i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 1 7 2 2 e1 7 3 3

Available online at www.sciencedirect.com

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

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

Simplified steady-state modeling for hermetic compressors

with focus on extrapolation

Wenhua Li*
Modeling, Analysis, Simulation and Computation (MASC), Carrier Corporation, Syracuse, NY 13057, USA

article info abstract

Article history: A simplified physical model for hermetic scroll and reciprocating compressors was
Received 10 August 2011 developed based on thermodynamic principles and was calibrated with experimental data
Received in revised form from publication. The model is able to calculate mass flow rate, power and discharge
16 January 2012 temperature. The extrapolation capability of the proposed model was investigated by
Accepted 9 March 2012 extending to rotary compressor and extrapolating the achieved model to predict the mass
Available online 17 March 2012 flow rate, power and discharge temperature at the conditions outside the given data range.
In an addition, the model was studied by predicting mass flow rate, power and discharge
Keywords: temperature for other refrigerants while experimental data of one refrigerant is used for
Scroll compressor calibration to get fitting parameters. The model exhibited very good extrapolation
Reciprocating compressor capability.
Model ª 2012 Elsevier Ltd and IIR. All rights reserved.
Volumetric efficiency
Extrapolation

Modélisation simplifiée en régime permanent des

compresseurs hermétiques avec l’accent mis sur
l’extrapolation
Mots clés : Compresseur à spirale ; Compresseur à piston ; Modèle ; Rendement volumétrique ; Extrapolation

1. Introduction system modeling. Under unsteady condition, the compressor

transients associated with changes in boundary condition are
The modeling of compressors plays a very important role in small compared to those of heat exchangers, and therefore,
the simulation of refrigeration or air conditioning system. the compressor could be modeled as a steady state process
Whether in the refrigeration system simulation of steady state (Negrão et al., 2011). A complex model of refrigeration system
or unsteady state, an accurate compressor model which can is coupled with a lot of components and the iteration is not
provide reasonable and relatively accurate extrapolation always within the range in which compressor model is built.
outside the compressor envelope is also essential for the This requires the extrapolation capability of a compressor

* Tel.: þ1 785 317 2624.

E-mail address: [email protected].
0140-7007/$ e see front matter ª 2012 Elsevier Ltd and IIR. All rights reserved.
doi:10.1016/j.ijrefrig.2012.03.008
 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 1 7 2 2 e1 7 3 3 1723

Nomenclature k refrigerant isentropic coefficient

Tevap saturation evaporating temperature [ C]
hv volumetric efficiency
Tcond saturation condensing temperature [ C]
Pr pressure ratio of discharge pressure with respect
Vsuc volume flow rate [m3 s1]
to suction pressure
Ncomp compressor rotation speed [rpm]
C clearance fraction
Vdisp compressor displacement [kg s1]
ma actual mass flow rate [kg s1]
c1, c2, c3 fitting parameters for discharge temperature
mi theoretical mass flow rate [kg s1]
a1, a2, a3 fitting parameters for shaft power model
hcl clearance volumetric efficiency
b1, b2 fitting parameters for mass flow rate model
n polytropic index
Wloss constant part of power loss [W]
Pdis discharge pressure [kPa]
hsuc, hdis enthalpy at suction and discharge [kJ kg1]
Psuc suction pressure [kPa]
Tshell compressor shell temperature [ C]
dp suction pressure drop [kPa]
Wt internal compressor power [W]

model so as to keep the iteration toward the right direction, experimental data. Navarroa et al. (2007) presented a physical
which is very important when Newton Raphson method is model which needs 10 unknown parameters to calculate
accepted to solve the non-linear equations. As next genera- volumetric efficiency and isentropic efficiency, and considers
tion of refrigerants is coming to the table (Calm, 2008) to have the influences from suction valve, discharge valve, leakage,
GWP (global warming potential) less than 150, the investiga- etc. Their purpose is to introduce a useful physical model
tion and upgrade of existing refrigeration and air conditioning which can be used to predict compressor performance under
system is in much need of the extrapolation capability of an operating conditions which are not tested or with different
efficient compressor model. refrigerants. However, they do not further explore the capacity
A detailed physical model (Chen, 2000; Jovane, 2007) which of extrapolation of their models under different conditions or
attempts to model the most important physical compressor with different refrigerants in the paper. Cuevas et al. (2010)
characteristics by the fundamental equations (conservation of discussed their semi-empirical compressor models under
mass, energy and momentum, etc.) can deliver considerable extended operating conditions. Their validation is still based
useful information about the principle which the compressor on the parameters fitting with experimental data including the
is working on, but it requires numerous inputs from manu- data under extreme conditions and the capability of extrapo-
factures. Also the computation is very time-consuming. It is lation which should use limited experimental data for
not suitable for modeling of refrigeration system but it is good parameter fitting is not explored. Jähnig et al. (2000) explored
for the optimization design of compressors. the extrapolation capability of the semi-physical models they
A steady state modeling of a complex refrigeration system developed for a small hermetic compressor. The investigation
or a transient system simulations generally avoid the detailed of extrapolation is limited to predicate mass flow rate and
physical compressor model. However, the polynomial fitting power at the higher condensing temperature or lower evapo-
such as 10-coefficient polynomial which ARI standard 540-91 rating temperature outside of the range of experimental data.
presents doesn’t necessarily provide reliable interpolations or In this work, the author tries to present a simplified physical
extrapolations for conditions not represented in the model of mass flow rate, power and discharge temperature for
compressor map, has no any physical meaning and should hermetic reciprocating and scroll compressors by introducing
not be used (Jähnig et al., 2000). the characterizing variables like the volumetric efficiency,
A semi-physical compressor model is preferred whether in isentropic efficiency. More importantly, the work will investi-
steady-state simulation or transient simulation of whole gate the extrapolation capability of the physical model in more
refrigeration system. The kind of model seeks to build empirical scenarios such as extreme conditions, different refrigerants,
and simplified characterized relations based on physical other kinds of compressors, and etc. A validation work is first
phenomena from the fundamental physical laws. The simpli- performed and the sources of experimental data come from
fied physical unknowns are acquired through the fitting with published literature, which cover a few reciprocating compres-
limited experimental data. Therefore, the model has not only sors and a few scroll compressors, and involve several kinds of
high accuracy inside the compressor data envelope, but also refrigerants such as R410A, R404A, R22, R134a and R12. By using
provides good extrapolation outside the envelope even with these sources, the author aims to build the semi-empirical mode
a limited number of experimental data. The advantage often based on first-hand experimental data to avoid any error by
makes numerical iteration in the system modeling avoid toward using the 10-coefficients data from manufactures.
wrong direction and get the system simulation converged
because the system computation may go anywhere unexpected.
Establishing a semi-physical compressor model is of 2. Formulation of simplified compressor
interest in a lot of literature (Negrão et al., 2011; Jähnig et al., model
2000; Kim and Bullard, 2002; Navarroa et al., 2007; Cuevas
et al., 2010; Winandy et al., 2002). However, most of these Some literature (Negrão et al., 2011; Navarroa et al., 2007;
literature only discussed and validated the capability (such as Cuevas et al., 2010; Winandy et al., 2002) developed the
relative error, RMS error and etc.) of their models fitted to simplified model of a scroll compressor based on a lot of
1724 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 1 7 2 2 e1 7 3 3

detailed information of experimental data such as machine for best C when experimental data is used for fitting. With the
suction temperature, scroll suction temperature, scroll suction pressure drop dp and the internal leakage, we refor-
discharge temperature, machine discharge temperature, scroll mulated Eq. (3) as
suction pressure drop, scroll discharge pressure difference,  1=k 
suction pressure, discharge pressure, ambient temperature Pdis
hv ¼ b1 þ b2 (4)
and shaft power. Here we’d like to develop the semi-empirical Psuc ð1  dpÞ
model only with experimental data as shown in Fig. 1. Such Where we have three unknown parameters b1, b2 and dp in Eq.
models can be used to predict mass flow rate, electric power (4), which need experimental data for fitting.
and discharge temperature when inputs of suction tempera-
ture, suction pressure and discharge pressure are given. 2.2. Model of electric power

2.1. Model of mass flow rate In ASHRAE Toolkit (Bourdhouxhe et al., 1999), the calculation
of compressor shaft power is proposed as follows:
For a compressor model, the volumetric efficiency is used to
calculate the suction volume flow rate. The actual volumetric W ¼ ð1 þ aÞWt þ Wloss (5)
efficiency is defined as the ratio of actual volume of refrigerant
entering compressor to the geometric displacement of the Where Wt is the internal compressor power, Wloss is the
compressor, which can be written as: constant loss due to electromechanical loss, and a is the ratio
of electromechanical losses to Wt.
V_ suc It is interesting to notice Negrão et al. (2011) defines the
hv ¼ (1)
Ncomp Vdisp
shaft power as
Where, Ncomp is compressor speed, Vdisp is compressor
displacement and Vsuc is volume flow rate. W ¼ Wisen =hall þ Wloss (6)
Physically, since the suction valve throttling, suction gas
heating, the clearance volume, and internal leakage, the
Where hall is overall efficiency. The overall efficiency is actu-
actual volumetric efficiency is less than 1.
ally the product of efficiency of the motor, mechanical effi-
Threlkeld (1962) viewed the compression and expansion
ciency and isentropic efficiency in a compressor. Wisen is the
processes in a reciprocating compressor as a polytropic
isentropic compression power defined as
process instead of isentropic process and proposed the
following Eq. (2) to calculate the actual volumetric efficiency,
 k1 

 k Pdis k
hv ¼ 1  C Pr1=n  1 (2) Wisen ¼ Psuc V_ suc 1 (7)
k1 Psuc
Where n is polytropic component.
Jähnig et al. (2000) and Cabello et al. (2005) introduced
Negrão et al. (2011) assumed the overall efficiency as constant.
suction pressure drop for compressor mass flow model
Jähnig et al. (2000) presented the calculation of the shaft
because of its significant effect on mass flow rate although
power as
small. Therefore, the actual volume efficiency in the paper of
Jähnig et al. (2000) has the formulation as
W ¼ Wisen =hall (8)
 1=n 
Pdis
hv ¼ 1  C 1 (3)
Psuc ð1  dpÞ
Here, Jähnig et al. (2000) didn’t consider the constant power
Jähnig et al. (2000) thought of the relation of dependence loss but assumed the variable overall efficiency instead of
between C and n, so he took n as isentropic coefficient k to look constant,

Fig. 1 e Conceptual schema of the compressor model.

 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 1 7 2 2 e1 7 3 3 1725

hall ¼ d þ e$expðf $Psuc Þ (9) compressor is also performed hereafter by aid of experimental
data. The parameters are determined by minimizing the
His analysis was based on the small reciprocating compressor. following objectives function to obtain optimized agreement
By observing and investigating the experimental data of with the measured data at given data points:
reciprocating and scroll compressor, we rather believes Eq. (6)
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
is a representing formula to calculate compressor shaft power Xmexp  mcal 2
and it is better to view the overall efficiency as variable instead dm ¼ min ;
mexp
of constant. The overall efficiency is a combination of isen- sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi
XWexp  Wcal 2
tropic efficiency, mechanical efficiency and motor efficiency. dW ¼ min ;
The overall efficiency is somehow dependant on the suction Wexp
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
pressure and discharge pressure. The proposed overall effi- XTexp  Tcal 2
dT ¼ min ; T is discharge temperature
ciency can be calculated as Texp
(15)
1=hall ¼ a1 þ a2 =Psuc þ a3 =Pdis (10)
Where Levenberg Marquardt method is used to solve the
least square problem and estimate the parameters for
Therefore, a complete calculation of compressor shaft
models of mass flow rate, electric power and discharge
power is proposed as
temperature.
 k1 
k Pdis k
W ¼ Psuc V_ suc 1 ða1 þ a2 =Pdis þ a3 =Psuc Þ þ Wloss
k1 Psuc
(11) 3. Validation of the models
Here we have four parameters a1, a2, a3 and Wloss to be
Experimental data which are used in the paper for model
calculated by fitting experimental data.
validation are listed in Table 1. They are all from public liter-
ature and from 6 different journal papers. These first-hand
2.3. Model of discharge temperature
test data were performed in different conditions and covered
four different refrigerant types of R12, R134a, R22 and R410A.
Generally, the energy balance for the entire compressor is
Data set 1e4 were for scroll compressor while data set 5e8
depicted by
were for reciprocating compressor.
The simulation results are depicted in Fig. 2 and Fig. 3. The
_ dis  hsuc Þ þ Q_ amb
W ¼ mðh (12) relative error on the mass flow rates, shaft power is within 5%
and absolute error on discharge temperature is within 5  C for
Where, Q_ amb is the heat transfer rate to the ambient, which
scroll compressor (Fig. 2) and for reciprocating compressor
can be calculated as:
(Fig. 3). The small error shows the calculated mass flow rates,
shaft power and discharge temperature are in good agreement
Q_ amb ¼ UAshell ðTshell  Tamb Þ (13) with measured values for the two kinds of compressors. It also
tells that the models are able to be employed to simulate the
The most difficult part is how to calculate Tshell, which may
scroll compressor although they are derived first from recip-
be related to a lot of factors such as pressure ratio, suction
rocating compressor.
pressure, discharge pressure, or discharge temperature. Kim
As further evidence, Table 2 is provided to check the
and Bullard (2002) suggested a liner relationship with the
simulation performance of the max. and min. relative error of
discharge temperature as Tshell¼a þ bTdis. Duprez et al. (2010)
mass flow rate, electric power and absolute error on the
considered shell temperature is related to condensing satu-
discharge temperature. Clearly, from the table, the relative
ration temperature and evaporating saturation temperature
error on the mass flow rate varies between 5.27% and 5.36%.
and chose Tshell ¼ a þ bTevap þ cTcond.
The relative error on the shaft power varies between 4.45%
A good correlation should be a good compromise between
and 4.14%. The absolute error on the discharge temperature
simplicity, reduced number of unknown parameters and
varies between 2.28  C and þ3.78  C. Max. RMS errors for
results accuracy, extrapolation capability. Our observance and
mass flow rate, shaft power and discharge temperature are
analysis of experimental data from publications which are
2.96%, 2.29% and 1.68  C, respectively.
over a wide range of operating conditions suggests the
calculation of shell temperature as

Pdis 4. Extension to other type of compressor

Tshell ¼ c1 þ c2 þ c3 =Pdis (14)
Psuc
Which tells the dependence of the shell temperature on the Jovane (2007) studied a novel rotary compressor and provided
pressure ratio is not enough to capture the physical, an indi- detailed experimental results in his thesis. The experimental
vidual discharge pressure is also needed to do better predic- data for refrigerant R410A has 16 data points and it covers the
tion of shell temperature. range of suction pressure from 600 kPa to 1000 kPa and the
In short, although models (Eq. (4), Eq. (11) and Eq. (14)) range of discharge pressure from 1500 kPa to 3400 kPa. Below
proposed in this article are derived from reciprocating (Fig. 4) are the simulation results with mass flow rate, shaft
compressor, the validation and exploration for scroll power and discharge temperature by using the proposed
1726 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 1 7 2 2 e1 7 3 3

Table 1 e List of sources of experimental data.

Name Source of Data Refrigerant Type Psuc[kPa] Pdis[kPa]

Data set 1 Chen, 2000 R410A Scroll 570e1100 1650e2700

Data set 2 Winandy et al., 2002 R22 Scroll 240e730 1190e2530
Data set 3 Cuevas and Lebrun, 2009 R134a Scroll 510e2040 1000e3910
Data set 4 Winandy and Lebrun, 2002 R22 Scroll 200e520 1680e2900
Data set 5 Compressor X, Negrão et al., 2011 R134a Recipe 66e164 1100e1700
Data set 6 Compressor Y, Negrão et al., 2011 R134a Recipe 66e164 1100e1700
Data set 7 Dataset B9A, Jähnig et al., 2000 R12 Recipe 100e170 900e1200
Data set 8 Dataset A4, Jähnig et al., 2000 R134a Recipe 90e180 950e1500

Fig. 2 e Comparison of measured and computed data for scroll compressor (dataset 1e4).
 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 1 7 2 2 e1 7 3 3 1727

Fig. 3 e Comparison of measured and computed data for reciprocating compressor (dataset 5e8).

Table 2 e Fitting error for the models.

Mass flow rate Electric power Discharge temperature

Min Rel. Max. Rel. RMS Min Rel. Max. Rel. RMS Min Abs. Max. Abs. RMS
error % error % error % error % error % error % error [ C] error [ C] error [ C]

Data set 1 4.30 5.36 2.96 4.45 3.71 2.29 1.62 2.94 1.12
Data set 2 3.26 4.38 1.65 3.36 2.29 1.23 1.87 2.52 1.15
Data set 3 1.80 2.66 1.13 4.05 4.14 2.26 0.42 2.58 0.74
Data set 4 2.48 3.24 1.87 0.80 0.49 0.41 1.40 1.50 0.86
Data set 5 4.37 2.61 2.07 1.72 1.80 1.02 1.80 2.82 1.59
Data set 6 5.27 3.39 2.73 0.84 1.09 0.56 2.27 3.78 1.68
Data set 7 1.87 1.69 1.18 0.50 0.67 0.38 2.28 2.92 1.59
Data set 8 2.81 2.37 1.28 1.51 1.01 0.77 N/A N/A N/A
1728 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 1 7 2 2 e1 7 3 3

models in Section 2. The relative error on mass flow rate and The results tell us that the models proposed in Section 2 are
electric power is between 1.14% and 0.98%, 1.07% and also able to be employed to model the type of rotary compressor.
0.36%, respectively. The absolute error on discharge temper-
ature is between 0.33  C and 0.48  C. It is not clear why the
presented models behave much better for the rotary 5. Extrapolation to outside conditions
compressor because the accuracy of experimental data may
be a good contribution. However, from the point of view of To check the extrapolation capabilities of the proposed models,
modeling, by examining Eq. (4) and Eq. (11), we can see the Dataset 2 (scroll compressor) is taken for the study. In Table 3,
constant terms of dp and Wloss may be more close to constant the data with suction pressure from 240 kPa to 620 kPa and
in reality when a very high efficient rotary compressor is correspondingly the discharge pressure from 1190 kPa to
modeled, resulting in a more stable solution. 1940 kPa are arranged for parameters fitting. The rest data with

Fig. 4 e Comparison of measured and computed data for rotary compressor.

 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 1 7 2 2 e1 7 3 3 1729

Table 3 e Comparison of extrapolation capability to higher condensing pressure and higher evaporating pressure for the
models (dataset 2).
Tsuc Psuc Pdis Rel. error  mass Rel. error  electric Abs. error  discharge
[ C] [kPa] [kPa] flow rate power temperature [K]

Used for model parameters 25 240 1190 1.30% 1.63% 0.97

fitting 25 350 1190 0.93% 1.06% 0.75
25 500 1190 0.97% 0.49% 0.38
25 620 1190 0.07% 0.16% 0.43
25 300 1350 3.01% 0.97% 0.36
25 420 1350 0.35% 0.47% 0.79
25 580 1350 0.55% 0.70% 0.71
25 300 1530 2.83% 1.49% 1.97
25 350 1530 2.21% 0.25% 1.00
25 420 1530 0.27% 0.05% 1.14
25 620 1530 0.19% 1.42% 0.13
25 350 1730 2.54% 0.86% 1.70
25 500 1730 1.13% 0.43% 1.29
25 420 1940 0.33% 1.13% 1.53
25 500 1940 0.18% 0.39% 1.95
25 580 1940 0.65% 0.75% 1.89
Not used for model parameters 18.3 630 2150 3.35% 1.39% 0.03
\fitting 25 500 2170 1.91% 1.41% 1.31
25 620 2170 1.57% 1.80% 0.87
25 620 2430 5.15% 0.85% 0.37
25 580 2530 0.55% 3.08% 0.74
25 680 1360 0.52% 0.69% 0.46
25 680 1940 0.92% 2.17% 1.21
25 680 2430 2.25% 1.70% 0.79
25 730 1190 0.27% 1.92% 0.97
25 730 1730 0.26% 2.08% 1.14
25 730 2170 1.10% 2.65% 1.52
25 730 2533 2.27% 1.70% 1.02

characteristics of suction pressure higher than 620 kPa or In Fig. 5(a), there is one experiment data (Tevap ¼ 7  C,
discharge pressure higher than 1940 kPa are purposely picked Tcond ¼ 60  C) whose mass flow rate is even smaller than the
for comparison with predicted results. We aim at looking at the predicted one at Tcond ¼ 70  C. But the relative error on the
extrapolation capabilities of the models at higher suction condition is 5.15%, which is acceptable. Furthermore, the
pressure or higher discharge pressure. relative errors of predicted mass flow rate for Psuc ¼ 500 kPa
By examining the relative error of mass flow rate and and 580 kPa in Table 3 are both less than 2%, indicating
shaft power, the absolute error of discharge temperature in possible experimental error for the experimental data at
Table 3, the models all exhibit very good extrapolation Tevap ¼ 7  C and Tcond ¼ 60  C. From the figure, we can see the
capabilities at higher suction pressure or higher discharge model of mass flow rate (Eq. (4)) performs very well outside the
pressure. At the predicted conditions which are outside the range of limit experimental data in capturing the physical
fitting data range, relative error of mass flow rate and shaft characteristics: mass flow rate is mainly dependent on the
power is within 4%, and absolute error of discharge temper- compressor inlet pressure and less relies on the discharge
ature is within 2 K. pressure. In Fig. 5(b) and (c), the electric power and discharge
To further explore the extrapolation capabilities, we used temperature are plotted verse evaporating temperature. In
the parameters which have been achieved in the above anal- Fig. 5(b), as evaporating temperature increase, the shaft power
ysis to calculate the mass flow rate, power, discharge per unit mass flow rate starts to be influenced less by
temperature at lower suction pressure to 164 kPa (30  C condensing temperature. This is understandable because the
saturation temperature), higher suction pressure to 910 kPa difference of pressure ratio between two different condensing
(20  C saturation temperature), lower discharge pressure temperatures is becoming smaller as evaporating tempera-
1044 kPa (25  C saturation temperature) and higher discharge ture increases. The trend is also well-captured by the
pressure 2998 kPa (70  C saturation temperature). The results proposed model of shaft power (Eq. (11)) through examining
are shown in Fig. 5 (values in the legend represent condensing the extrapolation outside the fitting data range in Fig. 5(b). In
temperature). The dotted points in the figure are from exper- Fig. 5(c), the experimental data indicated that as evaporating
iment and thin lines are from calculated results. Each thin line temperature increases or condensing temperature decreases,
is corresponding to a condensing temperature and results at discharge temperature decreases. The trend is almost
several different condensing temperatures such as 25  C, consistent with that of power per unit mass flow rate illus-
40  C, 55  C and 70  C are plotted in the figure. Although more trated in Fig. 5(b). The proposed model of discharge temper-
test data are plotted, only some test data pointed out in Table 3 ature (Eq. (14)) also captured the trend even outside the
are used to achieve the fitting parameters for each model. experimental data used for fitting.
1730 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 1 7 2 2 e1 7 3 3

Overall, from these figures (Fig. 5(a)e(c)) it can be seen that absolute error of discharge temperature is within 1.5 K after
the predicted results by using the proposed models are in good extrapolation to higher suction pressure (saturation temper-
agreement with the trend of experimental data even outside ature is 5.3  C higher) higher and higher discharge pressure
the range of experimental data used for fitting. The physical (saturation temperature is 12  C higher), one can defer that
characteristics of mass flow rate, shaft power and discharge when suction saturation temperate is 10  C or discharge
temperature mainly related to condensing temperature and saturation temperature is 15  C away from experimental data
evaporating temperature are well-captured by the proposed which is used for parameters fitting, extrapolation of the
models. Based on the fact from Table 3 that relative error of models are still within a good accuracy (about 5%).
mass flow rate and electrical power are within 3.5% and

6. Extrapolation to different ambient

temperature

The models we presented calculate mass flow rate, shaft

power and discharge temperature by considering the
compressor suction pressure, temperature and discharge
pressure as inputs. The models don’t take ambient tempera-
ture as direct input to predict mass flow rate and shaft power.
The ambient temperature is a direct input when calculating
discharge temperature.
In the following analysis, we will use data from the paper of
Jähnig et al. (2000) to examine the extrapolation capability of
mass flow rate model and shaft power model when ambient
temperature changes. Since the data doesn’t provide
discharge temperature, we will not cover discharge tempera-
ture model. In Table 4, calorimeter tests were performed at
three ambient temperatures of 32.2  C, 43.3  C and 15.6  C. At
15.6  C, the condensing temperatures were much lower than
ones at 32.2  C.
In this study, data at ambient temperature of 32.2  C is used
to fit parameters for the models and then the achieved
parameters for each model are employed to predict the mass
flow rate and shaft power at ambient temperature of 43.3  C
and 15.6  C. From Table 4 we can see that simulation results
are in good agreement with measured data even when
condensing temperature changes a lot, which is due to
ambient temperature (see rows at ambient temperature of
15.6  C). The good result of prediction tells us that presented
mass flow model is able to handle minor change of mass flow
rate due to the variation of ambient temperature (as Jähnig
et al. (2000) and Negrão et al. (2011) described, the mass flow
rate generally is not sensitive to ambient temperature).
Although the condensing temperature at ambient 15.6  C were
much lower than ones at 32.2  C which was used for calibra-
tion to get parameters, the relative error of extrapolating to
15.6  C for power range is less than 4.4%. Jähnig et al. (2000)
used his model to study the data set and got relative errors
for power range from 8% to 12%. He thought the inaccuracies
are confounded in this case and didn’t give further analysis.
By examining the Table 4, it is evident that when ambient
temperature is away from model setup ambient temperature
by 10  C, the relative error of electric power is within 2% and
the error is 4.4% when it is away from the baseline ambient
temperature by 17  C. That indicates that the extrapolation
capacity of the models presented can be 20  C away from the
baseline ambient temperature if accuracy of relative error
Fig. 5 e (a) Extrapolation of mass flow rate model for within 5% is the target of prediction.
a scroll compressor. (b) Extrapolation of electrical power As further analysis of extrapolation for the reciprocating
model for a scroll compressor. (c) Extrapolation of compressor, we used the parameters achieved by fitting the
discharge temperature model for a scroll compressor. data of 9 points to calculate the mass flow rate and shaft
 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 1 7 2 2 e1 7 3 3 1731

Table 4 e Comparison of relative error for the models (data from the paper of Jähnig et al., 2000).
Ambient Tevap Tcond Measured mass Rel. error  mass Measured electric Rel. error 
temperature [ C] [ C] flow rate [g s1] flow rate power [W] electric power
[ C]

Used for model 32.2 28.9 54.4 1.148 0.64% 118.71 0.35%
parameters fitting 32.2 28.9 48.9 1.184 0.26% 115.95 0.13%
32.2 28.9 43.3 1.225 0.49% 112.85 0.66%
32.2 23.3 54.4 1.511 1.59% 136.03 0.09%
32.2 23.3 48.9 1.586 0.59% 132.44 0.03%
32.2 23.3 43.3 1.62 0.25% 128.02 0.25%
32.2 17.8 54.4 1.97 1.61% 153.73 0.44%
32.2 17.8 48.9 2.039 0.26% 148.70 0.14%
32.2 17.8 43.3 2.113 1.33% 143.12 0.36%
Not used for model 43.3 28.9 54.4 1.091 0.19% 118.71 1.81%
parameters fitting 43.3 28.9 48.9 1.144 1.00% 115.99 1.48%
43.3 23.3 54.4 1.49 1.25% 136.14 1.22%
43.3 23.3 48.9 1.502 0.67% 132.25 1.42%
15.6 28.9 26.7 1.376 1.90% 102.20 4.39%
15.6 28.9 23.9 1.436 1.34% 100.28 4.13%
15.6 23.3 26.7 1.846 0.86% 113.50 3.24%
15.6 23.3 23.9 1.885 2.11% 110.93 3.62%

Fig. 6 e (a) Extrapolation of mass flow rate model for Fig. 7 e (a) Comparison of measured and computed mass
a reciprocating compressor. (b) Extrapolation of electrical flow rate for ZB26KCE-TFD (R22). (b) Comparison of measured
power model for a reciprocating compressor. and computed electric power for ZB26KCE-TFD (R22).
1732 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 1 7 2 2 e1 7 3 3

power with evaporating temperature from 40  C to 5  C and is 3500 rpm. The experimental data were recorded at ambient
condensing temperature from 23.9  C to 65  C. Fig. 6(a) and (b) 35  C with compressor suction temperature of 18.3  C. The
are the results. Some test points are overlapped in the figure. data doesn’t provide the measurement of discharge
The predicted trend of mass flow rate and shaft power is well temperature.
aligned with that of experimental data. The trend is very First, the data for refrigerant R22 is used for parameters
similar as that in Fig. 5(a) and Fig. 5(b) for a scroll compressor. fitting and for the baseline analysis. Fitting parameters for the
It should be noted that, from the figure, test data at models: mass flow rate, b1 ¼ 1.028, b2 ¼ 0.0272, dp ¼ 2.6e-
Tcond ¼ 23.9  C and 26.7  C are deviated from the corre- 5 kPa; shaft power, a1 ¼ 1.366, a2 ¼ 2327.7 kPa, a3 ¼ 40.3 kPa,
sponding extrapolation line, which do not indicate bad Wloss ¼ 3.196 W. Fig. 7(a) and (b) are the comparisons of
extrapolation. All extrapolation curves are plotted based on calculated mass flow rate, shaft power with measured ones.
the same suction temperature 32.2  C but the corresponding The models can be fitted to experimental data within 5%
ambient temperature at Tcond ¼ 23.9  C and 26.7  C should be (actual max. relative error is less than 2.3%).
used as the input of suction temperature for the numerical As an extrapolation, the above parameters achieved using
calculation. This is the reason why the presented model the measured data of refrigerant R22 are employed to predict
can perform a good extrapolation to different ambient the performance of mass flow rate and shaft power for
temperature. refrigerant R404A. Fig. 8 is the illustration of the simulation
results. The relative error on mass flow rate for R404A is
within 5% and the error on shaft power for R404A is within 5%
7. Extrapolation to different refrigerants except one point which has relative error of 8%. The results
are what we expected because the models we presented are
A Copeland Scroll compressor ZB26KCE-TFD is used to study irrelevant to refrigerant type and they are developed in the
the capability of the models in extrapolating to different context of semi-empirical relation.
refrigerants. Its displacement is 12 m3 h1 and nominal speed

8. Conclusion

This paper presents a detailed analysis of semi-empirical

methods to calculate mass flow rate, shaft power and
discharge temperature for reciprocating compressor and scroll
compressor. The analysis is focused on their extrapolation
capabilities while the models are validated within 5% relative
error of calculated mass flow, shaft power and discharge
temperature compared to experimental ones. The study of
exploration includes extending to rotary compressor, extrap-
olating the achieve models to predict the mass flow rate, shaft
power and discharge temperature at the conditions outside the
given data range, extrapolating to other ambient temperature,
predicting mass flow rate, shaft power, discharge temperature
for other refrigerants when experimental data for one refrig-
erant is used to get fitting parameters. The proposed models
exhibit a very good capability of extrapolation.

references

Bourdhouxhe, J.P., Grodent, M., Lebrun, J., 1999. HVACTOOLKIT:

a Toolkit for Primary HVAC System Energy Calculation.
ASHRAE, ISBN 1-883413-73-7.
Cabello, R., Navarro, J., Torrella, E., 2005. Simplified steady-state
modeling of a single stage vapor compression plant. Model
development and validation. Appl. Thermal Eng. 25, 1740e1752.
Calm, J.M., 2008. The next generation of refrigerants e historical
review, considerations, and outlook. Int. J. Refrigeration 31,
1123e1133.
Chen, Y., 2000. Mathematical Modeling of Scroll Compressors,
Ph.D. thesis, Ray W. Herrick Laboratories, Purdue University.
Cuevas, C., Lebrun, J., 2009. Test and modeling of a variable speed
Fig. 8 e (a) Comparison of measured and predicted mass flow
scroll compressor. Appl. Thermal Eng. 29 (2e3), 469e478.
rate for ZB26KCE-TFD (R404A) using the model fitted for R22. Cuevas, C., Lebrun, J., Lemort, V., Winandy, E., 2010.
(b) Comparison of measured and predicted electric power for Characterization of a scroll compressor under extended
ZB26KCE-TFD (R404A) using the model fitted for R22. operating conditions. Appl. Thermal Eng. 30, 605e615.
 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 1 7 2 2 e1 7 3 3 1733

Duprez, M.E., Dumont, E., Frère, M., 2010. Modeling of scroll compress Negrão, C.O.R., Erthal, R.H., Andrade, D.E.V., Silva, L.W., 2011.
ors e improvements. Int. J. Refrigeration 33 (2010), 721e728. A semi-empirical model for the unsteady-state simulation
Jähnig, D.I., Reindl, D.T., Klein, S.A., 2000. A semi-empirical of reciprocating compressors for household refrigeration
method for representing domestic refrigerator/freezer applications. Applied Thermal Engineering 31 (6e7),
compressor calorimeter test data. ASHRAE Trans. 106 1114e1124.
Jovane, M., 2007. Modeling and Analysis for a Novel Rotary, Ph.D. Threlkeld, J.L., 1962. Thermal Environmental Engineering.
dissertation, Purdue University. Prentice-Hall, Englewood Cliffs, NJ.
Kim, M.-H., Bullard, C.W., 2002. Thermal performance analysis of Winandy, E., Saavedra, O.C.S., Lebrun, J., 2002. Experimental
small hermetic refrigeration and air-conditioning analysis and simplified modeling of a hermetic scroll
compressors. JSME Int. J. Ser. B 45 (4). refrigeration compressor. Appl. Thermal Eng. 22 (2), 107e120.
Navarroa, E., Granrydb, E., Urchuegu, J.F., Corber, J.M., 2007. A Winandy, E.L., Lebrun, J., 2002. Scroll compressors using gas and
phenomenological model for analyzing reciprocating liquid injection: experimental analysis and modeling. Int. J.
compressors. Int. J. Refrigeration 30, 1254e1265. Refrigeration 25 (8), 1143e1156.