Purdue University

Purdue e-Pubs
International Refrigeration and Air Conditioning
School of Mechanical Engineering
Conference

2008

Simulation and Validation of a R404A/CO2

Cascade Refrigeration System
Jonathan Michael Winkler
University of Maryland

Vikrant Aute
University of Maryland

Reinhard Radermacher
University of Maryland

Doron Shapiro
Ingersoll-Rand Climate Control

Follow this and additional works at: http://docs.lib.purdue.edu/iracc

Winkler, Jonathan Michael; Aute, Vikrant; Radermacher, Reinhard; and Shapiro, Doron, "Simulation and Validation of a R404A/CO2
Cascade Refrigeration System" (2008). International Refrigeration and Air Conditioning Conference. Paper 942.
http://docs.lib.purdue.edu/iracc/942

This document has been made available through Purdue e-Pubs, a service of the Purdue University Libraries. Please contact [email protected] for
additional information.
Complete proceedings may be acquired in print and on CD-ROM directly from the Ray W. Herrick Laboratories at https://engineering.purdue.edu/
Herrick/Events/orderlit.html
 2273, Page 1

Simulation and Validation of a R404A/CO2 Cascade Refrigeration System

Jon WINKLER1,*, Vikrant AUTE2, Reinhard RADERMACHER3

Center for Environmental Energy Engineering
Department of Mechanical Engineering, University of Maryland
College Park, MD 20742 USA
1,2
Tel: 301-405-8726, 3Tel: 301-405-5286, Fax: 301-405-2025
Email: [email protected], [email protected], [email protected]

Doron SHAPIRO
Ingersoll-Rand Climate Control
Bridgeton, MO 63044 USA
Tel: 314-298-4730, Fax: 314-298-4765
Email: [email protected]

*Corresponding Author.

ABSTRACT
The interest in natural refrigerants, such as carbon dioxide, has been growing in recent years due to the high direct
global warming potential of common HFC refrigerants. Despite the environment-friendly characteristics of CO2 as a
refrigerant, due to high heat rejection temperatures and transcritical operation, CO2 cannot match the high energy
efficiency associated with current HFC technology. Thus, additional measures must be taken to achieve high COP
when using CO2. One approach is to use CO2 as one of the fluids in a cascade system along with a HFC refrigerant
as the high side fluid. Such systems may have roughly 75% less HFC refrigerant charge, and the global warming
potential is reduced compared to a baseline system using only HFC refrigerant. When used as a second fluid in a
cascade system, the CO2 cycle remains in the subcritical region, thus increasing the cycle’s COP. In this paper an
approach to model cascade systems is presented. The model is validated using experimental data for a R404A/CO2
cascade system and results are discussed. The simulation tool is developed to account for additional components in
the system such as multiple condensers and compressors. Some test points in the experimental system were run
using a set of parallel compressors, which can be easily addressed by the simulation tool.

1. INTRODUCTION
Low temperature vapor compression refrigeration systems are often very energy intensive due to the high pressure
ratios required to pump heat from a low temperature source. Such systems are designed to operate at evaporating
temperatures typically below -30.0°C with condensing temperatures dictated by the ambient temperature. Low
temperature systems are generally limited to conventional HFC refrigerants due to the large temperature lift. The
thermodynamic properties of certain HFC refrigerants allow for the most efficient operation of a conventional low
temperature vapor compression system, however HFC refrigerants are associated with high direct global warming
potential. Large-scale commercial low temperature vapor compression systems often use large amounts of
refrigerant charge (roughly on the order of 100’s kg of refrigerant) in order to provide cooling to multiple zones,
which increases the environmental impact of refrigerant leaks from the system.

Due to the increasing interest in environmental protection and conservation, there has been a recent resurgence in
the interest of natural refrigerants. Carbon dioxide, for example, has been embraced as an environment-friendly
refrigerant due to its low global warming potential. However, due to the low critical temperature of CO2, heat
rejection occurs at a high temperature well above the critical point, reducing the system performance. The use of

International Refrigeration and Air Conditioning Conference at Purdue, July 14-17, 2008
 2273, Page 2

CO2 as the sole refrigerant in a low temperature system is not necessarily a preferred solution; however the use of
CO2 as the low temperature refrigerant in a cascade refrigeration system is a very promising alternative.

Primary
TC Condenser

Expansion High Stage

Valve Compressor
TC Condenser

Tint , E Cascade
Expansion Compressor Tint ,C Condenser
Valve

Expansion Low Stage

Valve
TE Evaporator
Compressor

TE Evaporator

(a) (b)
Figure 1: a) Single stage vapor compression, and b) Cascade vapor compression system

A simple cascade vapor compression system is shown in Figure 1b. The system consists of two conventional vapor
compression systems coupled through a cascade condenser. The evaporator in the low temperature system absorbs
heat from the conditioned space. The heat absorbed by the low temperature evaporator along with the work input
from the low stage compressor is rejected to the high temperature vapor compression system through the cascade
condenser. The heat transferred to the high temperature vapor compression system in the cascade condenser and the
work input from the high stage compressor is rejected in the primary condenser. One of the main benefits of such
systems is that various refrigerants suitable for the corresponding temperature range can be used in each stage.

8.0E+6 360
CO2 R404A
7.0E+6 340
6.0E+6 320 TC
5.0E+6 300

4.0E+6 280
R404A Tint,C
3.0E+6 260 Tint,E
2.0E+6 240 TE
1.0E+6 220 CO2

0.0E+0 200
0 200 400 600 500 1000 1500 2000 2500
Enthalpy (kJ/kg) Entropy (J/K)
Low Temp. System (CO2) High Temp. System (R404A) Low Temp. System (CO2) High Temp. System (R404A)

(a) (b)
Figure 2: (a) Pressure enthalpy and (b) temperature entropy diagram of a theoretical cascade system

The use of a conventional HFC refrigerant, such as R404A, as the high temperature refrigerant along with CO2 as
the low temperature refrigerant is a suitable combination for low temperature refrigeration. In an R404A/CO2
cascade system the heat rejection of the low temperature system occurs at a much lower temperature than a
conventional CO2 vapor compression system. The heat rejection occurs at a temperature below the critical
temperature resulting in a higher COP.

International Refrigeration and Air Conditioning Conference at Purdue, July 14-17, 2008
 2273, Page 3

In a conventional vapor compression system, as shown in Figure 1a, there are two temperature levels, namely the
evaporating temperature, TE, and the condensing temperature, TC. For a properly designed system, these temperature
levels are ultimately dictated by the temperatures of the conditioned space and ambient. However, in a cascade
system there are four temperatures levels; the two additional temperatures being the condensing temperature of the
low temperature system, Tint,C, and the evaporating temperature of the high temperature system, Tint,E. The two
intermediate temperatures drive heat transfer from the low temperature system to the high temperature system. The
four temperature levels for an ideal cascade vapor compression system are shown on the T-s diagram depicted in
Figure 2b. The corresponding P-h diagram is shown in Figure 2a. Despite the carbon dioxide cycle operating in the
subcritical region, the operating pressures remain higher than the corresponding pressures in the R404A high
temperature system.

Similar to the conventional system, the condensing and evaporating temperatures (TE and TC) are dependent on the
surrounding environmental conditions. However, the intermediate temperatures for a given operating point are
dependent on the design of the system and ideal temperatures can be determined through the use of optimization.
Lee et al. (2006) points out that optimization studies of cascade systems are lacking, however only carries out a
thermodynamic analysis of a cascade system and details regarding heat exchanger characteristics are not included in
the simulation. Prior to conducting optimization studies of cascade systems, a validated simulation tool must be
developed. This paper presents a generalized simulation technique to modeling a cascade vapor compression system
and presents validation results using experimental data from an R404A/CO2 cascade system.

2. MODELING PROCEDURE
One approach to simulating a cascade vapor compression system is to simulate the low and high temperature
systems individually. In the system shown in Figure 3, the heat rejection in the condenser is being absorbed by a
secondary fluid. Prior to simulating this system, the secondary fluid inlet condition and mass flow rate is typically
specified prior to executing the condenser model. If the system shown in Figure 3 was being used as the low
temperature system in a cascade vapor compression system, the secondary fluid inlet condition and mass flow rate
would be provided through simulation of the high temperature system. The system shown in Figure 3a contains
additional components compared to a conventional vapor compression system. Thus, a generalized solution
approach to solve non-cascade vapor compression system containing multiple components should be implemented
to handle complex system configurations.

mS , TS ,in , PS ,in TS ,out , PS ,out

n m
Condenser
m Exp, i mCom , j
3 2 P1 i 1 j 1
P2 n m
m Exp,i hEvap ,out ,i h1 mCom , j
P3 i 1 j 1
x r PEvap ,1,out P1
P4
Compressor P5 PEvap , 2 ,out P1
4 1 Bank
Evaporator 1
h1 PCond , out P3
5
hSC hCond ,out
Evaporator 2

(a) (b) (c)

Figure 3: (a) Diagram of a multi-component vapor compression system, (b) vector of unknown solution variables,
and (c) corresponding set of residual equations

2.1. Vapor Compression System Simulation

An approach to simulate a multi-component vapor compression system has been presented by Winkler et al. (2007)
and the reader is asked to refer to this paper for a more complete description about the solution algorithm. In
general, to simulate a vapor compression system the solution algorithm must ascertain the refrigerant condition at
each state point within the cycle along with the mass flow rate through each component. Particular state points and

International Refrigeration and Air Conditioning Conference at Purdue, July 14-17, 2008
 2273, Page 4

mass flow rates can be calculated using various component models. For example, a compressor model typically
calculates a mass flow rate given an inlet pressure and enthalpy along with the outlet pressure. The total mass flow
rate calculated by the compressor bank in Figure 3a can be used to execute the condenser model. Similar to the
compressor, an expansion valve model typically calculates a mass flow rate which can be used to execute the
evaporator model connected downstream of the valve. The suction enthalpy and state point pressures are selected as
the unknown values, and the remaining state points and refrigerant mass flow rates are determined as the system
solver iteratively solves for the unknown valves. The unknown values can be placed in a solution vector, as shown
in Figure 3, which is initially filled with appropriate guess values.

Of course, a set of equations must be formulated with the number of equations equivalent to the number of unknown
variables. The solution procedure initiates with running the compressor bank using the suction enthalpy (h1) and the
suction and discharge pressures (P1 and P2). The total mass flow rate and mean discharge enthalpy, calculated using
adiabatic mixing, are propagated to the condenser prior to the model being executed. The condenser model
calculates the enthalpy at state point 3 and an outlet pressure using pressure drop correlations. Using the pressure at
state point 3 (P 3) and the outlet enthalpy of the condenser, each expansion valve model calculates a mass flow rate.
The corresponding expansion valve mass flow rate and outlet enthalpy along with the pressures at state points 4 and
5 (P4 & P5) are then used as inputs to run the corresponding evaporator model. Each evaporator model calculates an
outlet enthalpy and pressure. The calculated outputs from the component models are used in formulating a vector of
residual values, shown in Figure 3c, which represent the set of equations being solved to determine the set of
unknown values. For this case, the system subcooling has been chosen to close the set of equations and thus a value
for the subcooling must be input to the solution algorithm. The set of unknown values are simultaneously adjusted
by a nonlinear equation solver until the residuals satisfy a specified tolerance.

2.2. Cascade Vapor Compression System Simulation

Once a solution algorithm to simulate a vapor compression system has been developed, extending it to handle
cascade systems is not difficult. As previously mentioned, in a cascade system heat is transferred from the low
temperature system to the high temperature system through the cascade condenser. The cascade condenser model
serves as the condenser in the low temperature system and evaporator in the high temperature system.

Calculate guess condition at point 8

7 Condenser 6
Simulate low temperature system
holding condition at 8 constant
High Stage
Compressor
Update condition at 8 Simulate high temperature system
holding condition at 2 constant
8 5
Cascade
Condenser Calculate cascade system residual
3 2
high low
QCC QCC
r high
Low Stage QCC
Compressor

4 Evaporator 1 No Yes
r tol Done

(a) (b)
Figure 4: (a)Diagram of a cascade system, and (b) a flow chart describing the cascade simulation solution approach
A flow chart describing the approach implemented to simulate the cascade system is shown in Figure 4b. The
solution procedure starts by guessing an inlet condition to the cascade condenser at the point of the high temperature
system evaporator inlet (state point 8 in Figure 4a). Using the guessed condition and holding that condition constant,
the solution algorithm discussed in Section 2.1 is used to solve the low temperature vapor compression system.
Upon arriving at the solution, the inlet condition to the cascade condenser in the low temperature system will be
determined (state point 2 in Figure 4a). This condition is then held constant while the high temperature vapor
compression system is solved. Since state point 8 is likely to change after simulating the high temperature system,
the cascade condenser heat load of the low temperature system and high temperature system will not match. Since

International Refrigeration and Air Conditioning Conference at Purdue, July 14-17, 2008
 2273, Page 5

under steady state operation the heat being rejected from the low temperature system must equal the heat being
absorbed by the high temperature system, state point 8 is successively updated until the calculated heat load of the
cascade condenser is equivalent in both the low and high temperature systems. A normalized residual, r, is used to
determine if the solution procedure has reached convergence, as shown in Equation 1.

QCC
high
QCC
low
r (1)
QCChigh

In the above equation, QCC is the cascade condenser heat load calculated when the model is being run as the
high

evaporator in the high temperature system, and similarly QCC is the cascade condenser heat load calculated when
low

the model is being run as the condenser in the low temperature system. A tolerance value of 1.0E-4 was used to
determine if the cascade solver converged, which corresponds to an error between the two capacities of less than 1W
for a 10kW system.

3. SIMULATION VALIDATION RESULTS

3.1. Experimental System
Experimental results from a cascade system using R404A as the high temperature refrigerant and CO2 as the low
temperature refrigerant were used to validate the simulation tool. The experimental system is shown in Figure 5. The
high temperature system consisted of two compressors of varying capacity and a primary and secondary condenser.
The water flow rate through the primary condenser was adjusted for each case to achieve the condensing
temperature listed in Table 1 and in cases 2-6 a secondary condenser had to be used to further reduce the R404A
condensing temperature. The glycol flow rate through the CO2 evaporator was adjusted to maintain a constant
evaporating temperature for all six cases. Each of the compressors in the high temperature system could run in either
a loaded configuration (all four cylinders used to compress refrigerant) or unloaded configuration (two of four
cylinders used to compressor refrigerant).

Water
Secondary Primary Table 1: Desired R404A saturation temperatures and R404A
7 Condenser Condenser
6
condenser configuration
R404A Cond. CO2 Evap. R404A 2nd
Case
Temp. (°C) Temp. (°C) Cond.
1 48.9 -31.7 Off
2 40.6 -31.7 Off
R404A1 R404A2
3 32.2 -31.7 On
8 5 4 32.2 -31.7 On
Cascade 5 21.1 -31.7 On
Condenser
3 2 6 10.0 -31.7 On

Table 2: Cascade system compressor configuration

Case R404A1 R404A2 CO21 CO22 CO23
CO21 CO22 CO23 1 Unloaded Off Off On Off
2 Loaded Off Off Off On
3 Unloaded Off On On Off
4 Evaporator 1
4 Unloaded Off On On Off
Glycol 5 Unloaded Off On On Off
Figure 5: Diagram of experimental cascade system 6 Off Unloaded Off Off On

Table 2 lists the compressor configurations for both the high temperature and low temperature systems. Though only
a single compressor is run for each case in the high temperature system, cases 2-5 utilize two CO2 compressors
running simultaneously in a parallel arrangement. The low temperature system compressor bank consisted of three
compressors of various sizes. Brazed plate heat exchangers were used as the secondary R404A condenser, cascade
condenser, and CO2 evaporator. The primary R404A condenser was a water-cooled shell and tube heat exchanger.

International Refrigeration and Air Conditioning Conference at Purdue, July 14-17, 2008
 2273, Page 6

3.2. Modeling Procedure and Assumptions

The compressors in the high and low temperature systems were modeled using manufacturer data. For the high
temperature system, compressor coefficients for the loaded and unloaded operation were used along with the
compressor polynomial equation provided by ARI Standard 540-1999. The CO2 compressors were specifically
manufactured to operate in the subcritical region with condensing temperatures in the range of 0 to -20°C and
coefficients according to ARI Standard 540-1999 were derived using manufacturer rated performance data.

The low temperature evaporator, cascade condenser, and high temperature condensers were modeled using overall
heat transfer coefficients and heat transfer areas. Details regarding the geometric details of the heat exchangers were
not available and only the total heat transfer area was known. Thus, a simple multiple zone UA-based heat
exchanger model was developed. Apportioning heat transfer area to a given phase regime is important to accurately
capture the effect of changing temperature gradients within the heat exchanger. Heat transfer areas were apportioned
based on the refrigerant and secondary fluid inlet conditions. For example, the cascade condenser as shown in Figure
6a, and on the condensing (CO2) side has superheated refrigerant flowing in from the low temperature compressor
bank which is de-superheated and eventually subcooled. On the other side (R404A), refrigerant enters as a two-
phase mixture which is eventually superheated. As shown in Figure 6b there are four heat transfer zones in the
exchanger; namely a subcooled liquid to two-phase zone, a two-phase to two-phase zone, a two-phase to
superheated zone, and a superheated to superheated zone. Each zone was modeled using effectiveness-NTU
relations as described by Incropera and DeWitt (1996). The low temperature evaporator and high temperature
condenser are essentially simplifications of the cascade condenser model since the secondary fluid remains in the
liquid phase.

Tlow,dis

Thigh,evap,in Cascade Thigh,suction Tlow,cond

Condenser
Tlow,cond,out Tlow,discharge Thigh,evap

0 1
Heat Exchanger Length
Low Temp. Evaporator High Temp. Condenser
(a) (b)
Figure 6: (a) Schematic showing cascade condenser inlet temperatures, and (b) diagram showing temperature profile
through the cascade condenser
As previously mentioned, the overall heat transfer areas were provided from manufacturer data. The overall heat
transfer coefficients were calculated using experimental data and were assumed constant. Since the pressure drop
though the heat exchangers was not measured during the experimental phase, the pressure drop through the heat
exchangers was neglected. Accurate modeling of the heat exchangers was the most critical step in conducting the
validation. Though the assumptions used to model the exchangers are arguably simplistic, the component-based
nature of the simulation tool allows for easy substitution with more detailed heat exchanger models once available
(Winkler et al., 2006). Winkler et al. (2006) and Winkler et al. (2008) provide validation results for different types
of air-to-refrigerant vapor compression systems in which detailed air-to-refrigerant heat exchanger models were
used in the simulation.

The experimental system utilized a thermostatic expansion valve as the throttling device in the low and high
temperature systems, and thus the validation was conducted using the experimental suction superheat as input. In all
cases for the high temperature system, the refrigerant was subcooled a few degrees at the condenser outlet, thus it
made sense to use the experimental subcooling as input to the simulation as described in Section 2.1. However in the
low temperature system, low quality two-phase refrigerant exited the cascade condenser and there was not a

International Refrigeration and Air Conditioning Conference at Purdue, July 14-17, 2008
 2273, Page 7

subcooling value to use as input to the simulation. Therefore, the system discharge pressure was selected as input to
the simulation in order to close the system of equations. Thus, the final residual in Figure 3c was replaced with
Equation 2, where Pdis,exp is the experimental discharge pressure for that particular case.

r Pdis ,exp P2 (2)

3.3. Validation Results

The low and high temperature system capacity and power consumption validation results are shown in Figure 7. For
the low temperature system, the average error in system capacity was 2.6% with a maximum error of 3.6%. Capacity
results for the high temperature system are similar with an average error of 2.5% and a maximum error of 2.9%. The
compressor power was accurately predicted for the low temperature system with an average error of 2.9% and a
maximum error of 5.1%. For the high temperature system, the power consumption was predicted with an average
error 4.5%, however the maximum error was fairly high with an error of 11.4%. It is difficult to determine the
source of this error since it was the only case in which R404A compressor 1 was being run fully loaded and pressure
ratio for this case was predicted to within 2.3%.

3.5E+4 System Capacity 1.8E+4 Compressor Power

3.0E+4 +4% Error +5% Error

1.4E+4
-4% Error -5% Error
2.5E+4
1.0E+4
2.0E+4

6.0E+3
1.5E+4

1.0E+4 2.0E+3
1.0E+4 1.5E+4 2.0E+4 2.5E+4 3.0E+4 3.5E+4 2.0E+3 6.0E+3 1.0E+4 1.4E+4 1.8E+4
Experimental Capacity (W) Experimental Power (W)
Low Temp. System High Temp. System Low Temp. System High Temp. System
(a) (b)
Figure 7: Cascade system validation results for (a) system capacity and (b) compressor power consumption
The low and high temperature system mass flow rate results are shown in Figure 8a. The average error for the low
temperature system mass flow rate was 2.7% with a maximum error of 3.6%. For the high temperature system the
average error was 4.4% with a maximum error of 6.8%. The maximum error in the mass flow rate prediction
occurred for the cases with the maximum error in predicted saturation temperatures.

The predicted saturation temperatures are shown in Figure 8b and all saturation temperatures were predicted to
within ±3°C. The error in the saturation temperatures is rather small compared to the temperature gradient between
the refrigerant and secondary fluid in the low temperature evaporator and high temperature condensers. There was
no error in the low temperature system condensing temperature since the experimental discharge pressure was used
as input to the simulation as described in Section 3.2. The high temperature system evaporating temperature was
predicted with an average error of 1.5°C, and thus the temperature gradient across the cascade condenser was
accurately accounted for. System saturation temperatures are the primary driving force behind the heat transfer
between the refrigerant and the surrounding environment and thus it is important that the saturation temperatures be
accurately predicted. However, it is possible to accurately predict the saturation temperatures while failing to
accurately predict the pressure ratio, which is the primary factor dictating the compressor power consumption. The
pressure ratio for the low temperature system was predicted with average error of 2.8% and a maximum error of
4.3%. Again, this is expected since the experimental discharge pressure was used as input to the simulation. The
pressure ratio of the high temperature system was not as accurately predicted and there was an average error of 9.4%
with a maximum error of 15.2% between the simulated and experimental values. This seems to be attributed to the

International Refrigeration and Air Conditioning Conference at Purdue, July 14-17, 2008
 2273, Page 8

simplifications used in modeling the condensers. The maximum error in the pressure ratio prediction did not
coincide with the maximum error in the high temperature system compressor power prediction.

Mass Flow Rate 55 Saturation Temperatures

45
0.3
+5% Error 35 +3°C Error

25 -3°C Error
-5% Error
0.2 15
5
5
0.1
15
25
0.0 35
0.0 0.1 0.2 0.3 35 25 15 5 5 15 25 35 45 55
Experimental Mass Flow Rate (kg/s) Experimental Temperatures (°C)
Low Temp. System High Temp. System Low Temp. System High Temp. System
(a) (b)
Figure 8: Cascade system validation results for (a) system mass flow rate and (b) saturation temperatures

4. CONCLUSIONS
In this paper, a cascade system simulation algorithm has been discussed and implemented with the help of a
component-based modeling tool for vapor compression systems. The low temperature and high temperature vapor
compression systems consisted of multiple compressors and the high temperature system utilized two condensers.
The simulation tool, despite using simple heat exchanger models, predicted the COP with an average error of 4.4%
and a maximum error of 11.3%.

REFERENCES
ANSI/ARI Standard 540-1999, 1999, “Positive Displacement Refrigerant Compressors and Compressor Units”, Air-
Conditioning and Refrigeration Institute, Arlington, VA

Incropera, F.P., and DeWitt, D.P., 1996, Fundamentals of Heat and Mass Transfer, John Wiley & Sons, New York,
NY, p. 582-618.

Lee, T.S., Liu, C.H., and Chen, T.W., 2006, “Thermodynamic Analysis of Optimal Condensing Temperature of
Cascade-Condenser in CO2/NH3 Cascade Refrigeration Systems”, International Journal of Refrigeration,
Vol. 29, p. 1100-1108.

Winkler, J., Aute, V., and Radermacher, R., 2006, “Component-Based Vapor Compression Simulation Tool with
Integrated Multi-Objective Optimization Routines”, International Refrigeration and Air Conditioning
Conference at Purdue, Purdue University, July 2006.

Winkler, J., Aute, V., and Radermacher, R., 2007, “Comprehensive Investigation of Numerical Methods in
Simulating a Steady State Vapor Compression System”, International Journal of Refrigeration,
doi:10.1016/j.ijrefrig.2007.08.008.

Winkler, J., Aute, V., and Radermacher, R., 2008, “Simulation and Validation of a Two-Stage Flash Tank Cycle
using R410A as a Refrigerant”, International and Refrigeration and Air Conditioning Conference at
Purdue, Purdue University, July 2008.

International Refrigeration and Air Conditioning Conference at Purdue, July 14-17, 2008