, especially in the period when the compressor was off. The model was also
applied in cycling operating conditions; again showing good agreement with measurements.

Acknowledgments
The present study was developed as part of a technical-scientific cooperation program between the
Federal University of Santa Catarina and EMBRACO. The authors also acknowledge the support
provided by EMBRACO and CNPq (Brazilian Research Council) through Grant No. 573581/2008-8
(National Institute of Science and Technology in Refrigeration and Thermophysics) and CAPES
(Coordination for the Improvement of High Level Personnel).

9
9th International Conference on Compressors and their Systems IOP Publishing
IOP Conf. Series: Materials Science and Engineering 90 (2015) 012068 doi:10.1088/1757-899X/90/1/012068

General and Greek symbols

A Heat transfer area (m2) p Pressure (Pa)
C Constant (-) Pr Prandtl number (-)
c Specific heat (J/kg.K) Q̇ Heat transfer rate (W)
f Frequency of operation (Hz) T Temperature (K)

3
h Enthalpy (J/kg) Gas volume (m )
H Heat transfer coefficient (W/m K)2
Ẇind Indicated Power (W)
k Thermal conductivity (W/m.K)  Thermal diffusivity (m2/s)
m Mass (kg)  Coefficient of thermal expansion (1/K)

m Mass flow rate (kg/s)  Specific volume (m3/kg)
N Time step number (-)  Dynamic viscosity (Pa.s)
Subscripts
amb External ambient ele Electrical pc Piston cylinder gap
b Backflow G Gas inside cylinder sol General solid surface
bea Bearings ie Internal environment sc Suction chamber
ch General gas chamber in Inlet sm Suction muffler
dis Discharge leak Leakage suc Suction
dc Discharge chamber mot Motor w Cylinder wall
dm Discharge muffler oil Oil
dl Discharge line out Outlet

References

[1] Hermes C J L and Melo C 2008 A first-principles simulation model for the start-up and cycling
transients of household refrigerators Int. J. Refrig. 31 1341-57
[2] Dutra T and Deschamps C J 2013 Experimental characterization of heat transfer in the
components of a small hermetic reciprocating compressor Appl. Therm. Eng. 58 499-510
[3] Ooi K T 2003 Heat transfer study of a hermetic refrigeration compressor Appl. Therm. Eng. 23
1931-1945
[4] Ribas Jr F A, Deschamps C J, Fagotti F, Morriesen A and Dutra T 2008 Thermal analysis in
reciprocating compressors – a critical review Proc. Int. Compressor Eng. Conf. (Purdue) Paper
1907
[5] Todescat M L, Fagotti F, Prata A T and Ferreira R T S 1992 Thermal energy analysis in
reciprocating hermetic compressor Proc. Int. Compressor Eng. Conf. (Purdue) 1419-28
[6] Raja B, Seckar S J, Lal D M and Kalanidhi A 2003 A numerical model for thermal mapping in
hermetically sealed reciprocating compressor Int. J. Refrig. 26 652-58
[7] Ribas Jr F A 2007 Thermal analysis of reciprocating compressors Proc. Int. Conf. on
Compressors and their Syst. (London) 277-87
[8] Annand W J D 1963 Heat transfer in the cylinders of reciprocating internal combustion engines
Proc. I. Mech. Eng. 117 973-96
[9] NIST 2007 Refprop – Reference fluid thermodynamic and transport properties (NIST)
[10] Deschamps C J, Prata A T and Ferreira R T S 2000 Modelling of Turbulent Flow Through
Radial Diffusers J. of Brazilian Society of Mechanical Sciences n 22 pp 31–41
[11] Ferreira R T S and Lilie D E B 1984 Evaluation of the Leakage Through the Clearance Between
Piston and Cylinder in Hermetic Compressor Proc. Int. Compressor Eng. Conf. (Purdue) pp 1–6
[12] ANSYS Inc 2011 User’s guide v.12.0.0.0 (ANSYS)
[13] Catton I 1978 Natural convection in enclosures Proc. Int. Heat Transfer Conf. (Toronto) 6 13-
31
[14] Drost R T and Quessada J F 1992 Analytical and experimental investigation of a scroll
compressor lubricating system Proc. Int. Compressor Eng. Conf. (Purdue) 551-60

10