Abstract 
 
 
Title of Document:   TRANSIENT PERFORMANCE EVALUATION  
     OF AUTOMOTIVE SECONDARY LOOP  
     SYSTEMS 
 
 
     Magnus Eisele, Ph.D., 2012 
 
 
Directed by:    Professor Reinhard Radermacher 
     Center for Environmental Energy Engineering 
     Department of Mechanical Engineering 
     University of Maryland, College Park 
 
Automotive air-conditioning is a high impact technology where 
improvements in energy consumption and environmental performance can make 
a significant difference in fuel efficiency and comfort. The mandatory phase out of 
R134a as refrigerant in the European Union has set the stage for new systems 
and alternative refrigerants. While some of these refrigerants, such as R152a or 
R290, have a low Global Warming Potential, their flammability requires 
secondary loop systems to be used. The added thermal mass of such systems 
may increase power consumption and delay cool down while benefitting thermal 
comfort during start/stop operation. The recent revival of electric vehicles, as well 
as the associated focus on air-conditioning energy consumption, provides new 
challenges and opportunities.  
 
This research focuses on the performance evaluation of refrigerants 
R152a and R290 during transient operation in secondary loop systems, 
quantification of thermal storage benefits for start/stop operation, and 
investigation of energy saving potentials in electric vehicles through the use of 
advanced air-conditioning system controls and cabin preconditioning.  
A test facility was built to dynamically test secondary loop systems over a 
wide range of pull down conditions and drive cycles using a passenger cabin 
model and associated controls. It was shown that R290 is a viable alternative in 
secondary loop systems and system performance may be on par or better 
compared to R134a direct expansion systems. The preservation of cooling 
capacity and thermal comfort during off-cycle periods were quantified for a 
secondary loop system, as well as a combined ice storage system. System 
efficiency increases with longer off-cycle periods compared to direct expansion 
systems. Advanced compressor control strategies and the use of cabin 
preconditioning can make use of this characteristic and improve energy efficiency 
by more than 50%. Ice storage may be used in combination with cabin 
preconditioning to preserve comfort for an extended driving time with reduced 
use of the vapor compression cycle. A Modelica model of the secondary loop 
system was developed and validated with experimental data. The model enables 
dynamic simulation of pull-down and drive cycle scenarios and was used to study 
the effects of coolant volume and coolant concentration on transient performance. 
 
 
 
 
 
TRANSIENT PERFORMANCE EVALUATION OF 
AUTOMOTIVE SECONDARY LOOP SYSTEMS  
 
By 
 
Magnus Eisele 
 
 
 
 
Dissertation submitted to the Faculty of the Graduate School of the 
University of Maryland, College Park, in partial fulfillment  
of the requirements for the degree of  
Doctor of Philosophy 
 2012 
 
 
 
 
 
Advisory Committee: 
Professor Reinhard Radermacher, Chair 
Professor Gregory Jackson 
Professor Kenneth Kiger 
Associate Professor Bao Yang 
Associate Professor Peter Sunderland  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
? Copyright by 
Magnus Eisele 
2012 
 
 
 
 
 
 
 
 
 
 
 
 
ii 
 
 
 
 
 
 
 
 
 
 
 
Dedicated to my wife Kyungha and our son Simon 
 with my deepest love and gratitude 
 
  
iii 
Acknowledgments 
 
It is often said that a dissertation is not written alone. I would like to take 
the opportunity to thank some of the people who have helped make this research 
succeed and who have shaped me throughout the last years. 
With my sincerest gratitude I would like to thank my advisor Dr. Reinhard 
Radermacher for giving me the opportunity to work and study at the Center for 
Environmental Energy Engineering. His experience and advise throughout the 
last years were invaluable and his guidance inside and outside the framework of 
research was much appreciated. I am no less grateful to Dr. Yunho Hwang, who 
directed my research in the heat pump laboratory. His organization of the ACTA 
consortium, his knowledge and guidance of the day to day research, and his 
valuable support in the revision of publications, theses, and reports will not be 
forgotten. I also wish to express my appreciation to my committee members who 
helped guide this research: Dr. Gregory Jackson, Dr. Kenneth Kiger, Dr. Bao 
Yang, and Dr. Peter Sunderland.  
I want to direct my gratitude to the CEEE staff and students, especially the 
heat pump lab group, whom I worked with in the past years. The spirit and team 
work in the group were deciding factors for me to embark on the PhD journey in 
this lab. Many thanks to the "old" group I started in: Cara Martin, Jan Muehlbauer, 
Ahmet Oers, Xudong Wang and Dae-Hyun Jin, as well as the "new" group: 
Daniel Leighton, Kyle Gluesenkamp, Hoseong Lee, Xu Xing, Jiazhen Ling, Ali 
iv 
Alalili, Sahil Popli, Daniel Spencer and Bracha Mandel, as well as all the others I 
worked with over the years. It was a pleasure to work with you. 
Special thanks go to Hongtao Qiao for his support and discussions about 
component and system modeling. I would also like to thank Gang Li for helping to 
build and install the ice storage box for the secondary loop system. A special 
thank you also goes to Toru Okuma for the fruitful discussions in the lab and the 
supply of compressors for my test facility. 
I want to extend my deepest gratitude to my family, my parents Brigitte 
Eisele and Hans-Werner Eisele, as well as my wife Kyungha Park and our son 
Simon Eisele. You were a constant source of support, motivation and joy.  
 
 
 
  
v 
Table of Contents 
1 Literature Review ......................................................................................... 1 
1.1 Dynamic Testing of Automotive Air-Conditioning Systems .................... 1 
1.2 Transient Modeling of Automotive Air-Conditioning Systems ................ 5 
1.3 Alternative Refrigerants in Air-Conditioning and Refrigeration ............ 10 
1.4 Secondary Loop Systems in Air-Conditioning and Refrigeration ......... 22 
1.5 Literature Review Summary ................................................................ 35 
2 Research Objectives .................................................................................. 38 
3 Experimental Test Facility ......................................................................... 41 
3.1 Direct Expansion Test Facility ............................................................. 41 
3.2 Secondary Loop Test Facility .............................................................. 44 
3.3 Secondary Loop combined with Ice Storage Test Facility ................... 47 
4 Data Acquisition and Analysis .................................................................. 52 
4.1 Data Acquisition Process ..................................................................... 52 
4.2 Data Reduction Process ...................................................................... 53 
4.3 Uncertainty Analysis ............................................................................ 59 
5 Automation and Control ............................................................................ 68 
5.1 Structure of the Data Acquisition Control ............................................. 68 
5.2 The Cabin Model ................................................................................. 69 
5.3 Drive Cycle Control .............................................................................. 73 
5.4 Component Control ............................................................................. 76 
6 Refrigerant Comparison ............................................................................ 78 
7 Experimental Results: Secondary Loop Versus Direct Expansion ....... 85 
7.1 Charge Optimization - Secondary Loop Versus Direct Expansion ....... 86 
7.2 Steady-state Performance - Secondary Loop Versus Direct Expansion
  ............................................................................................................ 89 
7.3 Transient Performance - Secondary Loop Versus Direct Expansion ... 94 
7.3.1 Pull-down Performance - 2LP Versus DX ..................................... 96 
7.3.2 New European Drive Cycle Performance - 2LP Versus DX ........ 101 
7.4 Summary of Results: Secondary Loop Versus Direct Expansion ...... 106 
8 Experimental Results: Propane in Secondary Loop Systems ............. 108 
vi 
8.1 Charge Optimization: Propane in Secondary Loop Systems ............. 109 
8.2 Steady-state Performance: Propane in Secondary Loop Systems .... 111 
8.3 Transient Performance: Propane in Secondary Loop Systems ......... 116 
8.3.1 Pull-down Performance: Propane 2LP ........................................ 116 
8.3.2 New European Drive Cycle Performance - Propane 2LP ............ 121 
8.4 Summary: Propane as Refrigerant in Secondary Loop Systems ....... 124 
9 Thermal Storage During Off-Cycle Test Results ................................... 129 
9.1 Secondary Loop Thermal Storage Performance ............................... 130 
9.1.1 Secondary Loop Start/Stop Drive Cycle Performance ................ 131 
9.1.2 Secondary Loop Off-Cycle Cabin Warm-up Performance ........... 139 
9.2 Ice Storage Thermal Storage Performance ....................................... 143 
9.2.1 Ice Storage Charging Performance ............................................. 146 
9.2.2 Ice Storage Pull-down Performance ............................................ 149 
9.2.3 Ice Storage Off-cycle Cabin Warm-up Performance ................... 153 
9.3 Summary: Thermal Storage in Secondary Loop Systems ................. 157 
10 Experimental Results - A/C Controls and Cabin Preconditioning ....... 160 
10.1 A/C Control Performance and Benefits .............................................. 163 
10.2 Cabin Preconditioning Performance and Benefits ............................. 170 
10.3 Summary: Benefits of A/C Controls and Cabin Pre-Conditioning ...... 177 
11 Transient Simulation of Secondary Loop Systems ............................... 179 
11.1 Cabin Model ...................................................................................... 179 
11.1.1 Cabin Model Equations and Structure ......................................... 179 
11.1.2 Cabin Model Validation ............................................................... 184 
11.2 Direct Expansion System .................................................................. 185 
11.2.1 Direct Expansion Model and Components .................................. 186 
11.2.2 R134a Direct Expansion Model Validation .................................. 190 
11.3 Secondary Loop System ................................................................... 197 
11.3.1 Coolant Cycle Model and Components ....................................... 197 
11.3.2 R290 Secondary Loop Model Validation ..................................... 203 
11.4 Influence of Coolant Volume on Transient Performance ................... 209 
11.5 Influence of Coolant Concentration on Transient Performance ......... 215 
vii 
12 Summary and Future Work ..................................................................... 222 
12.1 Research Contributions ..................................................................... 222 
12.2 Summary of Research Outcomes ...................................................... 223 
12.3 Recommendations for Future Work ................................................... 226 
Appendix A Specification of Test Facility Components ......................... 228 
Appendix B Control Flow Charts .............................................................. 232 
Appendix C Drive Cycle Development ..................................................... 245 
Appendix D Steady-state Results Summary: Secondary Loop Versus 
Direct Expansion ............................................................................................ 254 
Appendix E Transient Test Procedure Illustrations ................................ 259 
Appendix F Steady-state Results Summary: Propane 2LP .................... 269 
Appendix G Glycol Property Equations.................................................... 274 
Appendix H Modelica Direct Expansion System Equations ................... 275 
Appendix I R134a DX Steady-state Reference Data ............................... 277 
References ...................................................................................................... 282 
 
  
viii 
List of Figures 
Figure ?3.1: Direct Expansion (DX) Test Facility Schematic ............................................43 
Figure ?3.2: Secondary Loop (2LP) Test Facility Schematic ............................................46 
Figure ?3.3: Ice Storage (IS) Test Facility Schematic ......................................................48 
Figure ?3.4: Ice Storage Heat Exchanger Schematic ......................................................49 
Figure ?3.5: Ice Storage Facility Operation Modes ..........................................................50 
Figure ?4.1: Data Acquisition Process Structure ..............................................................53 
Figure ?4.2: Moist Air / Refrigerant / Glycol Property Sub Routines .................................55 
Figure ?4.3: Systematic and Random Uncertainty ...........................................................60 
Figure ?4.4: Transient Uncertainty Example - New European Drive Cycle ......................66 
Figure ?5.1: Main Data Acquisition Structure and Control in LabVIEW ............................69 
Figure ?5.2: Cabin Model Operational Schematic ............................................................71 
Figure ?5.3: NEDC Cycle Input to LabVIEW Cycle Reader .............................................74 
Figure ?5.4: NEDC Cycle PID Control Verification ...........................................................76 
Figure ?6.1: Refrigerant P-h Diagram ..............................................................................81 
Figure ?6.2: Refrigerant T-s Diagram ..............................................................................84 
Figure ?6.3: Alefeld Refrigerant Efficiency with Respect to R134a ..................................84 
Figure ?7.1:Estimated Charge Savings for R152a Secondary Loop Systems ..................88 
Figure ?7.2: Secondary Loop Steady-state COP Illustration ............................................89 
Figure ?7.3: System Temperature and Pressure State Point Profiles (2LP Versus DX) ...90 
Figure ?7.4: Steady State Performance Results (2LP Versus DX) ...................................93 
Figure ?7.5: Secondary Loop Transient Cooling Capacity Illustration ..............................95 
Figure ?7.6: Secondary Loop Transient Capacity and COP Illustration............................95 
Figure ?7.7: Pull-down - Time to Cabin Comfort (2LP Versus DX) ...................................97 
Figure ?7.8: Pull-down Performance Results (2LP Versus DX) .......................................99 
Figure ?7.9: Pull-down Time-normalized Performance Results (2LP Versus DX) .......... 100 
Figure ?7.10: NEDC Supply Temperature Trend (2LP Versus DX) ................................ 103 
Figure ?7.11: NEDC Performance Metrics Trend (2LP Versus DX) ............................... 104 
Figure ?7.12: NEDC Accumulated Performance Metrics Results (2LP Versus DX) ....... 106 
Figure ?8.1: R290 Drop-in Performance in the Direct Expansion System ...................... 109 
Figure ?8.2: Estimated Charge Savings for R290 Secondary Loop Systems ................. 111 
ix 
Figure ?8.3: System Temperature and Pressure State Points (R290 2LP) .................... 112 
Figure ?8.4: Steady-state Performance Results (R290 2LP) ......................................... 114 
Figure ?8.5: Theoretical R290 2LP Performance Results with Increased Idling 
Compressor Efficiencies ..................................................................................... 115 
Figure ?8.6: Pull-down Time to Cabin Comfort (R290 2LP) ........................................... 117 
Figure ?8.7: Pull-down Performance Results (R290 2LP) .............................................. 119 
Figure ?8.8: Pull-down Time-normalized Performance Results (R290 2LP) .................. 120 
Figure ?8.9: NEDC Supply Temperature Trend (R290 2LP) .......................................... 121 
Figure ?8.10: NEDC Performance Metrics Trends (R290 2LP) ...................................... 123 
Figure ?8.11: NEDC Accumulated Performance Metrics Results (R290 2LP) ............... 124 
Figure ?8.12: Impact of Secondary Loop Systems on Life Cycle Climate Performance of a 
Mid-size Passenger Car ..................................................................................... 128 
Figure ?9.1: Start/Stop Drive Cycle Supply Air Profiles .................................................. 133 
Figure ?9.2: Start/Stop Drive Cycle Accumulated Performance Results ........................ 134 
Figure ?9.3: Start/Stop 3x Idling Drive Cycle Supply Air Profiles ................................... 136 
Figure ?9.4: Start/Stop 3x Idling Drive Cycle Accumulated Performance Results .......... 137 
Figure ?9.5: Drive Cycle Accumulated Performance Results Comparison ..................... 139 
Figure ?9.6: Secondary Loop Off-cycle Cabin Warm-up Profiles ................................... 141 
Figure ?9.7: Secondary Loop Off-cycle Relative Humidity Profiles ................................ 143 
Figure ?9.8: Enthalpy of Fusion Based on Mass of Ice .................................................. 145 
Figure ?9.9: Ice Storage Charging Performance ............................................................ 148 
Figure ?9.10: Ice Storage Pull-down Cabin Air Profiles ................................................. 150 
Figure ?9.11: Ice Storage Pull-down Transient Performance Profiles ............................ 152 
Figure ?9.12: Ice Storage Pull-down Accumulated Performance Results ...................... 153 
Figure ?9.13: Ice Storage Off-cycle Cabin Warm-up Temperature Profiles .................... 155 
Figure ?9.14: Cabin Warm-up Accumulated Sensible Capacity Comparison ................. 156 
Figure ?9.15: Ice Storage Off-cycle Cabin Warm-up Humidity Profiles .......................... 157 
Figure ?10.1: A/C Controls Power Consumption and Cabin Air Profiles Comparison .... 167 
Figure ?10.2: A/C Controls Accumulated Performance Results Comparison ................. 169 
Figure ?10.3: Theoretical Change in A/C Ctrl Accumulated Performance when Using an 
Indirectly Cooled Condenser for Supply Air Reheating ....................................... 170 
x 
Figure ?10.4: Preconditioning - Accumulated Performance Results Comparison .......... 173 
Figure ?10.5: Preconditioning - Power Consumption Transient Profile Comparison ...... 174 
Figure ?10.6: Preconditioning Ice Storage Performance Comparison ............................ 176 
Figure ?11.1: Modelica Passenger Cabin Model ........................................................... 183 
Figure ?11.2: Cabin Model Validation Results ............................................................... 185 
Figure ?11.3: Direct Expansion Steady-state System Model ......................................... 187 
Figure ?11.4: R134a Direct Expansion Transient System Model ................................... 189 
Figure ?11.5: R134a Direct Expansion NEDC Validation Results .................................. 194 
Figure ?11.6: R134a Direct Expansion Pull-down Validation Results ............................ 196 
Figure ?11.7: Secondary Loop Steady-state System Model .......................................... 201 
Figure ?11.8: Secondary Loop Transient System Model ............................................... 202 
Figure ?11.9: Secondary Loop NEDC Validation Results .............................................. 206 
Figure ?11.10: Secondary Loop Pull-down Validation Results ....................................... 208 
Figure ?11.11: Influence of Coolant Volume on Cabin Pull-down  (R290 2LP, T30/RH50, 
12 K soak, 1550 rpm, 250 g/s) ........................................................................... 211 
Figure ?11.12: Influence of Coolant Volume on Cabin Warm-up  (R290 2LP, T30RH50, 12 
K soak, 1550 rpm, 250 g/s) ................................................................................ 213 
Figure ?11.13: Influence of Coolant Volume on Performance Metrics  (R290 2LP, 
T35RH50, 12 K soak, 1550 rpm, 250 g/s) .......................................................... 214 
Figure ?11.14: Change of Freezing Temperature of an Aqueous Ethylene Glycol Mixture 
based on Mass Fraction ..................................................................................... 216 
Figure ?11.15: Influence of Coolant Concentration on Cabin Pull-down  (R290 2LP, 
T30RH50, 12 K soak, 1550 rpm, 250 g/s) .......................................................... 218 
Figure ?11.16: Influence of Coolant Concentration on Cabin Warm-up  (R290 2LP, 
T30RH50, 12 K soak, 1550 rpm, 250 g/s) .......................................................... 219 
Figure ?11.17: Influence of Coolant Concentration on Performance Metrics  (R290 2LP, 
T30RH50, 12 K soak, 1550 rpm, 250 g/s) .......................................................... 221 
Figure ?B.1: Cabin Model Flow Chart ............................................................................ 233 
Figure ?B.2: Thermostat Control Flow Chart ................................................................. 234 
Figure ?B.3: Drive Cycle Control Flow Chart ................................................................. 235 
Figure ?B.4: Electronic Expansion Valve Flow Chart ..................................................... 236 
Figure ?B.5: Coolant Pump Control Flow Chart ............................................................. 238 
xi 
Figure ?B.6: Compressor Control Flow Chart ................................................................ 239 
Figure ?B.7: Manual Compressor Control Flow Chart .................................................... 240 
Figure ?B.8: Pump Temperature Control Flow Chart ..................................................... 240 
Figure ?B.9: Compressor Pulldown Control Flow Chart ................................................. 241 
Figure ?B.10: Defrost Cycling Control Flow Chart ......................................................... 242 
Figure ?B.11: Relative Humidity Cycling Control Flow Chart ......................................... 243 
Figure ?B.12: Relative Humidity Continuous Control Flow Chart ................................... 244 
Figure ?C.1: Gear Equations for vehicle speed to compressor rpm conversion ............. 246 
Figure ?C.2: Superimposed EPA FTP Drive Cycle and Shift Cycle ............................... 247 
Figure ?C.3: NEDC Compressor Speed Profile ............................................................. 248 
Figure ?C.4: I-95 Cycle TMY2 Interpolated Climate Data .............................................. 251 
Figure ?C.5: I-95 Cycle - Cooling Capacity and Power Consumption ............................ 253 
Figure ?C.6: I-95 Cycle - Cabin Temperature ................................................................ 253 
Figure ?C.7: I-95 Cycle - Cabin Relative Humidity ......................................................... 253 
Figure ?D.1: Steady-state Performance Metrics Summary (2LP Versus DX) ................. 256 
Figure ?D.2: Steady-state Sensible/Latent Performance Summary (2LP Versus DX) .... 257 
Figure ?D.3: Steady-state Compressor Efficiency Summary (2LP Versus DX) .............. 258 
Figure ?E.1: Pull-down Test Procedure Illustration ........................................................ 260 
Figure ?E.2: NEDC Test Procedure Illustration .............................................................. 261 
Figure ?E.3: Pull-down Off-cycle Test Procedure Illustration ......................................... 262 
Figure ?E.4: Start/Stop Test Procedure Illustration ........................................................ 263 
Figure ?E.5: Start/Stop 3x Idling Test Procedure Illustration .......................................... 264 
Figure ?E.6: A/C Ctrl - Frost Cycling Test Procedure Illustration ................................... 265 
Figure ?E.7: A/C Ctrl - RH Cycling Test Procedure Illustration ...................................... 267 
Figure ?E.8: A/C Ctrl - RH Continuous Test Procedure Illustration ................................ 268 
Figure ?F.1: Steady-state Performance Metrics Summary (Propane 2LP) ..................... 270 
Figure ?F.2: Steady-state Sensible/Latent Performance Summary (Propane 2LP) ........ 272 
Figure ?F.3: Steady-state Compressor Efficiency Summary (Propane 2LP) .................. 273 
Figure ?I.1: Steady-state Performance Metrics Summary (R134a DX) .......................... 278 
Figure ?I.2: Steady-state Sensible/Latent Performance Summary (R134a DX) ............. 280 
Figure ?I.3: Steady-state Compressor Efficiency Summary (R134a DX) ....................... 281 
xii 
 
List of Tables 
Table ?1.1: Alternative Refrigerants and Characteristics (Wang et al.) [26] .....................11 
Table ?1.2: Comparison of Performance of Hydrocarbons per Application [26] ...............14 
Table ?1.3: Characteristics of Aqueous Solutions as Secondary Refrigerants [26] ..........30 
Table ?1.4: Comparison of R152a to R134a in an Automotive Secondary Loop System 
[39] .......................................................................................................................34 
Table ?3.1: Sensing Equipment Specifications ................................................................51 
Table ?4.1: Relative Systematic and Random Uncertainties (R290 2LP, T35RH40) .......62 
Table ?4.2: Summary of Steady-state Relative Uncertainty of Performance Parameters .64 
Table ?6.1: Refrigerant Property Table ............................................................................79 
Table ?7.1:Test Matrix - Secondary Loop Versus Direct Expansion System ....................85 
Table ?7.2: Charge Optimization Operating Conditions ...................................................87 
Table ?7.3: Charge Optimization Results: 2LP Versus DX ..............................................87 
Table ?8.1: Charge Optimization Results: R290 2LP ..................................................... 110 
Table ?10.1: Comparison of Energy Consumption for a 25 Minute Commute Relative to 
2LP F cyc (S) ..................................................................................................... 176 
Table ?11.1: Cabin Model Equations - Index Explanation .............................................. 180 
Table ?11.2: Cabin Model Validation Physical Parameters ............................................ 184 
Table ?11.3: R134a Direct Expansion Steady-state Model Validation Results ............... 191 
Table ?11.4: R134a Direct Expansion NEDC Accumulated Performance Validation ..... 195 
Table ?11.5: R134a Direct Expansion Pull-down Acc. Performance Validation ............. 197 
Table ?11.6: Secondary Loop Stationary System Validation Results ............................. 204 
Table ?11.7: Secondary Loop NEDC Acc. Performance Validation ............................... 205 
Table ?11.8: Secondary Loop Pull-down Acc. Performance Validation .......................... 207 
Table ?A.1: Evaporator Specifications ........................................................................... 228 
Table ?A.2: Condenser Specifications ........................................................................... 229 
Table ?A.3: Intermediate Heat Exchanger Specifications .............................................. 229 
Table ?A.4: Cooler Specifications .................................................................................. 230 
Table ?A.5: Compressor Specifications ......................................................................... 230 
Table ?A.6: Expansion Valve Specifications .................................................................. 231 
xiii 
Table ?A.7: Coolant Pump Specifications ...................................................................... 231 
Table ?C.1: Automobile Parameters for Speed to RPM Conversion .............................. 246 
Table ?C.2: Shift Point Table for Vehicle Acceleration and Deceleration ....................... 247 
Table ?C.3: I-95 Cycle - Route and Travel Information .................................................. 249 
Table ?G.1: Ethylene Glycol Property Equation Coefficients ......................................... 274 
  
xiv 
Nomenclature 
Symbols Explanations Units 
( A ) Ambient Cabin Preconditioning  
( C ) Comfort Cabin Preconditioning  
( S ) Soak Cabin Preconditioning  
2LP  Secondary Loop System  
A Area [m2] 
A/C Air-Conditioning  
AHU Air Handling Unit  
air Air-side  
AR Actual Reading  
ATL Atmospheric Life Time  
C+P Compressor + Pump Power  
C+P+H Compressor + Pump + Heater Power  
col Coolant-side  
COP Coefficient Of Performance [-] 
cp Specific Heat Capacity [J/kg K] 
D Diameter [m] 
DAQ Data Acquisition  
dP Pressure Drop [kPa] 
DX Dirext Expansion System  
E in Energy Consumed by the System [Wh] 
EB Energy Balance [-] 
EXV Electronic Expansion Valve  
F cyc Frost Cyclic Operation  
FS Full Scale  
FTP Federal Test Procedure  
GWP Global Warming Potential  
h enthalpy [J/kg] 
HC Hydrocarbon  
HFC Hydrofluorocarbon  
ht height [m] 
htc Heat Transfer Coefficient [W/m2 K] 
HVAC Heating, Ventilating, and Air-Conditioning  
hwy Highway Operation  
HX Heat Exchanger  
idle Idling Operation  
IS Ice Storage Facility  
k Thermal Conductivity [W/m K] 
L Latent Heat [J/kg] 
LEL Lower Explosion Limit [vol%] 
LFL Lower Flammability Limit [vol%] 
m/M Mass [kg] 
MAC Mobile Air-Conditioning  
MFR/mdot Mass Flow Rate [kg/s] 
n Number of ..  
NBP Normal Boiling Point [?C] 
NEDC New European Drive Cycle  
ODP Ozone Depletion Potential  
xv 
Symbols Explanations Units 
P Power [W] 
p/P Pressure [kPa] 
PID Proportional Integral Deriviative  
Q Capacity [W] 
Q (air) Energy used in Form of Cooling Capacity [Wh] 
ref Refrigerant-side  
RH Relative Humidity [%] 
RH ctn Relative Humidity Continuous Operation  
RH cyc Relative Humidity Cyclic Operation  
rpm Revolutions Per Minute  
s Entropy [J/kg K] 
S/S Start/Stop Cycle  
SHF Sensible Heat Factor [-] 
SL Secondary Loop  
SS3xI Start/Stop Three Times Idling Cycle  
STD Standard Deviation  
t/T Temperature [?C] 
TPF Transient Performance Factor [-] 
V Volume [m3] 
v Velocity [m/s] 
VCC Vapor Compression Cycle  
VFR Volume Flow Rate [m3/s] 
w Humidity Ratio [kg/kg] 
wt% Percent by Weight  
? Efficiency [%] 
? Dynamic Viscosity [mPa s] 
? Density [kg/m3] 
? Relative Humidity [%] 
? Uncertainty  
 
 
1 
1 Literature Review 
Chapter 1 reviews the pertinent literature in the fields of dynamic testing of 
automotive air-conditioning systems, the use of alternative refrigerants in air-
 conditioning systems, and the use of secondary loop systems in air-conditioning 
and refrigeration.  
 
1.1 Dynamic Testing of Automotive Air-Conditioning Systems 
Early studies on transient testing of automotive air-conditioning systems 
were performed by J. E. Rubio-Quero et al. [1] who used a Ford 1994 Crown 
Victoria R134a system for transient testing of mobile air-conditioning systems. 
Rubio-Quero et al. stated that while steady state tests are common in the 
automotive industry, they often do not accurately reflect the operating conditions 
of mobile air-conditioning systems (MACs). Since typical control schemes involve 
thermostatic expansion valves, and/or compressor clutch cycling, MACs most 
often operate in transient conditions.  
C. D. Collins and N. R. Miller [2] studied transient behavior of MACs. Their 
focus was on the development of a refrigerant charge loss diagnostic tool using 
measures of the transient behavior of the system to reduce premature clutch 
failure, and the experimental characterization of transient and steady-state 
performance of MACs. A narrow range of operating conditions was simulated, 
corresponding to a vehicle cruising at highway speeds with the MAC operating in 
full recirculation mode in a low humidity, high ambient temperature environment. 
2 
The test facility lacked the means of controlling humidity, since air loops were 
used, which controlled temperature by the power input to electric heaters and no 
air handling unit was present. The influence of humidity was neglected during this 
study. 
P. G. Weston et al. [3] studied the design and construction of a MAC test 
facility for transient studies. The major operating transients that should be 
addressed by a test facility were found to include: changes in compressor speed 
due to variations in vehicle engine speed; changes in condenser air flow rate due 
to ram-air effects at the front of the vehicle; changes in the passenger 
compartment air temperature during pulldown; and cycling of the compressor 
clutch to prevent evaporator frosting. Drive cycle tests were performed by using a 
square wave generator for the input to a compressor motor inverter and 
condenser fan inverter. The setting of the duration of square wave plateaus 
controlled compressor behavior during the drive cycle. The concepts of an 
"environment controller" and a "system controller" were briefly introduced, but 
details on the development of both were stated to be future research. 
P. Hrnjak ([4], [5], [6]), B. Hill ([7], [5]), J. Wertenbach [8], and Ward 
Atkinson ([9], [10]) presented work on the Society of Automotive Engineers (SAE) 
Alternate Refrigerant Cooperative Research Project. Steady-state tests to 
evaluate the coefficient of performance (COP) and cooling capacity, as well as 
transient tests to investigate soak and cool down performance were performed 
on an R134a system, as well as an enhanced R134a system and an R744 
system. Transient experimental analysis on secondary loop systems using 
3 
hydrocarbons as refrigerant was planned but has not been published at the 
present time. Three environmental chambers were used (indoor, outdoor, and 
compressor chamber) which enabled the initial temperature conditioning of all 
components. The indoor, as well as the outdoor chamber initially housed open 
wind tunnels. The automotive passenger cabin was simulated by adding a closed 
loop to the evaporator chamber. A water to air heat exchanger with controllable 
water flow rate was used to simulate the cooling rate of vehicle thermal mass. 
The water could be heated according to changes in temperature of the thermal 
mass of the vehicle. Coefficient of performance (COP), as well as evaporator 
capacity of the different systems were compared in steady-state performance 
tests, as well as transient New European Drive Cycle (NEDC) tests. Tests were 
conducted only up to 2,500 rpm. A model, which was validated with experimental 
results, extrapolated capacity and COP data at higher compressor speeds. 
However, the NEDC cycle and the USA equivalent, Federal Test Procedure (FTP) 
include significant portions of compressor speeds higher than 2,500 rpm. 
J. A. Baker et al. [11] presented research on building, testing and 
demonstrating a commercially viable, energy efficient secondary loop R152a 
MAC system. A test vehicle was equipped with a secondary loop system, and 
climatic wind tunnel tests, as well as road tests for cooling performance, were 
conducted. The wind tunnel tests included soak, cool down, as well as steady 
state performance tests and extended idle tests at an ambient temperature of 
40?C and 40% relative humidity. Results were presented in the form of average 
cabin air temperature and average air-conditioning (A/C) vent outlet temperature.  
4 
A. Gado ([12], [13]) built a dynamic simulator to simulate a passenger 
cabin while testing actual air-conditioning components under laboratory 
conditions on a test bench. Inputs to the dynamic simulator included: supply 
temperature, supply air humidity ratio, evaporator airflow rate, and various other 
user inputs, such as cooling loads and physical characteristics of the cabin. The 
mass and energy balance equations which were the core of the dynamic 
simulator were explicitly stated. It was shown that the dynamic simulator was 
able to run drive cycles, such as the NEDC, as well as pull down tests and cyclic 
tests with compressor on/off cycling. Instead of focusing only on compressor 
power consumption, the dynamic simulator allowed for real time monitoring and 
recording of refrigerant pressure and temperature, as well as air-side 
temperature and relative humidity. Cabin temperatures during NEDC cycle, as 
well as superheat and subcooling information were presented. 
M. B. Yahia and C. Petitjean [14] performed dynamic tests on a test bench. 
NEDC cycle, as well as steady state tests, based on European climate were 
performed. Two windtunnels, together with a variable speed compressor stand 
were used to test an NEDC cycle. The research focused on energy consumption 
of the compressor, however the control of the compressor speed was not well 
refined. Yahia and Petitjean found that average compressor power consumption 
during an NEDC differed about 5% from the steady state weight calculated 
average value. Test bench results and results from climatic windtunnel tests with 
actual cars did not agree well, and test conditions of the test bench had to be 
adjusted so that test bench results would resemble climatic windtunnel results 
5 
tested under the original test conditions. The authors concluded that if the A/C 
system could be controlled in short enough time intervals there should be little 
effect of transients on energy consumption, however no proof of this theory was 
provided in the test data shown. 
 
1.2 Transient Modeling of Automotive Air-Conditioning Systems 
C. Huang [15] conducted dissertation research on a dynamic simulation 
model for automobile passenger compartment climate control and evaluation. 
The objective of the dissertation was to develop a mathematical model to 
simulate dynamic features of mobile air-conditioning systems while predicting 
temperature and relative humidity inside the passenger cabin. The model 
consisted of two modules, one of them being the A/C network, and the other one 
being the passenger compartment. The passenger compartment was modeled 
as a lumped capacitance model. The model consisted of energy and mass 
balance equations that took the various loads, as well as the thermal and 
physical parameters of the cabin into consideration. Huang was able to 
transiently model the passenger cabin, showing the dynamic behavior of cabin 
temperature and relative humidity during a pull down procedure. 
M. Wang, D. Farley, and L. Leitzel [16] investigated head pressure spikes 
during vehicle acceleration by experiment and simulation. Head pressure spikes 
occur during gear shifts due to sudden acceleration of the compressor rpm and 
are a concern for system durability and passenger comfort requirements. Head 
6 
pressure spikes are mainly a mass balance phenomenon between the expansion 
device and the compressor, where the compressor pumps refrigerant faster than 
the expansion device can accommodate. The research focused on a 
mathematical transient model with application to A/C systems with scroll 
compressor. The developed model was able to simulate head pressure spikes 
effectively and allow conclusions on the phenomenon. It was found that scroll 
compressors have significant head pressure spikes due to their high volumetric 
efficiency. It was also found that the type of expansion device, as well as ambient 
conditions and heat transfer performance at the condenser during acceleration 
significantly affect the magnitude of head pressure spikes.  
T. Hendricks [17] conducted optimization of vehicle air-conditioning 
systems using transient air-conditioning performance analysis. The Sinda/Fluint 
analysis software was used to capture two-phase flow effects in evaporator and 
condenser, as well as system mass effects, air-side heat transfer, vehicle speed 
effects and temperature dependent properties. The A/C model was integrated 
with a simplified cabin model. Single-variable and multi-variable design 
optimizations were performed on COP, cabin cool down time, and system heat 
load capacity. The simplified cabin model was able to predict cabin and panel 
outlet temperatures during transient cool down periods and steady state 
operation. Combining the A/C system model and the cabin model allowed the 
prediction of drive cycle behavior and vehicle idle performance. Examples were 
presented in form of an optimization of COP, based on condenser tube diameter 
and transport line diameter. 
7 
R. Marzy, J. Hager, and C. Doppelbauer [18] investigated the optimization 
of vehicle warm-up, using the 1-D simulation tool KULI. To investigate vehicle 
warm-up, sub models for the vehicle engine, the HVAC system, and the vehicle 
cabin were combined into one model. A 3-D computational fluid dynamics (CFD) 
airflow simulation for the front end to model complex flows at the radiator and 
engine was added. Resistance curves were used on the air-side for the various 
heat exchangers and elements of the HVAC system in the 1-D model, which 
allowed for flexibility in adding or taking out components to model different 
system designs. Although the implementation of this software model was 
described and some transient simulation results were shown, validation of the 
model with experimental data was not provided. 
C. Tian and X. Li [19] evaluated the transient behavior of automotive air-
 conditioning systems using a variable displacement compressor. They built a 
mathematical model to evaluate variation of stroke length, based on change of 
suction pressure. As a result of their simulation effort, a time lag between system 
change and compressor wobble plate response of less than three seconds was 
found. Similarly, a time lag between piston stroke length and refrigerant mass 
flow rate was found to be less than five seconds. Experimental evaluation of 
transient response of variable displacement compressor parameters, such as 
rotary speed, piston stroke length, suction pressure and MFR was conducted. 
Both, experimental results and modeling results, showed that transient effects of 
the wobble plate are negligible and a variable displacement compressor can be 
modeled as a steady state component for transient modeling of HVAC systems. 
8 
H. Tummelscheit and D. Limperich [20] published information about the 
Air-Conditioning Library in the Dymola development environment for simulation of 
advanced A/C systems. The simulation tool Dymola, based on the language 
Modelica, was chosen by German original equipment manufacturers (OEMs) as 
the preferred tool for model development and library development. The A/C 
library is a commercial package which contains a complete range of component 
models and templates of typical A/C system architectures and refrigerants. 
Dymola has steady-state, as well as transient simulation capabilities and includes 
a dynamic process interface to show dynamic ph-diagrams and other 
visualization techniques. A model encryption allows the selective hiding of model 
source code, which unfortunately prevents convenient modification of existing 
models and addition of own code. Tummelscheit and Limperich used the Air-
 Conditioning Library with Dymola to compare the performance of R134a and 
Fluid "H". They concluded that fluid "H" shows lower capacity and COP and 
higher superheat and therefore cannot be used as drop-in refrigerant for R134a. 
I. Bayraktar [21] conducted research on time dependent simulation 
methods for vehicle thermal management. Underhood and HVAC optimization 
were considered, and both 1-D and 3-D calculations of component temperatures 
were conducted. Air flow and heat transfer were evaluated using 3-D CFD tools, 
namely Fluent and RadTherm software packages. Model results were compared 
to experimental results from a passenger compartment model in a climate 
chamber, instrumented with more than 100 thermocouples and a large number of  
pressure transducers on air- and refrigerant-side. To verify the model, the target 
9 
was to minimize difference between experimental and computational results for a 
cabin cool down process at four checkpoint locations (10, 20, 40 and 50 minutes). 
The author concluded that 1-D simulation can provide valuable first predictions, 
but higher fidelity 3-D CFD models are needed to analyze transient behavior of 
the entire thermal system of a vehicle. 
B. Li and A. Alleyne [22] presented a full dynamic system model with 
switched moving boundary components, which was able to accommodate severe 
transients in heat exchanger dynamics. The research was focused on start/stop 
cycles, specifically the phenomenon of compressor clutch cycling. Heat 
exchanger models were implemented in Thermosys, a Matlab toolbox. The 
dynamic system model was created in Matlab/Simulink and was successfully 
validated using an experimental test stand. In the moving-boundary modeling 
framework, heat exchangers were divided into control volumes or zones in terms 
of fluid phase. The location of the boundary between zones is a key dynamic 
variable that captures the essential multi- phase flow dynamics. It was shown 
that switching schemes between different representations (i.e. subcooled liquid, 
superheated vapor, or two-phase) handle the transitions of dynamic states while 
keeping track of vapor and liquid refrigerant regions during start-stop transients. 
The authors concluded that their model enables the adjustment of hysteresis set 
points on compressor cycling (i.e. at which temperatures the compressor turns 
on and off) in simulation before experimental testing is required. System 
responses like space temperature, refrigerant cycle pressure and superheat 
condition were successfully predicted by the model. 
10 
K. Sandhu, C. Chatham and A. Milosevic [23] developed a model to study 
the impact of various A/C circuit configurations on A/C performance of a Land 
Rover vehicle. In hybrid or electric vehicles, the A/C system design and resulting 
power consumption significantly affects driving range and cooling performance of 
batteries and electric systems. The simulation software GT-Suite was used. The 
model was validated at several steady-state conditions against experimental 
results from an R134a A/C system. A multi-evaporator design was studied with 3 
cooling loads, one of them being an intermediate heat exchanger cooling a 
secondary loop for battery and inverter cooling. Refrigerant R1234yf was used in 
the model. Transient pull down tests were simulated by integrating the A/C 
system model with a vehicle cabin model. A target evaporator temperature was 
calculated, based on ambient temperature, blower speed, solar intensity, 
recirculation mode and battery cooling demand. The target evaporator 
temperature was subsequently used to control compressor speed. Results of a 
pull down simulation in form of transient cabin temperature and transient 
compressor power were presented. However, no comparison with experimental 
data was provided. 
 
1.3 Alternative Refrigerants in Air-Conditioning and Refrigeration 
Table ?1.1 shows properties of five flammable refrigerants, along with 
R134a properties as baseline. Isobutane (R600a) is a hydrocarbon refrigerant, 
mostly used for medium- and low-temperature domestic refrigeration applications. 
With a low Global Warming Potential (GWP) of 8, an Ozone Depletion Potential 
11 
(ODP) of 0, and an Atmospheric Life Time (ALT) of less than a year, R600a has 
excellent environmental properties. R600a is in the safety classification A3 of the 
ASHRAE Standard 34 [24], meaning that it is highly flammable. It has a lower 
flammability limit (LFL) of 1.7 vol.%, which makes it the easiest to ignite among 
the refrigerants in the table. The acute toxicity exposure limit (ATEL) is a value 
used by ASHRAE Standard 34 [24] and ISO 817 [25] to establish the maximum 
refrigerant concentration limit for a refrigerant in air. It is calculated from the 
acute toxicity data using methods determined in accordance with the standards.  
Table ?1.1: Alternative Refrigerants and Characteristics (Wang et al.) [26] 
Refrigerants 
R134a 
(HFC) 
R290 
(HC) 
R600a 
(HC) 
R290/R600a 
mix. (HC) 
R152a 
(HFC) 
R1234yf 
(HFO) 
Global Warming Potential (GWP) 1,430 11 8 7 140 4 
Ozone Depletion Potential (ODP) 0 0 0 0 0 0 
Atmospheric Life Time [a] 14 <1 <1 <0.04 2 <0.05 
Acute Toxicity Exp. Level [ppm] 50,000 50,000 25,000 40,000 50,000 101,000 
Lower Flammability Limit [vol.%] - 2.2 1.7 2.0 3.9 6.5 
Minimum Ignition Energy [mJ]  - 0.25 0.25 N/A 0.38 >1000 
Safety Group A1 A3 A3 A3 A2 A2L 
Molar Mass [kg/kmol] 102.03 44.096 58.122 N/A 66.051 114.042 
Vapor Density (25?C) [kg/m3] 32.35 20.65 9.12 N/A 18.47 37.94 
Critical Temperature [?C] 101.1 96.8 135 114.8 113.3 94.7 
Critical Pressure [kPa] 4,061 4,247 3,647 4,040 4,522 3,382 
Normal Boiling Point [?C] -26.11 -42.11 -11.78 -31.5 -24.0 -29.48 
Vapor Spec. Heat (25?C) [kJ/kg K] 1.0316 2.0724 1.8189 1.77 1.2536 1.053 
 
Propane (R290) is a hydrocarbon (HC) refrigerant used for high-, medium-
 and low-temperature applications in commercial and industrial refrigeration, air 
conditioning, as well as heat pumps and chillers. With a GWP of 11, an ODP of 0, 
and an ALT of less than a year, R290 is an environmentally friendly choice. A 
12 
LFL of 2.2 vol.% puts it in the safety classification A3 with a MIE of 0.25 mJ 
required for ignition.  
An HC zeotropic mixture, a blend of R290 and R600a, is mostly used in 
high- and medium-temperature commercial, automotive and residential 
applications. As its constituents are R290 and R600a, it has a small GWP of 7, 
an ODP of 0, and a negligible ALT. Its LFL of 2.0 vol.% is in the typical range of a 
HC refrigerant, so that the assigned safety classification is A3. Due to 
flammability and safety concerns, this HC mixture, as its constituents, should be 
used in a SL system when used in MACs.  
1,1-Difluoroethane (R152a) is a HFC refrigerant, and was investigated for 
use in MACs, as well as household refrigerators. Its GWP of 140 is quite high 
compared to that of HCs, but is beneath the upper limit of 150, set by the 
European Council[27]. An ALT of about 2 years makes it less environmentally 
friendly compared to HCs, yet far superior to the currently used R134a. The 
advantage of R152a is the higher LFL of 3.9 vol.% and the resulting safety class 
A2. 
2,3,3,3-Tetrafluoropropene (HFO1234yf), proposed by American chemical 
companies, is currently investigated for use in MACs by Spatz and Minor [28]. Its 
GWP of 4 is the lowest among the reviewed flammable refrigerants. As R1234yf 
has zero ODP and a negligible ALT, it is the most environmentally friendly of the 
reviewed flammable refrigerants. A LFL of 6.5 vol.% and a MIE of greater than 
1000 mJ make it the safest choice among the assorted classification of 
13 
refrigerants in terms of flammability. The safety class to be designated is 
expected to be A2L. Due to its low LFL and MIE as compared to HCs, chemical 
companies proposed to apply R1234yf to DX systems without using a SL as a 
safety measure when used in MACs.  
Granryd[29] and Corberan et al. [30] summarized the environmental safety 
considerations and standards applied for the safe use of flammable refrigerants. 
Both ASHRAE Standard 34 [24] and European standard prEN378 (CEN, 2006) 
[31] classify refrigerants in three classes 1?3, where Class 1 is used for non-
 flammable fluids and Class 3 for highly flammable fluids. The group of Class 3 
refrigerants, which includes HCs, is currently limited in use for industrial 
applications in the USA and France, due to restrictive regulations. Some 
European standards, such as prEN378 (CEN, 2006), British standard BS EN378-
 1 (BSI, 2008) [32] and German standard DIN 8975 (DIN, 2004) [33], allow for a 
broader range of applications if certain safety requirements are met. As a result, 
the use of HCs in household refrigerators, freezers and small heat pumps has 
increased in European countries. 
Furthermore, Granryd [29] compared the performance of HCs, such as 
R600a and R290 and their mixtures to the well established refrigerants R12, R22 
and R134a. Blends of R290 with R600a could achieve vapor pressures close to 
that of R134a and would hence be a suitable replacement. While R290 has a 
higher volumetric cooling capacity, R600a and the HC blends have a lower 
volumetric cooling capacity compared to R134a. As a refrigerant with a high 
14 
volumetric cooling capacity will provide a high cooling capacity for a given swept 
volume of the compressor, R290 shows superior characteristics.  
Colbourne [34] summarized a study analyzing over 50 published technical 
documents comparing the performance of fluorinated refrigerants and HCs. A 
significantly higher number of tests showed an increase in performance when 
using HCs as compared to using fluorinated refrigerants, as shown in Colbourne 
and Suen [35]. The average improvements in Coefficient of Performance (COP) 
resulting from the use of HCs were 6.0% for domestic refrigeration applications, 
15.0% for commercial refrigeration applications, 8.8% for air-conditioning and 9.6% 
for heat pumping as shown in Table ?1.2.  
Table ?1.2: Comparison of Performance of Hydrocarbons per Application [26] 
Applications 
Proportion of cases where HCs improved COP 
Average 
Improvement < 10% 
Improvement 
10% - 20% 
Improvement 
> 20% 
Improvement 
Total 
Domestic Refrigeration 63.9% 13.9% 5.6% 83.7% 6.0% 
Commercial Refrigeration 51.6% 12.9% 25.8% 90.3% 15.0% 
Air-Conditioning 63.0% 25.9% 3.7% 92.6% 8.8% 
Heat Pumping 58.6% 37.9% 3.4% 100% 9.6% 
 
Colbourne and Ritter [36] investigated the compatibility of non-metallic 
materials with HC refrigerant and lubricant mixtures. Experiments were 
performed in compliance with European standards for the testing of elastomeric 
materials and ASHRAE material compatibility test standards. Test results were 
presented for swell rates, hardness rating, mass changes and the change of 
tensile strength. A R290/R600a HC blend in combination with various lubricants 
15 
VG32 Mineral Oil, VG100 Mineral Oil and VG46 Polyol Ester Oil, was tested. The 
chosen materials for testing were common materials, such as Buna N, HNBR, 
EPDM, Viton and Neoprene. Results for Buna N, HNBR, Viton and Neoprene 
indicated that linear and volumetric swell was minimal and acceptable. Buna N, 
HNBR, Neoprene and Silicone were found to show only minimal amount of mass 
change. The authors concluded that Buna N, Viton, Neoprene and HBNR, as 
well as Mylar as a plastic, were all suitable elastomers for use with HC 
refrigerants.  
Maclaine-Cross and Leonardi [37] compared the refrigerant performance 
of HCs based on refrigerant properties and concluded that the COP 
improvements, commonly reported in literature, were consistent with better 
thermodynamic properties of HCs. R600a properties and their influences on 
system performance were discussed. In refrigeration systems, the evaporator 
has to withstand pressures similar to those in the condenser during the off period. 
Since the typical condensing pressures of R600a is about 50% smaller than that 
of R134a and other refrigerants, the thickness of the heat exchanger material can 
be reduced. This helps the heat transfer in heat exchangers through a reduction 
of thermal resistance, and thus reduces capital costs and environmental impact. 
Additionally, the low compressor discharge temperature of R600a allows for a 
cheaper and more efficient design of the electric motor. They suggested that 
although performance differences between ideal cycles using R600a and 
common refrigerants were small, the flow and heat transfer parameters were 
typically better by a factor of two for R600a. This explains the 20% energy 
16 
savings reported in literature. Through redesign of components, similar 
improvements can be achieved in other small applications.  
Joudi et al. [38] studied the performance of MAC systems with alternative 
refrigerants. A computer model was developed to determine the most suitable 
alternative refrigerant to R12. The influence of evaporating temperature, 
condensing temperature and compressor speed in an ideal cycle was considered. 
Four refrigerants were studied: R134a, R290, R600a and an R290/R600a 
mixture. The model predicted the R290/R600a mixture as the most suitable 
alternative to R12. A subsequent experiment with a prototype MAC system 
compared R290, the HC blend and R12 performance. It was observed that R290 
outperformed R12 by approximately 8.3% in terms of COP, depending on 
condensing temperatures, whereas the R290/R600a mixtures had a similar 
performance to R12. For the pull-down time at soaking temperatures, a superior 
performance of the HC blend was recognized for all test conditions. The power 
consumption of the HC mixtures was slightly higher for all operating conditions 
investigated, resulting in slightly lower COPs between 0.86% and 2.27%, based 
on operating conditions. 
Ghodbane [39] investigated the use of R152a and HCs in MACs. Based 
on thermophysical data, a quantitative analysis of MACs with flammable 
refrigerants was proposed. R290 having the lowest normal boiling point (NBP) 
was expected to have the highest operating pressure, while R600a was expected 
to have the lowest operating pressure. R152a was expected to have a similar 
operating pressure to R134a since its NBP is closest to R134a. Critical 
17 
temperatures of R152a, R290 and R600a are all above 93.3?C. An ideal 
performance comparison with respect to R134a was performed over a wide 
range of evaporating and condensing temperatures. The cooling capacity of the 
system was set to be 3.7 kW. Pressure drop was neglected and the compressor 
isentropic efficiency was set to be 100%. Results showed that over the whole 
range of evaporating temperatures, R152a had the highest COP in the range of 
37.8?54.4?C condensing temperature, which is the desired range for MACs. The 
COP of R152a was increased about 6?19% as compared to R134a, depending 
on condensing temperature. R152a showed a higher hot discharge gas 
temperature compared to other refrigerants, resulting in 6.6?14.1 K higher 
temperatures compared to R134a. Despite the higher discharge temperature, 
R152a was chosen by Ghodbane as the most suitable flammable refrigerant to 
replace R134a in MACs, due to its superior performance. A more realistic 
evaluation of the refrigerants under operating conditions in a vehicle with 
simulation of road load conditions followed. R152a remained the best performer 
with a performance increase over R134a of 11% at road load conditions and 8% 
at idling conditions. R152a had a condensing heat transfer coefficient of about 
1.33 times higher than R134a under the prescribed conditions for idling (1.26 for 
road load). The evaporative heat transfer coefficient of R152a increased to 1.04 
times of R134a at the same ambient conditions under idling (1.14 times for road 
load). The highest increase in evaporative heat transfer coefficient was reported 
for R290 with 1.43 times of R134a for idling (1.51 times for road load). Two-
 phase flow pressure drop data showed that the HCs, as well as R152a, have a 
18 
considerably low pressure drop, with R152a having 50% less pressure drop than 
R134a. 
 
Risk assessment 
In their research ??Hydrocarbon?Refrigerant?Risk?in?Car?Air-Conditioning??,?
 Maclaine-Cross and Leonardi [40] provided an overview of the potentially 
dangerous scenarios that need to be considered when employing flammable HC 
refrigerants in MAC systems. Principal hazards associated with refrigerants were 
determined to be: explosion in space, fire, asphyxiation or poisoning, flying debris 
resulting from an explosion, cold burns and damages due to chemical reactions. 
Several safety standards were adopted in Europe, Australia, Asia and the USA, 
which restrict the use of flammable refrigerants based on the charge amount, 
safety design of the system, as well as the individual properties of the 
refrigerants. Based on the necessary precautions demanded by these standards, 
the authors believe that R600a and mineral oil mixtures require the least 
expenditure on precautions for MAC systems. This combination is already 
commonly used in household refrigerators.  
Razmovski and Rajasekariah [41] experimentally evaluated possible 
ignition sources in a car by connecting a welding torch to a HC refrigerant 
cylinder. The refrigerant streaming out of the welding torch was then tested for 
more than 15 min on the hot engine, electrical wiring, exhaust system, ignition, 
switches, fan motor and internal cigarette lighter. It was found that none of the 
19 
aforementioned sources ignited the torch and a lit up match or external butane 
cigarette lighter was the only possible ignition source for the HC refrigerant in a 
car. One scenario that particularly causes concerns is an explosion of flammable 
refrigerant within the passenger compartment through rupturing of a pipe just 
before the expansion valve, which releases a refrigerant cloud into the passenger 
compartment. It is expected that a fatigue fracture of the liquid line could be a 
likely cause of this scenario.  
Maclaine-Cross and Leonardi [40] concluded from their experiments that 
safe operation of HC refrigerants in cars is possible. In a study about HC 
refrigerant leakages in car passenger compartments, Maclaine-Cross referred to 
a report made by Arthur D. Little Ltd. [42], who noted that serious injury to 
occupants through use of flammable refrigerant would only be possible if the car 
crashed, due to overpressure in the compartment after a fatigue damage of the 
liquid line. The risk level of such an event, which could also occur with non-
 flammable refrigerants, was estimated to be 4.16x1010 per car year. Most 
refrigerants, not only HCs, are asphyxiants and causes drowsiness and driver 
fatigue. The potential hazards implied are thousands of times more frequent than 
the ignition of refrigerants and have a higher potential to entail damage through 
possible crashes. Peak concentrations in fractions of the LEL were determined 
for a variety of Australian cars, built from 1970 to 1989. For models built in the 
1970s, the concentration in the cabin was found to be above the LEL only for an 
instantaneous complete fracture. In all cars built in the 1980s, the possible 
concentration after one of the aforementioned events was well below the LEL.  
20 
A report from Dieckmann et al. [43] for the USA Department of Energy 
was reviewed, which assessed the risk of using flammable refrigerants in MACs. 
Field data from car crashes and car fires was used as basis for the analysis. The 
extent of damage to the refrigerant circuits in crashed cars was visually inspected. 
A fault tree analysis allowed for risk estimates. 
A similar risk assessment, performed by Elbers and Verwoerd [44], 
considered an R290 heat pump system used for residential heating. The 
estimated risk in terms of fatalities is provided in above table. To provide a 
context for these safety estimates, Ritter and Colbourne [45] published a review 
on HC risk assessment from 1991 to 1998. The use of background risks as a 
basis for comparison of the risk of fire with HC was presented. The background 
risks are then taken as a datum for comparison with the risk of a fire due to the 
use of flammable refrigerant. Several systems were considered by Ritter and 
Colbourne, including MAC, supermarket cold storage rooms, supermarket line 
chest freezers and integral unit chillers. A typical MAC system with a charge of 
400 g of HC mixture was assessed to have a risk level of 3.0x107 per car year. 
The risks considered included minor burns from jet fires, cloud fires and 
overpressure. The risk level of a car crash, resulting from leakage of a HC 
mixture was subsequently estimated to be 9.91x108 per car year. This number is 
negligible compared to the background risk of fire frequency in cars, which is 
3.29x103. Based on the foregoing comparison, Ritter and Colbourne [45] 
concluded that the risks of flammable refrigerant leakages and possible 
21 
subsequent fires, evaluated for numerous systems, are mostly negligible when 
compared to background risks.  
Jetter et al. [46] used a fault tree analysis to estimate the number of 
refrigerant exposures of automotive service technicians and vehicle occupants in 
the USA. The estimated number of exposure for occupants in a vehicle is smaller 
than the exposure for technicians. Furthermore the largest best estimate was a 
leakage following a collision. The authors claimed that a common concern, the 
leakage from the evaporator in the passenger compartment following a collision, 
might be overrated because for parked vehicles and those standing still after an 
accident, refrigerant leaking out of the evaporator would flow through the 
condensate drain due to its high density compared to air and would not enter the 
passenger compartment in a significant amount.  
A quantitative risk assessment model was developed by Colbourne and 
Suen [47] to examine the influence of design, installation of equipment and 
external conditions on the frequency of ignition and the associated 
consequences for indoor refrigeration and air-conditioning units using HC 
refrigerants. The output of the model was evaluated with a sensitivity analysis of 
the parameters to show their relative impact on the overall risk level. Key aspects 
in terms of design were shown to be refrigerant charge quantity and installation 
height due to their impact in the event of quick dispersion. Also the refrigerant 
selection itself showed a noticeable impact on the risk value, where R600a 
showed the highest risk. 
22 
The above literature review on low GWP refrigerants showed that 
flammable refrigerants are safely used in a variety of industrial, commercial and 
residential (mostly Europe/ Asia) settings. Cars with MACs running with Propane 
(R290) exist in Australia. Simulation and experimental research on flammable 
refrigerants in MACs shows that R152a and R290, as well as mixtures of R290 
and R600a are promising candidates to be used in future automotive air-
 conditioning systems. Hydrocarbons, such as R290 and R600a, are superior with 
regard to environmentally friendly operation. At the same time the supply of HCs 
can be reliably guaranteed in most parts of the world. Risk assessment studies 
showed that the risk of fire due to the use of hydrocarbons as refrigerant in direct 
expansion systems (DX) is low compared to background risks. It is generally 
agreed upon that it is safe to use hydrocarbons in MACs if safety measures are 
taken. The secondary loop is one of these safety measures which, in addition to 
fire hazard mitigation, includes other advantages, such as cold storage and 
enabling of multiple cooling loads.  
 
1.4 Secondary Loop Systems in Air-Conditioning and Refrigeration 
The vapor compression cycle of a secondary loop (SL) system consists of 
a compressor, a condenser, an expansion device and an intermediate heat 
exchanger, often called chiller. The chiller cools down a working fluid in the 
secondary loop, which in turn cools down the air by exchanging heat in the cooler, 
a liquid to air heat exchanger. One or more pumps cycle the coolant in the 
secondary cooling circuit. 
23 
Primary Loop Equipment 
Most of the components used in DX systems can also be used for SL 
systems, which is also true for the compressor. Palm [48] reported that HC 
producers listed the compressor manufacturers whose compressors are 
compatible with HCs, including 23 compressor manufacturers. After taking an 
internal evaluation of safety aspects, Danfoss [49] decided to supply 
compressors used in small hermetic systems with HC charge of 150 g or less. 
Corberan et al. [50] investigated the performance of a positive displacement 
hermetic refrigerant piston compressor working with R290 as refrigerant. The 
cooling capacity of an R22 compressor that was switched to R290 operation 
decreased by 13% to 19%. However, the COP of the system increased from 2 to 
6% at the same time. 
The intermediate heat exchanger, often called chiller, is of special interest 
in SL systems. Pellec et al. [51] tested two types of heat exchangers (brazed- 
and platular-welded-plate heat exchangers) working with ammonia and silicone 
heat transfer fluid as a secondary refrigerant. With the same mass flow rate on 
the secondary refrigerant-side, the heat transfer coefficient is about five times 
less for the platular- welded-plate heat exchanger compared to the brazed plate 
heat exchanger, due to the plate patterns. Setaro et al. [52] tested and compared 
heat transfer and pressure drop for a brazed plate heat exchanger and a tube-
 and-fin coil using two different refrigerants, R22 and R290 in an air-to- water heat 
pump system. Both, the heat transfer coefficient and pressure drop of R290, are 
lower than R22 in the two types of heat exchangers which were studied. Hrnjak 
24 
and Hoehne [53], who studied charge minimization in SL systems, reported that 
an air to R290 minichannel heat exchanger developed for a 2 kW cooling 
capacity refrigeration system needed less than 0.13 kg of R290 due to its smaller 
internal volume compared to traditional fin-and-tube heat exchangers. Fernando 
et al. [54] studied liquid-to-refrigerant heat exchangers using flat multiport with 
1.4 mm hydraulic diameter tubes and showed a lower charge compared to plate 
heat exchangers. Fernando et al. ([55], [56], [57]) also carried out comprehensive 
tests on performance of minichannel aluminum tube heat exchangers working as 
evaporator and condenser. Correlation equations of refrigerant-side (R290) heat 
transfer coefficient were compared and validated with experimental results.  
 
Primary Loop Refrigerants 
Primary refrigerants for commercial systems can be distinguished into 
Ammonia, hydrofluorocarbons (HFCs) and HCs. A lot of examples exist for 
Ammonia and HFCs as primary refrigerants, including American manufacturers, 
as shown by Likes [58], who established an early SL system for supermarkets. 
The primary refrigerant was R22 and the secondary refrigerant was a water-
 propylene glycol solution, 65/35% by mass. Preliminary test results showed that 
the performance of the secondary system for the medium temperature case was 
satisfactory, but the energy consumption was higher than expected. Rolfsman 
[59] reported a supermarket in Sweden which was converted to a SL system. 
NH3 was used as the primary refrigerant and CO2 was used as the secondary 
refrigerant for freezing. Five of the same systems were installed and had been 
25 
operating successfully for a year. Evenmo [60] cited a supermarket in the United 
Kingdom using R407C as the primary refrigerant and a commercial fluid as the 
secondary fluid, since first opened in February 1997. In the same year, Horton et 
al. [61] tested a drop-in SL refrigeration system for medium temperature 
supermarket applications. The primary refrigerant was R22 and the secondary 
refrigerants were a 50/50% mixture of water-propylene glycol, and hydro-
 fluoroether. Comparison of the performance of the R22 baseline vapor 
compression cycle and SL systems showed that the COP of the primary loop in 
the SL setup was about 1.5?2 times higher compared to the baseline. Nyvad and 
Lund [62] reported on a supermarket in Denmark replacing its existing (H)CFC-
 plant with a new indirect SL system. In this system, NH3 was used as the primary 
refrigerant and Tyfoxit (Tyforop Chemie Gmbh [63]) was used as the secondary 
refrigerant. The experimental results showed that average energy consumption 
of the SL system with NH3 as the primary refrigerant was 35% less compared to 
the old system using (H)CFC as the refrigerant. Arias and Lundqvist [64] reported 
field test results of advanced systems in three supermarkets. Two of them used a 
cascaded SL system with R404A as primary refrigerant and potassium formate 
brine solution as secondary refrigerant. The third system had individual SL 
refrigeration units in each display case which were all connected to a central 
building chiller for heat rejection. Faramarzi and Walker [65] tested the 
performance of a SL refrigeration system in USA supermarkets. The primary 
refrigerant was R507 and secondary refrigerant was Dynalene. The energy 
consumption of this SL refrigeration system was 4.9% less compared to the 
26 
baseline multiplex system. Minea ([66], [67]) reported a supermarket refrigeration 
system with SLs installed near Montreal, Canada. This advanced system 
involved secondary fluid loops on both refrigerating and condensing sides, heat 
recovery with brine- to-air heat pumps and passive heat exchangers. The primary 
refrigerant was R507 and the secondary refrigerants for low temperature freezing 
loop and medium temperature refrigeration loop were potassium formate and 
propylene glycol (50/50% by mass), respectively. The secondary fluid on the 
condensing side was ethylene glycol/ water mixture (50/50% by mass). The total 
quantity of the primary refrigerant was reduced by 61% compared to a baseline 
multiplex refrigeration systems.  
Examples for HCs as primary refrigerant include Rivers [68], who reported 
on an SL refrigeration system designed for a supermarket in Greenwich, England. 
An HC was chosen as primary refrigerant. Propylene glycol was employed as the 
medium temperature secondary refrigerant, while an organic potassium salt 
solution was chosen for the low temperature loop. Baxter [69] reported a case 
study for a small Danish supermarket where the old refrigeration plant had been 
replaced with a cascade plant. R290 was used as the high temperature 
refrigerant (14/30?C) while CO2 was used as the low temperature refrigerant 
(32/11?C). CO2 was used directly to perform the cooling in the freezers while 
propylene glycol was used in the coolers. Total energy consumption was 
decreased by 10% with the new plant. 
Primary refrigerants mostly used in residential and automotive air-
 conditioning are HCs, CO2, or R152a (used in MACs). Examples for 
27 
hydrocarbons include Choi et al. [70], who evaluated the performance of R22, 
R290, R290/600a (70/30%), and R32/152a (50/50%) used in a water-to-water 
residential heat pump for space cooling and heating. The secondary fluid was a 
60/40% mixture of water and ethylene glycol. For the same system capacity, 
R32/152a proved to be the best performer due to a good temperature glide 
matching in the heat exchangers and its excellent thermodynamic and 
transporting properties. The HC mixture R290/600a had the highest COP at a 
given compressor speed, but its COP at the constant-capacity criterion was the 
lowest. Chang et al. [71] reported the performance and heat transfer 
characteristics of a heat pump system charged with HC refrigerant (R290, R600a, 
R1270 and binary mixture of R290/ R600a and R290/R600). The secondary fluid 
was ethyl alcohol. Test results showed the heating and cooling capacity of R290 
was slightly less than that of R22. The COP of R290 was slightly higher 
compared to R22. The heating and cooling capacities of R290/600a and 
R290/600 mixtures were nearly linear with respect to mass fraction of R290. 
Pelletier and Palm [72] tested a domestic heat pump using R290 as compared to 
the R22 baseline system. For R290, the heating capacity was 7?10% lower, 
while the heating COP was 4?5% higher than R22. Payne et al. [73] investigated 
and compared the performance of R22, R290 and zeotropic mixtures of 
R32/R290 and R32/152a. The secondary fluid was 70/30% mixture of water and 
ethylene glycol. In the cooling mode at a constant capacity, the R32/290 (50/50%) 
mixture resulted in the highest COP, 8% higher than R22. In the heating mode, 
the COP of R290 was the highest, 6?8% higher than that of the remaining fluids.  
28 
Examples using CO2 as primary or secondary refrigerant in residential 
systems include Stene [74], who investigated the performance of a residential 
brine-to-water CO2 heat pump for combined low temperature space heating and 
hot water heating. The system heating capacity was 6.5 kW. Yanagisawa et al. 
[75] investigated a SL refrigeration system, using a vapor compression NH3 
cycle as the primary loop and a CO2 thermosiphon loop as the SL. In the SL, CO2 
was circulated between the intermediate heat exchanger and the cooler by the 
difference in density between liquid and vapor lines. The experimental SL with 
the CO2 liquid head of 0.9 m operated at a cooling capacity of 3.5 kW and COP 
of 3.7. 
In automotive air-conditioning, natural refrigerants, such as HCs, are 
considered due to the currently used refrigerant, R134a being phased out by 
2011 in the European Union. HCs present a potential alternative to R134a due to 
their good thermodynamic and transport properties, heat transfer characteristics, 
material compatibility, low cost, low toxicity and low GWP (Corberan et al. [30]; 
Domanski and Yashar [76]; Fernando et al. [54]; Mani and Selladurai [77], Palm 
[48]). The chemical industry has been working to synthesize new chemicals and 
mixtures of chemicals that have a GWP below the required threshold, such as 
R152a (Ghodbane [39]) and R1234yf (Hill [78]; Spatz and Minor [28]). The HCs, 
R152a, and NH3, have received less attention compared to other refrigerants 
because of safety concerns, such as flammability or toxicity. The USA 
Environmental Protection Agency (EPA) does not explicitly forbid the retrofit of 
R134a systems with Hydrocarbons, though some states do [79]. However, HCs 
29 
are declared unacceptable substitutes on the basis of a current lack of adequate 
assessment of their flammability risk by the U.S. EPA as of 2010 [80]. Therefore, 
most published papers on these refrigerants being employed in automotive direct 
expansion systems are theoretical in nature. However, one way of safely 
employing HCs, R152a or NH3 in MACs is to use them in conjunction with a SL 
system. An SL system would prevent the primary refrigerants from entering the 
passenger cabin and thereby mitigate any fire hazard for vehicle occupants.  
 
Secondary Refrigerants 
There are two kinds of secondary refrigerants: single phase fluids and 
two-phase fluids. Single phase fluids generally include some kind of antifreeze 
solution, corrosion inhibitor, or biocides. A series of glycol-, potassium acetate-, 
and potassium formate-based secondary refrigerants have been applied in SL 
refrigeration systems (Ure [81]). There are few SL refrigeration systems that use 
a two-phase mixture as secondary refrigerant. Two-phase secondary refrigerants 
take advantage of the high latent heat during the phase change process from 
liquid to solid or from liquid to gaseous state (Delventura et al. [82]). The higher 
the percentage of refrigerant undergoing the phase change, the higher the 
energy stored per kilogram of refrigerant. The freezing point is the most important 
physical property to consider when choosing among fluids. Usually the freezing 
point of the chosen secondary refrigerant is not less than 5?10 K below the 
system operating temperature. Aqueous solutions, due to their high specific heat 
and good heat transfer coefficients, are widely used as working fluids in the 
30 
secondary circuit. Most commonly used aqueous solutions with their 
thermophysical properties are listed in Table ?1.3 (Melinder [83]).  
Table ?1.3: Characteristics of Aqueous Solutions as Secondary Refrigerants [26] 
Aqueous Solutions 
cA  
[wt.%] 
tF 
 [?C] 
t  
[?C] 
?  
[kg/m
 3
 ] 
cp  
[kJ/kg K] 
k 
[W/m K] 
?  
[mPa s] 
Ethylene Glycol 36.19 -20 -10 1058 3.494 0.415 7.99 
Propylene Glycol 39.41 -20 -10 1045 3.620 0.384 21.67 
Ethyl Alcohol 29.73 -20 -10 970 4.127 0.393 12.22 
Methyl Alcohol 24.87 -20 -10 969 3.914 0.431 5.42 
Glycerol 46.28 -20 -10 1129 3.155 0.460 2.94 
Ammonia  13.46 -20 -10 953 4.239 0.460 2.94 
Potassium Carbonate 30.95 -20 -10 1322 2.877 0.519 6.25 
Calcium Chloride 20.82 -20 -10 1196 2.992 0.530 4.48 
Magnesium Chloride 16.59 -20 -10 1151 3.192 0.505 5.58 
Sodium Chloride 22.62 -20 -10 1182 3.305 0.530 4.22 
Lithium Chloride 13.66 -20 -10 1083 3.405 0.522 3.97 
Potassium Acetate 27.91 -20 -10 1155 3.232 0.467 5.67 
Potassium Formate 28.94 -20 -10 1192 3.121 0.506 3.55 
 
Other secondary refrigerants, which are widely used include ice slurries 
and CO2. Ice slurries consist of a number of ice particles in an aqueous solution 
where the diameter of ice particles is equal or smaller than 1mm, as stated by 
Egolf and Kauffeld [84]. The smaller the ice particles are, the better the slurries 
can be transported. The cooling capacity of an ice slurry is four to six times 
higher than that of conventional chiller water, depending on the ice fraction (Ure 
and Mashrae [85]). Often cooling point depressants, such as salts or ethanol are 
used, as researched by Meewisse and Ferreira [86] who compared two freezing 
point depressants, sodium chloride and ethanol. The results showed that 
economic performance of SL cooling cycle utilizing sodium chloride as freezing 
point depressing substance was at most 8% higher than that of ethanol. 
31 
Examples of successful integration of ice slurries in SL systems include 
Wang and Goldstein [87], who applied an ice slurry system for cooling in railway 
transportation, Christensen and Kauffeld [88], who described the application of 
ice slurry as secondary refrigerant in a SL system with the use of an ice slurry 
accumulation tank, and Wang et al. [89] who studied an SL ice slurry system 
using ethylene glycol-water binary solution in the Ritz Carlton Plaza Hotel in 
Japan. Fukusako et al. [90] reviewed studies related to the cold thermal storage 
systems and components using ice slurries and recent research activities on ice 
slurries in Japan. The types of ice storage systems were categorized into five 
groups by the process of ice manufacturing. Saito [91] reviewed recent research 
on cold thermal energy storage including a SL ice slurry system. The ice making 
process, ice storage process, transportation and utilization of stored ice, as well 
as the whole process were identified as important future research topics. 
Recently more and more attention has been paid to Carbon Dioxide (CO2), 
which shows several benefits as a two-phase secondary fluid. The advantages of 
utilizing CO2 as a secondary refrigerant are lower pumping power (zero pump 
power when thermosiphon is used), smaller pipe size, excellent heat transfer 
properties, and good material compatibility with the additional benefit of the low 
cost of the fluid compared to conventional single-phase secondary refrigerants. 
The main disadvantage of CO2 appears to be relatively low critical temperature 
and limited availability of components (Hinde et al. [92]). A few applications that 
utilize CO2 as a volatile secondary refrigerant have been implemented in low-
 temperature applications in supermarket systems in Sweden (Melinder [93]; 
32 
Pachai [94]; Pearson [95]). Pearson [96] submitted patents on the use of CO2 as 
a volatile secondary refrigerant, including a novel hot gas defrost system.  
Further examples on CO2 as secondary refrigerant include Christensen 
[97], who investigated a SL system using CO2 as primary and secondary 
refrigerant in supermarket applications. Tests showed that CO2 was a suitable 
alternative to HFCs in supermarket applications. Pachai [94] reported a SL 
system installed in Helsingborg, Sweden. The primary refrigerant was a HC, a 
mixture of R290 and R170, and the low- and intermediate-temperature side 
secondary refrigerants were CO2 and propylene glycol, respectively. The system 
had successfully been running since 1997 and more than fifty shops have been 
installed with similar systems in Sweden. Nilsson et al. [98] reported an ice rink 
refrigeration system with CO2 as the secondary fluid with a cooling capacity of 15 
kW. Also in 2006, a British company implemented an air-conditioning system to 
protect computer servers against overheating for a bank at its London office, as 
reported by Jahn [99]. The chilled water circuit condensed CO2 at 6?C. Then, 
CO2 was pumped out to the server cabinets by centrifugal pumps, evaporated, 
and returned to the condenser to begin the process again. The cooling capacity 
of the whole system was 300 kW. Hinde et al. [92] reported that at least nine low-
 temperature CO2 systems were operational in the USA and Canada in early 2008. 
All systems utilized CO2 as a low-temperature two-phase secondary refrigerant. 
The stores ranged in size from small neighborhood markets to large 
supercenters and warehouse-style stores. The system's cooling capacities 
ranged from 22 to 160 kW.  
33 
Secondary Loop Systems in Automotive Applications  
Examples of secondary loop systems in MACs include Ghodbane[39], 
who investigated the potential of R152a and HC refrigerants as alternative 
refrigerants to R134a and published a comparative assessment of a SL system 
applied to MACs. The assessment showed that the SL for MACs requires less 
development time compared to other alternatives, such as CO2, air cycle, etc. 
Ghodbane expected the SL system to have higher mass and cost than the 
baseline R134a system, but to offer the potential of being lower mass and cost 
compared to the other alternatives to R134a systems. The performance of a SL 
MAC with R152a or HCs as primary refrigerant and ethylene glycol-water mixture 
in 50/50 wt.% as secondary refrigerant was compared theoretically to a 
traditional DX system using R134a. Based on the theoretical comparison, the 
author decided to experimentally test a SL system using R152a as primary 
refrigerant. Comparison of the performance of the SL system with a production 
R134a system as baseline is shown in Table ?1.4 at 87.5 km/h road load and 
ambient conditions of 37.8?C and 40% RH. The SL system COPs were 
decreased by 5?19% while the condenser capacities were increased by 6% 
when R152a was used as a primary refrigerant. The contributor to higher 
compressor power and higher condenser capacity for the SL system was a low 
suction condition, which was dictated by the cooling capacity requirement. Added 
pump power in combination with the indirect contribution from low suction 
conditions were found to result in a low COP compared to DX systems.  
34 
Table ?1.4: Comparison of R152a to R134a in an Automotive Secondary Loop 
System [39] 
Parameter 
Refrigerant / System 
R134a DX R152a DX R134a 2LP R152a 2LP 
Ambient Temperature [?C] 37.8 37.8 37.8 37.8 
Ambient Relative Humidity [?C] 40 40 40 40 
Humidity Ratio [kg/kg dry air] 0.017 0.017 0.017 0.017 
Evap. Air Flow Rate [m
 3
 /h] 424.8 424.8 424.8 424.8 
Evap. Air Mass Flow Rate [kg/min] 7.8 7.8 7.8 7.8 
Evap. Outl. Air Temperature [?C] 10 10 10 10 
Cond. Air Flow Rate [m
 3
 /h]  3,398 3,398 3,398 3,398 
Cond. Air Mass Flow Rate [kg/min] 62.6 62.6 62.6 62.6 
Cond. Outl. Air Temperature [?C] 47.8 47.2 48.3 47.8 
Refrigerant Charge [kg] 0.96 0.62 0.96 0.62 
Refr. Mass Flow Rate [kg/min] 3.14 1.82 3.28 1.88 
Comp. Suction Pressure [kPa] 300 295 214 213 
Comp. Suction Temperature [?C] 3.6 4.4 -5.6 -5.0 
Comp. Discharge Pressure [kPa] 1,723 1,544 1723 1,544 
Comp. Discharge Temperature [?C] 90.6 107.2 95.6 123.3 
Comp. Isentropic Efficiency 60% 60% 60% 60% 
Comp. Power [kW] 3.24 2.89 4.03 3.89 
Cond. Outl. Pressure [kPa] 1,640 1,510 1,640 1,510 
Cond. Outl. Temperature [?C] 50.6 51.7 50.6 51.7 
Cond. Capacity [kW] 10 9.6 10.8 10.6 
Evap. Inl. Pressure [kPa] 368 326 263 235 
Evap. Inlet Temperature [kPa] 6.7 5.6 -2.8 -3.3 
Evaporator Effectiven./Overall Eff. 85% 85% 62% 62% 
Evap. Latent Load [kW] 3.1 3.1 3.1 3.1 
Evap. Sensible Load [kW] 3.6 3.6 3.6 3.6 
Evap. Cooling Capacity [kW] 6.7 6.7 6.7 6.7 
Comp. Power [kW] 3.2 2.9 4.0 3.9 
Comp. Power Loss [kW] 0.097 0.087 0.12 0.12 
A/C Blower Power [kW] 0.28 0.28 0.28 0.28 
System Power [kW] 3.6 3.3 4.6 4.4 
Cycle COP [-] 2.072 2.322 1.666 1.726 
System COP [-] 1.863 2.065 1.464 1.51 
 
Dentis et al. [100] compared the SL system with R152a and HC 
refrigerants to an R134a system on a test bench, and demonstrated that the 
performance of SL system was similar to, and in some cases exceeded, the 
35 
performance of an R134a system. Ghodbane [101] also compared the 
performance of MAC SL systems to conventional R134a systems used in a small 
size passage car under the same test conditions in a windtunnel and in road 
performance tests. The SL system delivered equal or better performance and 
comfort levels compared to the production R134a system.  
Ghodbane and Baker [102] received a patent on the SL system for 
passenger compartment heating and cooling. Another patent was issued to 
Kadle and Ghodbane [103] for a heat pump using SL air-conditioning system. An 
R152a SL system with pre-prototype components installed in a sport utility 
vehicle was demonstrated by Ghodbane et al. [104]. They built and tested the 
performance of an R152a SL system, in a compact passenger car. The energy 
consumption of the MAC compressor in the SL system was shown to be 13?16% 
lower compared to the production R134a system when applying a capacity 
control algorithm. The global warming impact of direct refrigerant emissions of 
the R152a SL system was claimed to be lower by at least 94% compared to an 
equivalent R134a system. 
 
1.5  Literature Review Summary 
? Few researchers did transient tests with MACs in controlled laboratory 
environments. Transient tests, such as drive cycle tests, are usually 
performed in a climatic windtunnel, which is associated with high 
expenses. At the same time, climatic windtunnel tests limit the amount 
of instrumentation and the accuracy of results that can be obtained.  
36 
? Researchers which performed dynamic tests with MACs usually 
operated in narrow operating conditions, often limited by the employed 
equipment. In few cases where drive cycle research was conducted, 
the drive cycle control was rigid and of limited accuracy. 
?  A. Gado developed a cabin model which enabled transient testing of 
MAC components without the need for industry scale climatic 
windtunnels.  
? Most pure modeling research and simulation studies are limited to 
simple transient studies, such as pulldown scenarios.  
? Commercial software was used in many cases which limits access to 
the source code and the flexibility of adding own component models.  
? The software Dymola shows a potential to be universally employable in 
transient studies of MAC systems. Dymola is based on a modular 
approach with a variety of component libraries which can be updated 
by adding new components or altering the source code of existing 
components. 
? In general, publications on research in automotive air-conditioning are 
rare when compared with other fields. The automotive industry often 
chooses to not disclose complete statements of testing conditions or 
exact system parameters.  
? The review on low GWP refrigerants showed that mildly flammable to 
flammable refrigerants are safely used in a variety of industrial, 
commercial and residential (mostly Europe/ Asia) settings. Cars with 
direct expansion MACs running with Propane (R290) exist in Australia.  
? Simulation and experimental research on flammable refrigerants in 
MACs shows that R152a and R290, as well as mixtures of R290 and 
R600a are promising candidates to be used in future automotive air-
 conditioning systems.  
37 
? Hydrocarbons, such as R290 and R600a, are superior with regard to 
environmentally friendly operation. At the same time the supply of HCs 
can be reliably guaranteed in most parts of the world.  
? Risk assessment studies showed that the risk of fire due to the use of 
hydrocarbons as refrigerant in direct expansion systems (DX) is low 
compared to background risks. It is generally agreed upon that it is 
safe to use hydrocarbons in MACs if safety measures are taken. The 
use of secondary loop system is one of these safety measures.  
? A variety of alternative  flammable refrigerants can be used as primary 
refrigerants in SL systems. Considering the desired characteristics for 
automotive systems, such as normal boiling point and specific volume, 
R152a and R290 have been shown to operate well in SL systems.  
? Experimental research on R152a SL MAC systems was done by 
Ghodbane et al., who showed the viability of automotive SL systems, 
whereas R290 has only been used in commercial applications and not 
yet in automotive applications.  
? Though limited experimental research was done, no studies on the 
typical characteristics and strengths of secondary loop system (i.e. 
cold storage, multiple cooling loads, delayed pull down, etc.) were 
found in published literature.  
? For simplicity and reliability, aqueous mixtures of water and ethylene 
glycol work well as secondary refrigerants in automotive SL systems. 
The high specific heat and high heat transfer coefficient of water favor 
mixtures with high water content, such as a 68/32% water ethylene 
glycol mixture. 
38 
2 Research Objectives 
The overarching goal of this thesis is to study the performance of 
secondary loop systems, in particular thermal storage potential and benefits in 
terms of transient performance. Based on the reviewed literature, as well as 
developments in state-of-the-art automotive air-conditioning units, the following 
research objectives have been identified: 
1. Characterize performance of secondary loop systems using low-
 Global Warming Potential (GWP) refrigerants 
o Build a test facility which enables steady-state and transient 
performance measurements with high accuracy for direct 
expansion, as well as secondary loop systems 
o Quantify performance characteristics of secondary loop 
systems, such as cooling capacity, coefficient of 
performance (COP), energy consumption and pull down time 
o Investigate the performance of environmentally friendly low-
 GWP refrigerants, such as R152a and R290 in secondary 
loop systems 
2. Quantify potential benefits of thermal storage, as well as measures 
to reduce A/C energy consumption 
o Quantify the effect of thermal storage on thermal comfort 
and power consumption during off-cycle operation 
o Investigate the role of phase change thermal storage to 
reduce energy consumption of  secondary loop systems 
o Evaluate alternative measures of reducing energy 
consumption, such as compressor and pump controls, as 
well as cabin pre-conditioning 
3. Model and simulate a secondary loop system for steady-state, as 
well as transient operation with Modelica 
o Model a passenger cabin, a direct expansion system and a 
secondary loop system 
o Validate models based on experimental data 
39 
o Simulate effects of coolant volume and composition on 
transient system performance 
 
A test facility was built to accommodate direct expansion and secondary 
loop systems. The test facility, the sensing equipment, and modifications made to 
the facility for different tests are described in Chapter ?3. Data acquisition and 
controls of the facility, as well as the cabin model needed for transient testing, 
are discussed in Chapters ?4 and ?5. 
Experimental results to quantify the performance of secondary loop 
systems in relation to direct expansion systems, as well the test procedures used 
for steady-state and transient performance testing are discussed in Chapter ?7. 
The performance of alternative refrigerants, such as R152a and R290 in 
the secondary loop system is discussed in Chapter ?8. Steady-state, as well as 
transient experimental results are provided with respect to the R134a direct 
expansion system.  
Thermal storage of secondary loop systems during off-cycle period, as 
well as the development of suitable test procedures is discussed in Chapter ?9, as 
well as an evaluation of ice storage with regards to thermal comfort and energy 
saving in electric vehicles. 
Chapter ?10 introduces possible control strategy to reduce energy 
consumption during the operation of the air-conditioning unit in electric vehicles. 
The benefits of different control strategies and cabin pre-conditioning are 
investigated and discussed. 
40 
Models of a direct expansion system, a secondary loop system, and an 
automotive passenger cabin were built in the Modelica language. Details of the 
models and their validation with experimental data are shown in Chapter ?11. 
Simulations were carried out to study the effect of coolant volume and coolant 
composition on transient system performance.  
   
41 
3 Experimental Test Facility 
Chapter ?3 is divided into three parts, each introducing a variation of the 
experimental facility, which was build to perform tests on mobile air-conditioning 
systems (MACs). While chapter ?3.1 acquaints the reader with the basic direct 
expansion (DX) system, also referred to as the baseline system, Chapters ?3.2 
and ?3.3 introduce variations which enable experimental research on secondary 
loop systems. Specifically, the secondary loop (2LP) test facility, and the 
combined secondary loop ice storage (IS) facility. Each facility allows for the use 
of different refrigerants in the vapor compression cycle. 
 
3.1 Direct Expansion Test Facility 
Figure ?3.1 shows the direct expansion (DX) baseline facility. The test rig 
features a basic vapor compression cycle (VCC), including a compressor (1), a 
condenser (2), an electronic expansion valve (3) and an evaporator (4). The 
automotive compressor is belt driven by an electric motor (5), similar to the 
operation in cars using internal combustion engines (ICEs). The motor is driven 
by a variable frequency drive (VFD), enabling a variable speed control of the 
compressor. A torque meter in between the electric motor and the compressor, 
combined with an rpm meter in front of the compressor clutch, allows for 
measurement of compressor power consumption. A receiver (6) was installed 
after the condenser for a stable refrigerant charge management during changing 
operating conditions. A mass flow rate meter, placed at the lowest point of the 
42 
test facility after the receiver, measured mass flow rate of subcooled liquid 
refrigerant to determine refrigerant-side cooling capacity.  
For the purpose of automotive air-conditioning testing, the test facility 
allowed for two independent climatic environments, namely the ambient and the 
passenger cabin. This was accomplished by housing the condenser windtunnel 
(9) in a climatic chamber and installing the evaporator in a psychrometric loop 
featuring an air handling unit (AHU) (8). The condenser windtunnel, as well as 
the evaporator psychrometric loop served the purpose of air flow rate and air-
 side capacity measurements at the respective heat exchangers. In addition to 
nozzles (7), needed for air flow rate measurement, both the windtunnel and the 
psychrometric loop featured thermocouple mesh grids, relative humidity 
measurements, flow turbulators, and flow straighteners in accordance with 
AHSRAE standard [105]. 
Figure ?3.1 contains a legend, which details system components and 
sensing equipment used in the direct expansion baseline facility. 
43 
 
Figure ?3.1: Direct Expansion (DX) Test Facility Schematic 
 
Component Symbols Component/Sensor 
1 Compressor 
2 Condenser 
3 Expansion Valve (electronic) 
4 Evaporator 
5 Electric Motor 
6 Receiver 
7 Air Flow Nozzle 
8 Air Handling Unit 
9 Fan 
Sensor Symbols  
m Mass Flow Rate Meter 
M Torque Meter 
P Pressure Transducer 
?P Differential Pressure Transducer 
r RPM Meter 
T Thermocouple 
? Relative Humidity Sensor 
 
 
T
 M
 m
 r
 ?
 ?PAHU
 ?P
 T
 T
 T
 T T T
 T
 T
 TT
 PPP
 PP
 P?
 ?
 ?
 1 2
 34
 5
 6
 7
 8
 9
 7
44 
3.2 Secondary Loop Test Facility 
Figure ?3.2 shows a schematic of the secondary loop (2LP) experimental 
test facility. Compared to the baseline direct expansion facility, several 
modifications were made to the facility. The evaporator of the VCC was replaced 
with an intermediate plate type heat exchanger (4). Instead of the evaporating 
refrigerant cooling down the air of the passenger cabin directly, the intermediate 
heat exchanger cools down a secondary working fluid.  
About 5.5 kg of a water-ethylene glycol mixture were used in the current 
research, prepared to 32 wt% ethylene glycol and 68 wt% water. The use of a 
water glycol mixtures is common in automobiles and was chosen due to its 
balance between frost prevention, high specific heat capacity, cost effectiveness, 
and safety. The mass fraction of water and glycol was chosen to prevent pipe 
bursts due to freezing at low ambient temperature while maximizing water 
content in the mixture, due to the superior heat transfer characteristic of pure 
water. Furthermore, this specific mass fraction was chosen to allow for 
comparison with previous research by Ghodbane et al. [104].  
A variable speed pump (8) was used to circulate the chilled coolant from 
the intermediate heat exchanger through a liquid to air heat exchanger, called the 
cooler (5). In the experimental test facility, the cooler cools down the air in the 
psychrometric loop. In an automobile, the cooler would be located underneath 
the dashboard to provide cool air to the passenger cabin. A reservoir (9) was 
used as pressure expansion tank. Temperatures and coolant mass flow rate 
45 
were measured to determine capacities of the intermediate heat exchanger and 
the cooler in the secondary loop. 
In automobiles, this configuration allows for the entire vapor compression 
cycle to be placed underneath the hood. This results in a significant reduction of 
refrigerant charge, a reduction of refrigerant-side pressure drop, and an increase 
in safety when using flammable or high pressure refrigerants. On the other hand, 
more components (heat exchangers, pump, coolant) result in an increase in 
mass and consequently fuel consumption, A/C power consumption, as well as a 
thermal delay in passenger cabin cool down. Some of these disadvantages may 
be offset by a smart choice of refrigerants and thermal storage controls to 
balance the need for capacity to deliver passenger comfort and the goal of 
achieving a reduction in long term A/C power consumption. 
46 
 
Figure ?3.2: Secondary Loop (2LP) Test Facility Schematic 
 
Component Symbols Component/Sensor 
1 Compressor 
2 Condenser 
3 Expansion Valve (electronic) 
4 Intermediate Heat Exchanger 
5 Cooler 
6 Electric Motor 
7 Receiver 
8 Coolant Pump 
9 Coolant Reservoir 
10 Air flow Nozzle 
11 Air Handling Unit 
12 Fan 
Sensor Symbols  
m Mass Flow Rate Meter 
M Torque Meter 
P Pressure Transducer 
?P Differential Pressure Transducer 
r RPM Meter 
T Thermocouple 
? Relative Humidity Sensor 
 
 
T
 M
 m
 r
 ?
 ?PAHU
 ?P
 T
 T
 T T T
 T
 T
 TT
 PPP
 PP
 ?
 ?
 ?
 1 2
 3
 4
 6
 7
 10
 11
 11
 8
 T T
 T T
 10
 9
 5
 m
47 
3.3 Secondary Loop combined with Ice Storage Test Facility 
Figure ?3.3 shows the test facility schematic for a secondary loop test 
facility including ice storage (IS). The figure omits the primary loop, i.e. the VCC, 
since the refrigerant-side was not modified. The main change to the secondary 
loop facility, introduced in Chapter ?3.2, is the addition of an ice storage box (3) 
and the integration of three sets of ball valves (6) to bypass each heat exchanger. 
The ice storage heat exchanger, shown in Figure ?3.4 was custom 
designed and built in the laboratory for the purpose of producing ice from 15 
Liters of water and using the ice to assist the VCC in cooling the passenger cabin. 
Based on transient performance data from the baseline and secondary loop 
system, the heat of fusion of 10 to 15 L of ice was found to provide enough 
energy for cooling the passenger cabin during a 20 minute commute without the 
need for a vapor compression system. However, this idealized approximation did 
not account for the challenge of distributing the heat fast and evenly throughout 
the ice so as to harvest the cooling capacity in a quick and homogenous fashion. 
During realistic operation, the ice would melt quickly around the tubes of the heat 
exchanger, giving rise to sensible heating of liquid water around the tubes. The 
estimated 10 to 15 L of ice would not suffice as a standalone source of cooling, 
but would rather be helpful in taking part of the load away from the VCC. To 
increase heat transfer, fins were added to the ice storage heat exchanger, 
though with a significantly larger fin spacing compared to air-side fins.  
48 
 
Figure ?3.3: Ice Storage (IS) Test Facility Schematic 
 
Component Symbols Component/Sensor 
1 Intermediate Heat exchanger 
2 Coolant Pump 
3 Ice Storage Heat Exchanger 
4 Cooler 
5 Coolant Reservoir 
6 Ball Valve 
7 Air Flow Nozzle 
8 Air Handling Unit 
Sensor Symbols  
m Mass Flow Rate Meter 
?P Differential Pressure Transducer 
T Thermocouple 
? Relative Humidity Sensor 
 
 
?PAHU
 T
 T
 ?
 ?
 1
 4
 T T
 T T
 3
 5
 2
 7
 6
 8
 m
49 
 
Figure ?3.4: Ice Storage Heat Exchanger Schematic 
 
The ice storage heat exchanger, made solely of copper, was 
manufactured in the laboratory according to the following design goals: 
? to allow for manual construction using soldering connections 
? to create a large surface area for heat transfer, while remaining 
structurally stable 
? to fit into and sufficiently use the volume of a 15 liter plastic box 
 
Figure ?3.5 provides information on the operational modes of the ice 
storage test facility. Bypasses, controlled by ball valves, allow for four operational 
modes of the facility. Mode (a) bypasses the ice storage heat exchanger, so as to 
use only the basic secondary loop system for cooling. Mode (b) is a charging 
50 
mode, in which the cooler is bypassed to allow the entire cooling capacity of the 
VCC to be used for ice production. Mode (c) enables the combination of VCC 
and ice storage to cool the passenger cabin. Mode (d) forces the ice storage box 
to cool the passenger cabin alone, by bypassing the chiller.  
 
Figure ?3.5: Ice Storage Facility Operation Modes 
 
51 
Specifications of the sensing equipment used throughout the three 
variations of the experimental test facility are provided in Table ?3.1. The type of 
sensing equipment, as well as manufacturer and model, are shown. The table 
further provides information on operating range and systematic uncertainty, 
based?on?manufacturer?s?data. 
Specifications of system components, such as compressor, expansion 
valves, and heat exchangers are provided in ?Appendix A.  
Table ?3.1: Sensing Equipment Specifications 
Sensor Manufacturer/Model 
Operating 
Range 
Systematic 
Uncertainty 
Refrigerant-side    
Thermocouple Omega / T-type -200 : 350 ?C ? 0.5 K 
Pressure Transducer 
Setra / 280E 0 : 250 psig ? 0.13% FS 
Setra / 280E 0 : 500 psig ? 0.13% FS 
Mass Flow Rate Meter MicroMotion  0 : 1,300 kg/h ? 0.05% AR 
Air-side    
Thermocouple-grid Omega / T-type -200 : 350?C ? 0.3 K 
Relative Humidity Sensor 
Vaisala / HMD 30YB 2 : 100% ? 2 % 
Vaisala / HMD 60Y 2 : 100% ? 2 % 
Differential Pressure Tr. Setra / 264 5.0" WC ? 1% FS 
Refrigerant-side    
Thermocouple Omega / T-type -200 : 350 ?C ? 0.5 K 
Mass Flow Rate Meter MicroMotion/ R025 0 : 2,720 kg/h ? 0.5% AR 
Compressor/Pump Power    
RPM Meter Shimpo / RS-220H 0 : 500 Hz ? 0.1% AR 
Torque Meter Himmelstein 28002T 56.5 Nm ? 0.1% FS 
Watt Meter Ohio Semitronics / PC5 0 : 750 W ? 0.5% FS 
52 
4 Data Acquisition and Analysis 
Chapter 4 introduces the data acquisition process, as well as the data 
reduction process and uncertainty analysis. 
 
4.1 Data Acquisition Process 
Figure ?4.1 shows the structure of the data acquisition process (DAQ). The 
sensing equipment (1), introduced in Chapter 3, sends output signals of 4-20 mA 
or 0-5 VDC to National Instrument Field Point modules (2) [106]. Field Point 
modules collect the signals and transfer them to a data acquisition personal 
computer (PC) (3). National Instrument's LabVIEW software reads in the signals 
for further processing. Several LabVIEW programs (4) were written, some of 
which are introduced in Chapter ?5. The programs allow for reading sensor 
signals, converting them to the unit of measure (i.e. pressure or mass flow rate) 
and employing data reduction techniques. In addition to the standard LabVIEW 
library, the PID Control Toolset, as well as a sub-program, XProps [107], were 
used. XProps is a software distributed by Optimized Thermal Systems and is 
based on NIST's REFPROP-8 [108] refrigerant property database. Furthermore, 
sub-programs for moist air and for water-glycol mixtures were used in the data 
reduction process. Part of the processed data was used to control  equipment in 
the test facility (5) (6), as well as visualize processes real time on the PC monitor 
(7). Controls can be distinguished into continuous controls, using analog output 
modules (used for EXV opening, compressor rpm, etc.), or the control of discrete 
53 
events, using digital output modules (used for compressor clutch, pump on/off, 
etc.). An array of all collected data and processed data was written to a Microsoft 
Excel spreadsheet (8). Custom Microsoft Excel analysis spreadsheets were used 
for further data analysis and visual formatting. 
 
Figure ?4.1: Data Acquisition Process Structure 
 
4.2 Data Reduction Process 
Data reduction was performed within the LabVIEW data acquisition 
program. Property subroutines are an essential part of converting measurables, 
such as temperature and pressure, into calculated parameters, such a density, 
specific heat capacity, enthalpy, and others. A refrigerant property subroutine, a 
moist air property subroutine, as well as a property subroutine for aqueous 
mixtures of glycol were used in LabVIEW to calculate fluid properties. The 
54 
refrigerant property subroutine is based on XProps [107]. The moist air property 
subroutine was developed by Hwang and Aute in the Center for Environmental 
Energy Engineering. The glycol property subroutine was written specifically for 
this research. Equations and coefficients used in the subroutine were borrowed 
from M. Conde Egineering [109]. A visual representation of all subroutines, 
showing the respective inputs and outputs, is provided in Figure ?4.2. 
Air flow rates across cooler and condenser were determined by pressure 
drop measurement across nozzles in the condenser windtunnel and the 
evaporator psychrometric loop. The calculation of air flow rates, and 
subsequently the calculation of air-side heat exchanger capacities, were based 
on the "Standard Methods for Laboratory Airflow Measurement", ANSI/ASHRAE 
41.2-1987 (RA92) [105]. 
Airs side mass flow rates were calculated based on air flow rate and local 
density using Equation (1).  
InNozzNozzleVm ,??? ??      (1) 
Air-side capacity was determined for the evaporator and cooler using 
mass flow rate and enthalpy difference, while condenser air-side capacity was 
determined using mass flow rate, specific heat capacity, and temperature 
difference, as shown in Equations (2) and (3), respectively. 
? ?hAirEvap,, ?? mQ AirEvap ??      (2) 
? ?TcmQ pAirCond ??? AirCond,, ??
      (3) 
55 
 
Figure ?4.2: Moist Air / Refrigerant / Glycol Property Sub Routines 
 
Evaporator or cooler air-side capacity can be further distinguished into 
sensible and latent capacity. Sensible capacity constitutes the portion of heat 
transfer which is used to decrease the temperature across the evaporator or 
cooler, while latent capacity denotes the portion used to dehumidify the air. A 
common parameter used in air-conditioning is the Sensible Heat Factor (SHF), 
which decribes the amount of total heat used for sensibly changing the 
56 
temperature of a given quantity of air. Sensible and latent capacity, as well as the 
sensible heating factor are shown in Equations (4), (5), and (6). 
                       (4) 
                                (5) 
    
          
       
     (6) 
Refrigerant-side capacities for evaporator, intermediate heat exchanger, 
and condenser were determined using refrigerant mass flow rate and enthalpy 
difference, as shown in Equation (7). 
? ?hRefRef ?? mQ ??      (7) 
Calculation of energy balance for refrigerant-to-air heat exchangers, as 
well as refrigerant-to-coolant and coolant-to-air heat exchangers, were 
determined to proof proper operation of the system and ensure consistency of 
measured data. Equation (8) gives an example of an energy balance calculation 
for the evaporator in the baseline DX system. 
? ?
 100
 Ref,
 ,Ref,
 Evap
 AirEvapEvap
 Evap
 Q
 QQ
 EB ?
 ?? ?
 ?
      (8) 
Compressor power was determined by torque and rpm measurements, as 
shown in equation (9). Pump electrical power was measured directly using a 
wattmeter.  
rpmTorquePComp ??
      (9) 
57 
Compressor isentropic and volumetric efficiencies were calculated using 
experimentally measured data. Isentropic efficiency was determined using 
Equation (10), where the numerator denotes ideal enthalpy difference achieved 
through isentropic compression, and the denominator describes the 
experimentally measured enthalpy difference. The volumetric efficiency was 
determined by Equation (11),?where?MFR?is?the?refrigerant?mass?flow?rate,??suction 
is the suction density, rpm is the compressor speed, and VD is the compressor 
displacement.  
            
                         
                           
     (10) 
            
   
            
   
  
     (11) 
System efficiency, determined by the coefficient of performance (COP), 
was  calculated by comparing refrigerating effect to the power consumed by the 
system. Equations (12) and (13) introduce the COP, where Q is cooling capacity, 
and P is power consumption. In case of a secondary loop system, two COPs can 
be distinguished. The primary COP, Equation (12), is the efficiency of the primary 
refrigeration cycle, where Q is refrigerant-side cooling capacity at the 
intermediate heat exchanger, and P is power consumption of the compressor. 
Total COP, Equation (13), was determined by air-side cooling capacity and 
combined power consumption of compressor and coolant pump. 
Comp
 Chiller
 prim
 P
 Q
 COP
 Ref,
 ?
 ?
      (12) 
58 
Total
 AirCooler
 Total
 P
 Q
 COP
 ,
 ?
 ?
      
(13) 
Up to this point, the discussed equations dealt with performance 
measurements during steady state operation. However, air-conditioning 
operation in an automobile is often highly transient, which is reflected by the 
majority of the experimental tests conducted in this research. To assess and 
compare transient performance, performance parameters, such as power 
consumption and cooling capacity, are integrated over the duration of the 
respective tests. Similarly to the coefficient of performance for steady state 
operation, a transient COP, the transient performance factor (TPF), was defined 
to compare system performance during transient operation. The TPF is defined 
in Equation (14). 
    
      
 
   
       
 
   
     (14) 
While ice storage is used in connection with the secondary loop system, 
two additional performance parameters can be introduced: cooling capacity of 
the ice storage heat exchanger, Equation (15), and ice storage charging COP, 
Equation (16). Cooling capacity of the ice storage heat exchanger can only be 
measured on the refrigerant-side and not on the ice-side, due to phase change. 
                                   (15) 
                
     
      
     (16) 
 
59 
4.3 Uncertainty Analysis 
Uncertainty analysis was performed on both, steady state, as well as 
transient data. The total uncertainty of a parameter can be distinguished into two 
contributors, the so called systematic uncertainty and random uncertainty. The 
systematic uncertainty is a bias error, which causes the mean of a measured 
parameter to be significantly different from its "true" value. Systematic uncertainty 
can be caused by imperfections in the measurement instrument, imperfections of 
the calibration of the instrument, and unexpected changes in the environmental 
conditions which cannot be accounted for. Calibration with a more precise 
instrument can be used to remove part of the systematic error of an instrument. 
Random uncertainty is based on the fact that repeated measurements of the 
same parameter using the same instruments under the same operating 
conditions will randomly and unpredictably differ in magnitude. Random 
uncertainty is closely related to the precision, i.e. the resolution of the 
instruments used for measurement. For a finite repeated set of measurements, 
the higher the precision of the instruments, the smaller the variance around the 
mean of the measurand. If a measurement under identical operating conditions 
and with identical instruments would be repeated indefinitely, the random 
uncertainty would be expected to be null, due to a statistically normal distribution. 
The total uncertainty of a parameter can subsequently be determined by linear 
summation of the systematic and random uncertainty. Figure ?4.3 strives to 
visualize the concepts of systematic uncertainty, random uncertainty, and  
calibration. 
60 
 
Figure ?4.3: Systematic and Random Uncertainty 
 
During the data reduction process, non-measurable fluid properties, such 
as enthalpy or specific heat, were calculated based on measured physical 
properties, such as temperature and pressure. These non-measurables were 
further used to determine system performance metrics, such as cooling capacity 
and COP. Random uncertainty of non-measurables and performance metrics can 
be computed as their respective standard deviation, using Equation (17).  
61 
                
 
 
       
        
           
   
 
 
      (17) 
For any calculated parameter, the systematic uncertainty needs to be 
propagated by means of Pythagorean Addition, shown in Equation (18) for a 
function f and coefficients xi from x1 through xn.  
                  
  
   
    
 
  
  
   
    
 
     
  
   
    
 
 
 
 
      (18) 
Equation (18) implies that the relationship of any calculated parameter, i.e. 
enthalpy, to its measurable components, i.e. temperature and pressure, be 
known. A flexible approach of finding the uncertainty of calculated parameters is 
the use software which allows for propagation of uncertainty of fluid properties, 
such as Engineering Equation Solver (EES).  
Table ?4.1 provides an example of systematic and random uncertainty for a 
steady state test of the secondary loop system (2LP), using Propane (R290) as 
primary refrigerant. The table shows systematic, random, and total uncertainty of 
performance parameters at idling compressor speed and at compressor speed 
related to highway driving. Uncertainties of three different cooling capacities are 
given: refrigerant-side cooling capacity at the intermediate heat exchanger, 
refrigerant-side cooling capacity at the cooler, and air-side cooling capacity at the 
cooler. Total uncertainty was lowest on the refrigerant-side, due to a low 
systematic uncertainty. High systematic uncertainty caused the air-side capacity 
to exhibit the highest total uncertainty. Random uncertainty was typically highest 
on the refrigerant-side, due to vibrations in the compressor belt drive and PID 
62 
control of the EXV. The thermal mass of the coolant allowed coolant and air-side 
measurements to have decreased fluctuations and consequently lower random 
uncertainty. In general, relative uncertainty of cooling capacity for idling speed 
tests was about twice as high compared to high way speed tests. This was due 
to reduced absolute magnitudes of measurands compared to fixed absolute 
uncertainties (e.g. decreased ?T across heat exchanger, while thermocouple 
uncertainty of 0.5 K stays constant). Uncertainty of power consumption was 
dominated by random uncertainty for the reasons mentioned above. Uncertainty 
of primary or refrigerant-side COP was about 5-6%, mostly dominated by random 
uncertainty. Uncertainty of total system COP on the refrigerant-side was about 6% 
for highway speed tests and about 10% for idling speed tests.  
Table ?4.1: Relative Systematic and Random Uncertainties (R290 2LP, T35RH40) 
Performance 
Parameter 
Idling Highway 
Systematic 
[%] 
Random 
[%] 
Total 
[%] 
Systematic 
[%] 
Random 
[%] 
Total 
[%] 
Capacity (refrigerant) 0.7 2.5 3.2 0.7 1.1 1.7 
Capacity (coolant) 4.6 1.3 5.9 2.2 0.5 2.8 
Capacity (air) 14.5 1.7 16.2 7.7 1.2 8.8 
Power Consumption 0.4 4.1 4.5 0.4 3.7 4.1 
COP (refrigerant) 0.9 4.8 5.7 0.8 3.9 4.6 
COP (coolant) 6.5 3.3 9.8 2.7 3.4 6.1 
 
Due to lower uncertainty compared to air-side measurements, refrigerant-
 side COP was used to describe steady-state system performance of the direct 
expansion system, while coolant-side COP was used to describe steady-state 
system performance of the secondary loop system. 
63 
Table ?4.2 provides a summary of total relative uncertainties of 
performance parameters during steady state operation. Uncertainties are given 
for different refrigerants (R134a, R152a, R290), different system layouts (DX, 
2LP), a range of ambient temperatures and relative humidities (15?C/80%, 
25?C/50%, 35?C/40%, 45?C/20%), and different compressor speeds (idle, 
highway). Across various refrigerants, system layouts, and operating conditions, 
several characteristics can be observed: Refrigerant-side capacity typically 
showed a relative uncertainty in the range of 3%, often less at highway speed 
and more at idle speed. For the secondary loop system, refrigerant-side capacity 
was in the range of 3% at highway speed and in the range of 6% at idle speed. 
Relative uncertainty of power consumption was on the order 3% at highway 
speed and on the order of 5-7% at idle speeds. In a few cases, moslty when 
idling at low ambient temperatures, relative uncertainty was significantly higher 
than above mentioned values. At these conditions the lack of cooling load, which 
made EXV PID control difficult, resulted in fluctuations in MFR, which increased 
random uncertainty of cooling capacity. 
64 
Table ?4.2: Summary of Steady-state Relative Uncertainty of Performance 
Parameters 
Refrig./System R134a DX 
Amb. Conditions T15 RH80 T25 RH50 T35 RH40 T45 RH20 
Speed idle hwy idle hwy idle hwy idle hwy 
Capacity (ref) [%] 1.2 - 1.3 1.3 1.4 1.6 1.5 1.5 
Capacity (col) [%] - - - - - - - - 
Capacity (air) [%] 12.8 - 15.5 8.5 19.1 10.0 20.0 13.6 
Power [%] 6.7 - 6.1 3.4 6.5 2.9 6.9 2.7 
COP (ref) [%] 7.0 - 6.4 3.6 6.8 3.3 7.4 3.2 
COP (col) [%] - - - - - - - - 
Refrig. /System R152a DX (drop-in) 
Amb. Conditions T15 RH80 T25 RH50 T35 RH40 T45 RH20 
Speed idle hwy idle hwy idle hwy idle hwy 
Capacity (ref) [%] 14.1  - 3.1 1.8 2.8 2.0 2.5 6.8 
Capacity (col) [%] - - - - - - - - 
Capacity (air) [%] 13.6  - 13.3 7.5 17.0 9.3 18.7 12.1 
Power [%] 42.2  - 6.6 3.7 6.3 3.2 7.0 3.0 
COP (ref) [%] 16.5  - 7.3 4.2 7.0 3.8 7.4 7.7 
COP (col) [%] - - - - - - - - 
Refrig./System R290 DX (drop-in) 
Amb. Conditions T15 RH80 T25 RH50 T35 RH40 T45 RH20 
Speed idle hwy idle hwy idle hwy idle hwy 
Capacity (ref) [%] - - 2.3 - 2.6 3.3 2.9 2.6 
Capacity (col) [%] - - - - - - - - 
Capacity (air) [%] - - 12.6 - 14.2 8.1 16.3 11.3 
Power [%] - - 4.6 - 4.8 2.3 5.0 2.0 
COP (ref) [%] - - 5.3 - 5.5 4.0 6.1 3.4 
COP (col) [%] - - - - - - - - 
Refrig. /System R152a 2LP (drop-in) 
Amb. Conditions T15 RH80 T25 RH50 T35 RH40 T45 RH20 
Speed idle hwy idle hwy idle hwy idle hwy 
Capacity (ref) [%] 4.4 - 2.2 2.7 4.3 1.5 4.3 2.7 
Capacity (col) [%] 7.0 - 6.5 3.5 5.4 2.9 5.1 2.8 
Capacity (air) [%] 11.4 - 14.0 7.8 17.0 9.6 18.9 12.4 
Power [%] 6.6 - 5.1 2.8 6.1 2.4 6.9 2.4 
COP (ref) [%] 7.8 - 5.5 4.0 7.5 2.8 8.2 3.7 
COP (col) [%] 13.1 - 10.9 5.2 10.2 4.6 9.9 4.4 
Refrig. /System R290 2LP (reduced speed) 
Amb. Conditions T15 RH80 T25 RH50 T35 RH40 T45 RH20 
Speed idle hwy idle hwy idle hwy idle hwy 
Capacity (ref) [%] 5.8 - 3.2 1.9 3.2 1.7 3.0 2.0 
Capacity (col) [%] 9.0 - 6.1 3.1 5.9 2.8 5.5 2.7 
Capacity (air) [%] 14.7 - 12.8 7.1 16.2 8.8 17.5 11.7 
Power [%] 4.7 - 4.7 4.1 4.5 4.1 4.1 4.5 
COP (ref) [%] 7.4 - 5.7 4.6 5.7 4.6 5.1 5.1 
COP (col) [%] 15.6 - 10.8 6.3 9.8 6.1 8.9 6.4 
 
65 
For transient measurements the principle of random uncertainty, by which 
a parameter is repeatedly measured under the same conditions, becomes 
irrelevant. since all parameters are deliberately changing in time. As a result, 
uncertainty of transient tests can only be evaluated by systematic uncertainty, 
though it's calculation is more involved. The absolute magnitude of a measurand 
can change significantly during the course of a test. Systematic uncertainty was  
evaluated at every timestep during the test and integrated over the time of the 
experiment to determine the total uncertainty of accumulated performance 
parameters, such as energy consumption or energy availability for cooling. 
Transient uncertainty analysis was performed using the uncertainty 
propagation table method in the Engineering Equation Solver software (EES). An 
example of transient uncertainty is given in Figure ?4.4, showing absolute 
uncertainty of performance parameters during a New European Drive Cycle test 
(NEDC) for the R152a secondary loop (2LP) system. The uncertainty of air-side 
capacity followed uncertainty of the temperature difference across the cooler, 
and was consequently decreasing as the test progressed. Uncertainty of 
refrigerant-side uncertainty followed fluctuations in MFR, which were determined 
by the compressor speed profile. During sharp transients superheat can be lost 
and refrigerant-side uncertainty can spike. Uncertainty of power consumption 
directly followed the NEDC compressor speed profile.   
 
66 
 
Figure ?4.4: Transient Uncertainty Example - New European Drive Cycle 
 
Similar to the analysis of transient performance data, discussed in 
Chapter ?4.2, transient uncertainty data can be integrated over time to find the 
accumulated uncertainty over the duration of a transient test. With accumulated 
uncertainties of air-side capacity (19) and power consumption (20), uncertainty of 
the transient performance factor (TPF) can be determined, based on 
Pythagorean summation (21). 
                              
 
   
     (19)  
                                  
 
   
     (20) 
       
 
          
           
 
  
        
          
              
 
     (21) 
67 
 Uncertainty of steady state experimental results, as well as accumulated 
uncertainty of transient experimental results, are provided by means of error bars 
in the charts shown in Chapters ?7, ?8, and ?9.  
  
68 
5 Automation and Control 
Chapter 5 provides information about automation and control of the data 
acquisition system and test facility. The chapter discusses the passenger cabin 
model, which allows transient testing of MAC components, as well as the controls 
implemented for drive cycle testing, and the controls implemented for research 
on energy saving strategies through alternative control of the air-conditioning 
system. 
 
5.1 Structure of the Data Acquisition Control 
Figure ?5.1 shows the structure of the Main LabVIEW Data Acquisition 
program (main DAQ). The main DAQ consisted of six modules, which followed 
separate tasks and can be scheduled independently.  
The first module scheduled the data reduction process, the control of the 
cabin model, the visualization of all data in real time on screen, and the recording 
of data to an Excel spreadsheet. The data reduction process was introduced in 
Chapter ?4.2, while the Cabin Model and associated controls will be introduced in 
Chapter ?5.2. 
The second module was used to manually control the temperature, 
relative humidity, and air flow rate in the condenser wind tunnel and evaporator 
psychrometric loop, respectively.  
69 
The third module scheduled the drive cycle controls. Various components 
and parameters in the system can be controlled by predefined drive cycles, as 
opposed to constant control or PID control. Details of the drive cycle control will 
be introduced in Chapter ?5.3 
Modules 4 through 6 controlled components in the system directly. While 
module 4 controled the compressor clutch and compressor speed, module 5 
controled the pump speed and it's actuation, and module 6 controled the opening 
of the EXVs, based on evaporator superheat. Component control modules will be 
introduced in Chapter ?5.4. 
 
Figure ?5.1: Main Data Acquisition Structure and Control in LabVIEW 
 
5.2 The Cabin Model 
The Cabin Model is a lumped model of the automotive passenger cabin to 
simulate thermal and psychrometric processes. The main function is real-time 
communication with the controller of the air-handling unit of the psychrometric 
70 
loop to control the state of the air upstream of the evaporator (DX) or cooler 
(2LP). 
The core of the model consists of a set of energy and mass balance 
equations, which are solved by the Gauss-Jordan Method. This method was first 
developed by Gado [12]. For details on the equations used in the model, the 
reader is referred to Gado [12]. Since the core of the model did not change, the 
current model was successfully validated against data from Huang [15], using the 
same procedures outlined by Gado. 
The Cabin Model enables transient testing by a process described in 
Figure ?5.2. At every timestep, temperature, relative humidity, and air flow rate 
downstream of the evaporator (DX) or cooler (2LP) were measured. Based on 
physical and thermal characteristics of the cabin, as well as thermal loads, the 
cabin model determined temperature of the cabin air, temperature of interior 
mass, and relative humidity inside the cabin. The cabin air was then allowed to 
mix with ambient air, based on a pre-defined recirculation ratio, which lead to the 
state of the return air. The return temperature and relative humidity determined 
by the cabin model were subsequently used as set points for the PID control of 
the air handling unit (AHU), which controlled the condition of the air upstream of 
the evaporator (DX) or cooler (2LP). After the air was cooled down by the 
evaporator/cooler, the process was reiterated at the next timestep. All thermal 
and physical characteristics of the passenger cabin can be set by the user in the 
LabVIEW graphical user interface, which makes the Cabin Model a versatile 
instrument for testing MAC components in various automotive scenarios.  
71 
 
Figure ?5.2: Cabin Model Operational Schematic 
 
A flow chart of the Cabin Model is shown in ?Appendix B. Several control 
strategies used in automotive air-conditioning were implemented to increase 
versatility of the Cabin Model for simulation of subjective evaluation drives.  
 An evaporator fan speed control was implemented to decrease air flow 
rate across the evaporator/cooler as cabin temperature approaches comfort 
temperature, set by the user. A five step control was used to vary face velocity 
from 2.47 m/s for hot soaked conditions to 1.83 m/s for cabin temperatures within 
2?C of comfort setting. 
A thermostat control, using a virtual heater core, was added to be able to 
keep the cabin at comfort temperature once the air-conditioning system pulled 
the cabin temperature below comfort setpoint. In an actual automotive air-
 conditioning unit, the total air stream would pass through the evaporator. After 
Energy / Mass Balance
72 
being cooled down and dehumidified, part of the air stream would be routed 
through a heater core for reheating in order to control for comfort temperature. 
The heater core utilizes hot engine coolant to reheat the air, while the supply 
temperature to the cabin is controlled by a mixing flap after the evaporator, which 
controls the fraction of flow rate passing through the heater core. As the current 
test facility lacked the ability to house a heater core and mixing chamber, a virtual 
heater core was added to the Cabin Model to enable thermostat functionality. 
The virtual heater core used a software side PID controller which added a certain 
magnitude to the actual, measured supply temperature downstream of the 
evaporator/cooler. The cabin model energy and mass balance equations 
therefore received an artificially higher supply temperature signal. This resulted in 
an increase of the calculated return temperature in the model, and therefore the 
upstream evaporator air temperature, controlled by the model. Therefore, within 
one timestep, the virtual heater core would lead to a real increase in process 
temperature, enabling the facility to work as if an actual heater core were present. 
Heater core load was subsequently determined by using the virtual temperature 
difference across the heater. 
Drive cycle controls, described in Chapter ?5.3, were added to the Cabin 
Model. Instead of using a fixed ambient temperature, relative humidity, and solar 
insolation, the user could choose to use pre-determined drive cycles as input. 
This allowed for transient changes in ambient temperature, humidity and 
insolation, which is a key feature of a new drive cycle test, developed to 
showcase the versatility of the test facility. The development of drive cycles, as 
73 
well as system operation during a long term evaluation drive, the I-95 test, is 
explained in detail in ?Appendix C. 
 
5.3 Drive Cycle Control 
Drive cycle modules were developed to control components of the test 
facility in a predefined way, using software side PID controls. The New European 
Drive Cycle (NEDC) is one example of a predefined sequence of vehicle speed 
over time. The cycle consists of several UN/ECE Elementary Urban Cycles and 
one UN/ECE Extra-Urban Driving Cycle, as provided by the USA Environmental 
Protection Agency[110]. This predefined sequence of vehicle speed was used to 
determine a compressor speed profile and condenser fan frequency 
profile. ?Appendix C gives a more detailed overview about the conversion process. 
Inputs to a drive cycle module are: 
? Process value  
? Set point array  
? Time step 
? Cycle length  
The set point array is a text file, which holds an array of time vs. set points. 
The cycle length could be manually set by the user, to abort drive cycle operation 
after a set amount of time. The set point fed to the PID control is an adjusted set 
point, as a feed forward factor was implemented in all drive cycle modules for a 
tighter control of the process value. The feed forward factor can be tuned by the 
user, in a similar manner as tuning PID gains. For a feed forward factor of zero, 
74 
the set point of the PID control is the original set point, while for a feed forward 
factor of one, the adjusted set point will be the set point for the next time step in 
the set point array. This allowed the PID control to "know" a set point ahead of 
time and therefore helped to smooth out over- and undershooting.  
To simplify the task of reading in drive cycle profiles into LabVIEW, drive 
cycle modules included a reader which read out setpoint and time information 
from an Excel spreadsheet. For long drive cycles specifying each setpoint and 
time individually can get tedious, as drive cycles were typically controlled to a 0.5 
second timestep. Instead, only the corner points of a profile were specified in the 
Excel spreadsheet, as seen in Figure ?5.3, and the drive cycle reader interpolated 
between these points. Based on the desired time step, the interpolation between 
the corner points of the profile was fine or coarse. 
 
Figure ?5.3: NEDC Cycle Input to LabVIEW Cycle Reader 
 
Drive cycles modules could be turned on and off individually. The following 
parameters could be controlled by drive cycle modules: 
0
 500
 1000
 1500
 2000
 2500
 3000
 3500
 4000
 0 100 200 300 400 500 600 700 800 900 1000 1100 1200
 Com
 pres
 sor 
Spee
 d [rp
 m]
 Time [s]
 Labview NEDC Cycle -Points
75 
? Compressor speed 
? Compressor clutch actuation 
? Condenser fan speed 
? Ambient temperature 
? Ambient relative humidity 
? Solar insolation 
For NEDC tests, only compressor rpm and condenser fan frequency 
cycles are necessary. For off-cycle Start/Stop and SS3xI tests a clutch actuation 
profile has to be added. A rather complex drive cycle was developed to simulate 
a car driving down the I-95 corridor from Maine to Florida on the United State 
east coast. The test facility was able to simulate changes in climate, day and 
night cycles, driving patterns, and rest stops, using fully automated drive cycle 
control. Towards this purpose, drive cycles for ambient temperature, relative 
humidity, and solar insolation were added, based on TMY2 data [111]. ?Appendix 
C provides more detail about the I-95 test and associated test results.  
During shakedown tests, the NEDC compressor rpm profile was used in 
the compressor rpm drive cycle module to verify the fidelity of the setpoint 
following of drive cycle modules. Figure ?5.4 shows compressor rpm, as well as 
condenser fan speed, and the discrepancy between set points (predetermined  
profile) and process value. The compressor rpm followed drive cycle set points 
with high fidelity. Deviations from the drive cycle profile occurred for a short time 
during fast acceleration and deceleration due to inertia of the compressor and the 
motor, as well as due to pressure balance across the compressor. The short 
76 
deviations had no significant effect on power consumption measurements. The 
condenser fan control showed a high fidelity, even during sharp transients.  
 
Figure ?5.4: NEDC Cycle PID Control Verification 
 
5.4 Component Control 
Two electronic expansion valves (EXVs) were used for superheat control 
at the evaporator (DX) or the intermediate heat exchanger (2LP). The two valves 
were installed in parallel and fitted with separate controls, which allowed for a 
-10
 -8
 -6
 -4
 -2
 0
 2
 4
 6
 8
 10
 10
 20
 30
 40
 50
 60
 70
 80
 90
 100
 110
 0 200 400 600 800 1000 1200
 Er
 ro
 r [
 H
 z]
 C
 om
 pr
 es
 so
 r F
 re
 qu
 . [
 H
 z]
 Time [s]
 Compressor RPM Cycle
 Set Point
 Process
 Error
 -10
 -8
 -6
 -4
 -2
 0
 2
 4
 6
 8
 10
 0
 10
 20
 30
 40
 50
 60
 70
 80
 90
 100
 0 200 400 600 800 1000 1200
 Er
 ro
 r [
 H
 z]
 C
 on
 d 
Fa
 n 
Fr
 eq
 u.
  [H
 z]
 Time [s]
 Condenser Fan Cycle
 Set Point
 Process
 Error
 -10
 -8
 -6
 -4
 -2
 0
 2
 4
 6
 8
 10
 10
 20
 30
 40
 50
 60
 70
 80
 90
 100
 110
 0 200 400 600 800 1000 1200
 Er
 ro
 r [
 H
 z]
 C
 om
 pr
 es
 so
 r F
 re
 qu
 . [
 H
 z]
 Time [s]
 Compressor RPM Cycle
 Set Point
 Process
 Error
 -10
 -8
 -6
 -4
 -2
 0
 2
 4
 6
 8
 10
 0
 10
 20
 30
 40
 50
 60
 70
 80
 90
 100
 0 200 400 600 800 1000 1200
 Er
 ro
 r [
 H
 z]
 C
 on
 d 
Fa
 n 
Fr
 eq
 u.
  [H
 z]
 Time [s]
 Condenser Fan Cycle
 Set Point
 Process
 Error
77 
continuous control of opening area of the smallest opening of one valve to the 
largest opening of both valves combined. This enabled the test facility to 
accommodate the different needs of various refrigerants and a wide range of 
operating conditions. ?Appendix B provides more detailed information about the 
EXV control, which was developed for this research. 
The coolant pump was controlled in either manual or automatic control 
mode. In manual mode, pump speed could be set by the user, while automatic 
control employed a software side PID control. Automatic control was used mostly 
when controlling cabin temperature during comfort control tests utilizing ice 
storage. More detailed information on coolant pump control, as well as a control 
flow chart, is provided in ?Appendix B. 
The control of compressor speed and clutch actuation was the most 
versatile, since each test needed a different control strategy. The different modes 
of compressor control are: 
? Manual 
? Pump - temperature control 
? Drive cycle 
? Pull down 
? Comfort (Frost cycling) 
? Comfort (Relative Humidity cycling) 
? Comfort (Relative Humidity continuous) 
A detailed summary of the compressor controls developed for this 
research, as well as their respective flow charts, is provided in ?Appendix B. 
78 
6 Refrigerant Comparison 
One of the objectives of the present research was the evaluation of 
performance of alternative refrigerants, in both a direct expansion system, as well 
as in a secondary loop system. Before conclusions can be drawn from 
experimental results, the thermophysical characteristics of the refrigerants have 
to be compared to establish a foundation for their evaluation. Two alternative 
refrigerants were compared with the properties of R134a as reference. HFC-
 134a has been used in MACs since the early 1990's as a replacement for the 
more environmentally harmful CFC-12. While it has been used for the past two 
decades, recent regulations (MAC directive) resulted in its ban in all new 
automotive units in the European Union, starting from 2011. As a possible 
alternative, HFC-152a was introduced by Ghodbane et al. in the late 1990's [39]. 
To mitigate the flammability of R152a, the use of a secondary loop system was 
proposed by Ghodbane et al. [112]. Due to a lack of incentives for the automotive 
industry, the use of R152a in secondary loop systems has not been investigated 
further until recently. Hydrocarbons, such as R290, R600a, or mixtures of the two, 
are yet another environmentally friendly alterative. R290 has been used by car 
owners in Australia, as well as USA, within the past decades for direct expansion 
system retrofits, which is illegal in some states. Both R152a, as well as R290, 
were investigated in this research for use in secondary loop systems MACs.  
A comparison of refrigerant properties is shown in Table ?6.1. Both, R152a, 
as well as R290 (Propane) have an Ozone Depletion Potential (ODP) of zero. 
79 
The Global Warming Potential (GWP) of R152a GWPR152a=140 is significantly 
lower than R134a GWPR134a=1430, and is within the limits set forth for new 
refrigerants by the European Union (GWP < 150). R290 has an even lower GWP 
GWPR290=11, making it the most environmentally friendly choice amongst the 
three refrigerants used in this research.  
Table ?6.1: Refrigerant Property Table 
Parameter R134a R152a R290 
ODP 0 0 0 
GWP 1400 140 11 
NBP -26.11 -24.0 -42.11 
Tcr 101.1 113.3 96.68 
Pcr 4059 4520 4247 
Cp(0?C) 0.894 1.147 1.779 
L (0?C) 198.7 306.6 374.5 
? vapor (0?C) 14.42 8.381 10.34 
? liquid (50?C) 1102 830.5 448.6 
? (0?C) 1.09x10-5 9.045x10-6 7.79x10-6 
Molar Mass 102.03 66.051 44.096 
 
For use in MACs, the normal boiling point (NBP) of a refrigerant is a good 
indicator for its suitability of operating pressure. NBP, The normal boiling point of 
R134a is NBPR134a=-26.11?C. The normal boiling point of R152a is very similar 
NBPR152a=-24.0?C, indicating that R152a will work at similar, though somewhat 
lower pressures as R134a. R290 has a significantly lower NBP of  
NBPR290=-42.11?C. It is therefore expected that R290 operates at higher 
pressures compared to R134a. During drop-in tests in the direct expansion 
system, R290 operated at about 43% increased pressure at the evaporator, 
compared to R134a. 
80 
Specific heat of saturated vapor at 0?C of R152a is about 28% increased 
compared to R134a. Specific heat capacity of R290 at the same conditions is 
about 99% increased compared to R134a. Both, R152a, as well as R290, will 
have lower superheat, as well as lower subcooling, all other things considered 
equal. However, since the test facility uses EXVs, superheat stayed the same, 
but EXV opening was about 20% to 35% increased for R152a, and about 45% to 
60% increased for R290.  
A density comparison shows that R290 has a saturated liquid density 
which is about 60% reduced from the saturated liquid density of R134a. The 
liquid density of R152a is about 24% reduced compared to R134a at a 
temperature close to 50?C (liquid line). As a result, the refrigerant charges in the 
BSL facility of both, R152a and R290, were decreased by 20% and 56%, 
respectively. Saturated vapor density at 0?C is reduced 28% and 42% for R290 
and R152a, respectively. When operating a compressor with same displacement 
volume and same rpm, MFR for these refrigerants is expected to decrease 
accordingly, sans minor deviations based on changes in volumetric efficiency. 
Experimental results for the DX facility show a decrease in MFR of -29% for 
R290 and -41% for R152a, which supports the theoretical values. 
A comparison of viscosity shows that liquid line viscosity can be reduced 
by 10% to 15% for R152a and as much as 50% for R290. For saturated vapor, 
R152a viscosity can be reduced by 10% to 15% in comparison to R134a, while 
R290 viscosity can be reduced by 25%to 30%. A smaller frictional pressure drop 
can be expected throughout the system for both R152a and R290 when 
81 
compared to R134a. These magnitudes could not be verified experimentally, as 
the pressure drop in the vapor and liquid lines were mostly within measurement 
uncertainty.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    
Figure ?6.1 shows the P-h Diagram for R134a, R152a and R290. For the 
temperature range of interest (~ -10?C through +15?C), the latent heat of R152a 
is increased by roughly 55% compared to R134a, whereas the latent heat of 
R290 is increased by roughly 88%. Figure ?6.1 also shows a comparison of 
isentropic lines (*) in the upper left corner of the figure. R290 isentropic lines 
show the smallest slope, possibly leading to an increase of discharge 
temperature and isentropic losses. 
 
Figure ?6.1: Refrigerant P-h Diagram 
* 
82 
Figure ?6.2 shows the T-s Diagram for R134a, R152a, and R290. The T-s 
Diagram can be very useful to compare expected refrigerant performance in 
terms of cooling capacity and efficiency. Didion [113] discussed the importance 
of the vapor dome shape in the T-s Diagram. An important characteristic location 
is the critical point. When comparing refrigerants, a high critical pressure typically 
indicates a decrease in cooling capacity due to a lower suction density, resulting 
in a decreased MFR and subsequently volumetric capacity. The theoretical 
volumetric capacity for R152a is decreased by 10%, while the volumetric 
capacity of R290 is increased 35% compared to R134a. Similarly, the normal 
boiling point (NBP) is related to the critical temperature and a good indicator for 
working pressure. Refrigerant property data, shown in Table ?6.1 supports these 
indications. On the other hand, McLinden and Didion [114] point out that high 
critical temperatures often indicate increased COP. Since the condensation 
process is further removed from the critical point, excessive compressor 
superheat, as well as flash gas losses, can be avoided. While the specific heat 
capacity affects the shape of the vapor dome, it often has less of an effect on 
performance than NBP. However, since drop-in refrigerants, or refrigerants for 
the same application, are likely to be chosen to have similar NBPs, specific heat 
capacity can be a first indicator of cycle performance. Vapor-liquid lines featuring 
a smaller inclination angle are usually an indicator for better cycle performance. 
The slopes of vapor liquid lines are compared in the upper right corner of 
Figure ?6.2 (**) for R134a, R152a, and R290. 
83 
An alternative way to compare cycle performance of different refrigerants 
was introduced by Alefeld [115]. Alefeld devised a method to determine 
theoretical cycle efficiency using the second law of thermodynamics. A more 
accurate statement of efficiency??2, based on enthalpy differences, as well as a 
highly simplified statement, ?6, were introduced, as shown in Equations (22) and 
(23)  
     
     
     
     (22) 
         
   
  
)     (23) 
where gc is? compressor? isentropic? efficiency,? ?r is carnot efficiency, c is 
specific heat capacity, and r is latent heat. The simplified statement has the 
advantage that in fact only very little information about the refrigerant is needed 
and prior experimental testing to determine cycle efficiency can be omitted.  
Various assumptions and simplifications were introduced on the way from ?2 to 
arrive at ?6. Values for both are shown in Figure ?6.3 for R152a and R290 with 
respect to R134a as function of evaporating temperature. It can be concluded 
that, assuming an isentropic efficiency of gc=0.75, R152a will result in a slightly 
increased COP, while the use of R290 will result in a slightly decreased COP. In 
both cases, the comparison of refrigerant efficiency based on Alefeld's method 
shows no significant variations beyond ? 5%. 
84 
 
Figure ?6.2: Refrigerant T-s Diagram 
 
 
Figure ?6.3: Alefeld Refrigerant Efficiency with Respect to R134a 
** 
85 
7 Experimental Results: Secondary Loop Versus Direct 
Expansion 
Chapter ?7 presents experimental results, which provide a foundation to 
compare automotive secondary loop and direct expansion systems. Experimental 
results were obtained from steady state, as well as transient tests. Table ?7.1 
shows a test matrix, including compressor speeds and ambient conditions, as 
well as air-side face velocities at the heat exchangers. Experimental tests on the 
direct expansion system were performed using R134a and R152a as refrigerants, 
while tests on the secondary loop system were performed using R152a. Test 
conditions were chosen based on typical conditions encountered in the industry. 
Table ?7.1:Test Matrix - Secondary Loop Versus Direct Expansion System 
System/ 
Refrigerant 
Test 
Compr. 
Speed 
Ambient 
Temp/RH 
Evap. 
Face Vel. 
Cond. 
Face Vel. 
R134a DX 
/R152a DX 
/R152a 2LP  
Steady-state 
Performance 
idle 15?C/80% 
2.3 m/s 1.5 m/s 
idle 25?C/50% 
idle 35?C/40% 
idle 45?C/20% 
highway 15?C/80% 
2.3 m/s 3.0 m/s 
highway 25?C/50% 
highway 35?C/40% 
highway 45?C/20% 
NEDC 
NEDC 
profile 
30?C/50% 2.3 m/s 
NEDC 
profile 
Pulldown 
Performance 
idle 30?C/50% 
2.3 m/s 1.5 m/s 
idle 41?C/32% 
highway 30?C/50% 
2.3 m/s 3.0 m/s 
highway 41?C/32% 
 
86 
The variable speed coolant pump used in the experimental facility is a low 
efficiency gear pump. The pump, along with the three phase motor, was 
oversized to allow for a wide range of mass flow rates and pressure drop 
conditions. Electrical pump power consumption measured with a wattmeter was 
in the range of 330 W. A typical secondary loop pump, recently developed for 
research and pre-development of a new kind of automotive thermal management 
systems, shows a peak power consumption of 90 W and a rated power of 30 - 80 
W at mass flow rates similar to the present research for a 50wt% water-ethylene 
glycol solution. This value was confirmed through [116], as well as personal oral 
communication with a person working on the development of such pumps. Due 
to the pump in the present research being oversized for an actual automotive 
system, system COP suffered, especially at idling conditions where compressor 
power consumption was reduced and pump power consumption had a larger 
impact on total power. Therefore, experimental power consumption and COP 
results for the secondary loop system in the present research were adjusted to 
70 W coolant pump power consumption instead of the experimentally measured 
330 W (average) throughout all steady-state and transient performance results. 
 
7.1 Charge Optimization - Secondary Loop Versus Direct Expansion 
Charge optimization for the baseline direct expansion system (DX), as well 
as for the secondary loop system (2LP) was conducted according to the SAE 
International?Surface?Vehicle?Standard?for?the??Procedure?for?Measuring?System?
87 
COP?of?a?Mobile?Air?Conditioning?System?on?a?Test?Bench? [117]. The operating 
conditions, detailed in the standard, are shown in Table ?7.2. 
Table ?7.2: Charge Optimization Operating Conditions 
Parameter Magnitude 
Ambient Temperature [?C] 40 
Ambient Relative Humidity [%] 40 
Compressor Speed [rpm] 2000 
Condenser Subcooling [K] >5 
Evaporator Face Velocity [m/s] 2.7 
Condenser Face Velocity [m/s] 3.5 
 
Charge optimization results for the baseline system are shown in 
Table ?7.2 for the R134a direct expansion system (DX), R152a DX, and R152a 
2LP. For R134a, a charge of 1,730 g was found to results in the best system 
performance, balancing cooling capacity and COP. The use of R152a reduced 
the charge by roughly 19%. Changing the system to a secondary loop system 
decreased the charge 28% from the original R134a DX system. When comparing 
refrigerant charges in this research, it is implied that the refrigeration cycles, both 
for DX and 2LP systems, are larger compared to typical MAC cycles. In the 
present research, the layout of the laboratory facilities resulted in an increased 
length of vapor and liquid lines, compared to an original automotive vapor 
compression cycle. 
Table ?7.3: Charge Optimization Results: 2LP Versus DX 
Refrigerant 
/System 
Optimized 
Charge [g] 
Charge 
(% R134a DX) 
R134a DX 1,730 0% 
R152a DX 1,400 -19% 
R152a 2LP 1,240 -28% 
88 
Figure ?7.1 shows a theoretical estimate of a more realistic charge for 
direct expansion and secondary loop systems. It was assumed that an original 
MAC system would not include the liquid line MFR meter, and would feature a 
reduced receiver and filter dryer volume compared to the present test facility. 
Heat exchanger volume, compressor volume, and vapor line volumes were kept 
the same as in the present research. Liquid line size was varied within 
reasonable boundaries. It can be observed that R152a charge may be reduced 
by as much as 25%, if all other components remain the same.  
 
Figure ?7.1:Estimated Charge Savings for R152a Secondary Loop Systems 
 
89 
7.2 Steady-state Performance - Secondary Loop Versus Direct 
Expansion 
Steady-state performance measurements were conducted to compare the 
2nd loop system to a DX system without the introduction of transient effects. 
Comparison of steady-state cooling capacity and power consumption are 
important measures to evaluate suitability of alternative refrigerants and 
operating systems.  
Figure ?7.2 shows an illustration of the difference between the expected 
COP of a secondary loop system and a direct expansion system. Given the same 
refrigerant, primary loop components, and operating conditions, the COP of a 
secondary loop system is expected to be smaller compared to a DX system, due 
to the additional pump work, more surface area for heat losses (or heat 
introduction), and a small amount of heat introduced to the coolant by the pump. 
 
Figure ?7.2: Secondary Loop Steady-state COP Illustration 
90 
Figure ?7.3 shows system temperatures and pressures at various state 
points throughout the vapor compression system. R152a is working at a 
significantly decreased pressure compared to R134a. When switching from direct 
expansion to secondary loop system, pressure was elevated only slightly, mostly 
due to reduced pressure drop and different heat transfer area at the intermediate 
heat exchanger. Temperatures were changing only slightly when using R152a 
instead of R134a. Condensing temperature was elevated by 1 K, while 
evaporating temperature remained the same. When operating the secondary 
loop system, condensing temperature stayed the same within the uncertainty of 
measurement, while evaporating temperature decreased by 2.5 K, due to the 
replacement of the evaporator with the intermediate heat exchanger. 
 
Figure ?7.3: System Temperature and Pressure State Point Profiles (2LP Versus DX) 
 
91 
Figure ?7.4 shows a summary of steady-state results for the comparison of 
R152a direct expansion (R152a DX) and R152a secondary loop system (R152a 
2LP) to an R134a direct expansion system (R134a DX). Results are provided  
relative to R134a DX results. For reference, absolute steady-state data for the 
R134a DX system is provided in ?Appendix I. R152a was used as a drop-in, 
effectively working under the same operating conditions in the same equipment. 
Measurements were recorded for a range of ambient temperatures from 15?C 
through 45?C. However, Figure ?7.4 shows only 35?C data, while a more 
comprehensive summary of steady-state data for the 2LP versus DX comparison 
is given in ?Appendix D. Compressor speed of 2,500 rpm was used to simulate 
highway driving, while a compressor speed of 900 rpm was used for idling tests. 
Figure ?7.4 a) shows a comparison of performance metrics, specifically 
cooling capacity, power consumption, and COP. Cooling capacity of R152a in a 
direct expansion system was virtually the same as R134a for highway speed 
tests, and about 8% decreased for idle tests. At idle speed, the secondary loop 
system showed a capacity similar to the R152a DX system, while a slightly 
decreased capacity was observed at highway speed, possibly due to increased 
heat introduction at the coolant piping and increased heat introduction by the 
pump. Power consumption of R152a DX was reduced by more than 10% at 
highway speed and idle speed, respectively. R152a 2LP power consumption 
increased only slightly compared to R152a DX, mainly due to power consumed 
by the coolant pump, but decreased more than 10% compared to R134a DX. At 
idling speed, pump power maked up for a greater fraction of total power, resulting 
92 
in a slightly larger deviation from R152a DX power. COP of R152a DX was more 
than 10% increased above R134a DX COP, for both highway and idle tests. COP 
of secondary loop system was reduced, compared to R152a DX, but between 5% 
and 10% greater compared to R134a DX. Due to smaller absolute magnitudes 
and more fluctuation in superheat control, relative uncertainties were greater 
during idling speed tests compared to highway speed tests. 
Figure ?7.4 b) shows a comparison of volumetric and isentropic efficiencies. 
While at idling speed the volumetric efficiencies for the R152a DX and R152a 
2LP were effectively the same compared to R134a DX, the volumetric efficiency 
of R152a was increased 5% and 12% for DX and 2LP, respectively. Isentropic 
efficiency of R152a DX wais about 10% increased while idling, while it was 
slightly decreased during highway driving. R152a 2LP isentropic efficiency was 
increased by 13% and 5% during idling and highway driving, respectively.  
Figure ?7.4 c) shows a comparison of sensible and latent cooling 
performance at the evaporator (DX) and cooler (2LP). Sensible cooling capacity 
of R152a DX was within 5% of R134a DX for both, idling and highway driving. 
The R152a 2LP system showed a  similar sensible performance compared to the 
R134a DX system for idling and about 7% decreased for highway driving. Latent 
capacity for both, R152a DX and R152a 2LP was similar to the R134a DX case. 
During idling, both R152a DX and 2LP showed a significant decrease of -20%, 
albeit these measurements are associated with high uncertainties and a concrete 
conclusion cannot be drawn. The sensible heating factor (SHF), both at idling 
and highway speed, did not deviate from the R134a DX SHF.  
93 
 
Figure ?7.4: Steady State Performance Results (2LP Versus DX) 
94 
7.3 Transient Performance - Secondary Loop Versus Direct 
Expansion 
Chapter ?7.3 provides experimental results about the transient performance 
of the R152a secondary loop system with respect to a direct expansion system. 
Secondary loop systems have an inherent transient performance penalty, when 
compared to a direct expansion system. The performance penalty is explained by 
Figure ?7.5, which illustrates typical transient trends of cooling capacity during a 
pull-down test. Cabin temperature is high initially, resulting in a high cooling 
capacity. As the cabin is cooled down temperature difference decreases, 
consequently resulting in a decrease in cooling capacity. Since thermal mass, 
such as pipes and heat exchangers, is being cooled down, only part of the total 
cooling capacity of  the refrigerant-side will be available to cool down the air. 
Eventually, as the system approaches steady-state, capacity available on the 
refrigerant-side is close to capacity delivered on the air-side. The area below the 
capacity curve is the total energy used for cooling over the test. Secondary loop 
systems have an increased thermal mass, due additional heat exchangers, 
coolant pipes, and the coolant itself. Therefore, an additional decrease of energy 
available to cool down the air can be expected.   
Figure ?7.6 shows an illustration of  COP over time during a pull-down test. 
Due to high initial cooling capacity, COP is high when the passenger cabin is 
warmed up. As the cabin cools down, COP decreases according to the decrease 
in cooling capacity. Therefore, pull-down tests tend to show a higher average 
COP for short tests, while longer tests tend to show a lower average COP. 
95 
 
Figure ?7.5: Secondary Loop Transient Cooling Capacity Illustration 
 
 
Figure ?7.6: Secondary Loop Transient Capacity and COP Illustration 
96 
7.3.1 Pull-down Performance - 2LP Versus DX 
Pull-down tests were performed to evaluate cooling capacity and time 
needed to pull the cabin down to comfort temperature. Tests were started with 
the cabin in a soaked condition, meaning that cabin temperature was elevated 
above ambient temperature. As the air-conditioning system was turned on, the 
compressor delivered cooling capacity to start decrease the supply air 
temperature to the cabin. Depending on compressor speed, ambient condition, 
and cooling loads, the time taken to pull the cabin down to comfort setpoint 
varied. The comfort temperature, after reaching which a test was aborted, was 
defined as 24?C.  
Figure ?7.7 shows a comparison of time taken to pull the cabin down to 
comfort temperature. Time is given relative to the time required by the R134a DX 
system. Using R152a as a drop-in in the direct expansion system either reduced 
or kept constant the time required to achieve comfort condition. At high ambient 
temperatures and highway speed, time to comfort was reduced by more than 
15%, while at 30?C time to comfort equaled the result for the R134a DX system. 
At 30?C ambient temperature idling, R152a DX decreases time to comfort by less 
than 10%. Using the secondary loop system increased time to comfort for 
highway driving speed. At 30?C, time increased by 22%, while at 40?C time 
increased by 12%. At 30?C idling, time to comfort for the secondary loop system 
was similar to the R152a DX system. The system under test lacked the capacity 
to pull the cabin down to comfort temperature while idling at 41?C ambient 
97 
temperature. Therefore, idling pull down at 41?C was aborted after 60 minutes for 
both systems and time to comfort showed no relative deviation. 
 
Figure ?7.7: Pull-down - Time to Cabin Comfort (2LP Versus DX) 
 
Figure ?7.8 shows accumulated pull-down performance results. As 
described in Chapter ?4.2, capacity and power consumption were integrated over 
the time of testing and a transient performance factor was determined. The 
accumulated performance metrics are therefore not decoupled from the time to 
comfort (i.e. testing time). Figure ?7.8 a) shows the total energy which is available 
to cool the cabin air, relative to the total cooling energy of the R134a DX system. 
For the R152a DX system, the trends directly correlate with the time to comfort 
results for the respective operating conditions. As less cooling energy is available 
on the air-side when using secondary loop systems, accumulated capacity 
98 
decreases compared to R134a DX and R152a DX when time to comfort is the 
same, i.e. when idling at 41?C ambient temperature. During other tests, it simply 
takes a longer time to pull down the cabin, which increases the total energy 
required to achieve comfort temperature. Figure ?7.8 b) shows a comparison of 
total energy provided by the compressor (DX) or compressor and coolant pump 
(2LP) during pull down tests. Accumulated power consumption follows the time to 
comfort trend. The R152a DX system required significantly less energy, between 
15% and 28%, depending on compressor speed and ambient temperature. Due 
to additional coolant pump power, as well as increased time to comfort, the 
secondary loop system required a higher energy input compared to the R152a 
DX system. At idling speed, the energy input was 10% to 15% below R134a DX 
energy input, while at highway speed the energy input was 6% to 17% increased. 
Since time to comfort is short at highway speed as compared to idling speed, the 
relative difference in energy input between direct expansion and secondary loop 
systems was more pronounced. The relative transient performance factor, shown 
in Figure ?7.8 c), was about 20% increased in all tests for R152a as a drop-in in 
direct expansion systems, when compared to R134a DX. For the R152a 
secondary loop system, TPF was within 5% of R134a DX TPF for both idling and 
highway speed, as well as for both ambient temperatures.  
99 
 
Figure ?7.8: Pull-down Performance Results (2LP Versus DX) 
100 
Another way to evaluate pull-down performance is to normalize 
performance parameters with respect to time. When time-normalized, as shown 
in Figure ?7.9 a) and b), transient performance factor of the secondary loop 
system was observed to be 20% and 13% lower at 30?C and 41?C ambient 
temperature, respectively. TPF of the R152a DX system was 25% and 38% 
increased above R134a DX TPF. 
 
Figure ?7.9: Pull-down Time-normalized Performance Results (2LP Versus DX) 
-40
 -30
 -20
 -10
 0
 10
 20
 30
 40
 Q (air) E in TPF (air)
 En
 er
 gy
 ; T
 P
 F 
[%
 ]
 Ambient Temp. [?C] ; Compressor Speed [rpm]
 R134a DX R152a DX R152a 2LP
 Time-Norm. Performance (T30 hwy, % R134a DX)
 -40
 -30
 -20
 -10
 0
 10
 20
 30
 40
 Q (air) E in TPF (air)
 En
 er
 gy
 ; T
 P
 F 
[%
 ]
 Ambient Temp. [?C] ; Compressor Speed [rpm]
 R134a DX R152a DX R152a 2LP
 Time-Norm. Performance (T41 hwy, % R134a DX)
 a)
 b)
101 
7.3.2 New European Drive Cycle Performance - 2LP Versus DX 
A comparison of transient behavior and performance during the New 
European Drive Cycle (NEDC) is provided in Chapter ?7.3.2. The NEDC is used to 
evaluate power consumption of the air-conditioning unit at quasi-realistic, but 
standardized driving conditions. Due to large transients in compressor speed and 
condenser air flow rate, considerable fluctuations in cooling capacity, and 
consequently cabin supply temperature are introduced. As a secondary loop 
system has a larger thermal mass, drive cycle performance, as well as cabin air 
conditions may vary from the performance of a direct expansion systems. 
Transient trends, as well as accumulated performance data are discussed. 
Figure ?7.10 shows a transient comparison of supply temperature. As can 
be observed, using R152a as a drop-in in the direct expansion system reduced 
cabin supply temperature slightly, but not significantly. Due to the additional 
thermal mass of the secondary loop system, initial cool down of supply 
temperature was hindered and can be increased as much as 8?C during the 
initial minutes of the cycle. Close to the end of the highway section of the NEDC, 
R152a 2LP supply temperature approached the supply temperature of the direct 
expansion system. At the same time, secondary loop supply temperature was not 
subject to the same fluctuations supply temperatures of the direct expansion 
system experiences. 
Further transient trends are shown in Figure ?7.11, including trends in air-
 side cooling capacity, system power consumption, and dehumidification. As can 
be observed in Figure ?7.11 a), R152a DX transient cooling capacity profile is 
102 
similar to the R134a DX capacity profile. R152a 2LP air-side capacity was 
significantly decreased during the first minutes of the cycle. During the middle 
section of the NEDC, the average capacity was observed to be close to the direct 
expansion system average, while showing less fluctuations. During acceleration, 
the secondary loop system capacity was slower to respond, while during 
deceleration, cooling capacity was preserved above the magnitude observed for 
the direct expansion system. After an extended time into the test, during the 
highway section of the cycle, secondary loop system capacity was on par with 
the direct expansion system capacity. Figure ?7.11 b) shows transient profiles of 
power consumption. R152a as a drop-in in the direct expansion system showed 
decreased power consumption throughout the drive cycle. The R152a 2LP 
system power consumption was strongly increased during the first minutes of the 
cycle, as the refrigerant-side capacity was significantly higher compared to the 
direct expansion system (although air-side capacity was decreased). The R152a 
2LP power consumption remained higher than power consumption of the R134a 
DX system throughout the drive cycle. 
 Figure ?7.11 b) shows a comparison of transient dehumidification during 
the NEDC. Dehumidification profiles follow inverted profiles of cooling capacity, 
with R152a performing slightly worse than R134a DX, especially during the high 
speed portions of the cycle. During the mid section of the NEDC cycle, the direct 
expansion system experienced re-evaporation of condensate from the 
evaporator surface during idling. In comparison, the secondary loop system 
103 
performed worse in the beginning, but prevented re-evaporation during the mid 
section of the cycle for the most part.  
 
Figure ?7.10: NEDC Supply Temperature Trend (2LP Versus DX) 
 
Figure ?7.12 provides a comparison of accumulated NEDC performance 
data, as well as a comparison of sensible and latent performance. Figure ?7.12 a) 
shows refrigerant-side, as well as air-side accumulated performance parameters. 
R152a DX refrigerant-side, as well as air-side energy available for cooling were 
similar to the R134a DX system. Due to a decreased energy input in terms of 
compressor power, transient performance factors of R152a DX were about 10% 
increased above R134a DX TPF. The secondary loop system showed a 
significant increase of 25% in the energy available for cooling on the refrigerant-
 side. However, considerably less energy, -10%,  was available on the air-side. 
-10
 -5
 0
 5
 10
 15
 20
 25
 30
 35
 40
 0 5 10 15 20
 Tem
 per
 atu
 re 
[?C
 ]
 Time [min]
 R134a DX R152a DX R152a 2LP
 NEDC Supply Temperature
104 
 
Figure ?7.11: NEDC Performance Metrics Trend (2LP Versus DX) 
-10
 -8
 -6
 -4
 -2
 0
 2
 4
 6
 8
 10
 0 5 10 15 20
 D
 e
 h
 u
 m
 id
 if
 ic
 a
 ti
 o
 n
  [
 g
  H
 2
 O
 /
 k
 g
  A
 ir
 ]
 Time [min]
 R134a DX R152a 2LP R290 2LP
 NEDC Dehumidification
 0
 500
 1000
 1500
 2000
 2500
 3000
 3500
 4000
 4500
 0 5 10 15 20
 P
 o
 w
 e
 r 
[W
 ]
 Time [min]
 R134a DX R152a DX R152a 2LP
 NEDC Power Consumption (C+P)
 1000
 20
 3000
 4000
 5000
 6000
 0 5 10 15 20
 C
 a
 p
 a
 c
 it
 y
  [
 W
 ]
 Time [min]
 R134a DX R152a DX R152a 2LP
 NEDC Cooling Capacity (air)
 a)
 b)
 c)
105 
Since the NEDC cycle is only little more than 18 minutes long, a significant 
amount of total energy is spent on cooling thermal mass of the system. Since the 
total energy input to the secondary loop system was increased by about 5%, the 
air-side transient performance factor was decreased by 14%. Figure ?7.12 b) 
shows accumulated sensible and latent capacity. It can be observed that the 
sensible performance of R152a DX was similar to R134a DX. 2LP accumulated 
sensible capacity was decreased by 10%, following the trend of the accumulated 
air-side capacity. Accumulated capacity for both the R152a DX and the R152a 
2LP system were reduced by 20%. The total amount of latent capacity integrated 
over the NEDC test time was small compared to sensible capacity. At the same 
time, measurement uncertainty of latent performance is high. The 
dehumidification performance follows the measurement of the latent capacity. 
The R152a DX as well as the R152a 2LP systems showed 20% reduced 
dehumidification performance as compared to R134a DX. 
106 
 
Figure ?7.12: NEDC Accumulated Performance Metrics Results (2LP Versus DX) 
 
7.4 Summary of Results: Secondary Loop Versus Direct Expansion 
? Charge optimization for the R152a 2LP system showed a 28% 
reduction in refrigerant charge compared to the R134a DX system. A 
theoretical estimation of charge reduction for units built into a small to 
mid-size vehicle showed that by changing from R134a to R152a the 
charge can be reduced by about 24%, based on liquid line density. 
a)
 b)
 -40
 -30
 -20
 -10
 0
 10
 20
 30
 40
 Q (air) Q (ref) E in TPF (air) TPF (ref)
 R
 el
 at
 iv
 e 
En
 er
 gy
 ; T
 P
 F 
[%
 ]
 Performance Metrics
 R134a DX R152a DX R152a 2LP
 Acc. Performance Parameters (NEDC, % R134a DX)
 -40
 -30
 -20
 -10
 0
 10
 20
 30
 40
 Dehumid. Q sensible Q latent
 D
 eh
 u
 m
 id
 . [
 %
 ] 
En
 er
 gy
  [%
 ]
 Performance Metrics
 R134a DX R152a DX R152a 2LP
 Sensible/Latent Energy (NEDC, %R134a DX)
107 
When switching from R152a DX to R152a 2LP, an additional 25% of 
charge can be saved. 
? Steady-state results showed that the cooling capacity of secondary 
loop systems was similar to the direct expansion system using the 
same refrigerant. R152a power consumption was more than 10% 
reduced compared to R134a, which allowed the R152a 2LP system to 
have 5% to 10% higher COP compared to the R134a DX system. 
Compressor isentropic and volumetric efficiencies were determined to 
be similar or higher compared to R134a. Sensible and latent cooling 
performance was within 10% of the R134a DX system. At idling 
conditions, latent cooling capacity might be decreased, but high 
uncertainties prevent  a definite statement. 
? Transient pull-down results showed that as expected, time to pull down 
to comfort set point was increased for secondary loop systems. For the 
amount of coolant in the present research (~ 5.5 kg), the increase can 
be as much as 25% from R152a DX to R152a 2LP at highway driving 
conditions. Consequently, the accumulated power during a pull down 
was increased by about 30% from R152a DX to R152a 2LP, and 
therefore also increased above R134a DX results. The transient 
performance factor for pull-down was within 5% of R134a DX and 
about 20% decreased compared to R152a DX. 
? During NEDC operation, the energy provided to cool down the supply 
air was decreased by about 10% from R134a DX. At the same time, 
energy consumption was increased by about 5%. Consequently, air-
 side TPF was decreased by more than 10%. Latent energy used for 
dehumidification was reduced by about 20% for both, R152a DX and 
R152a 2LP. Sensible energy for the secondary loop system was 
reduced by 10%, due to sensible cooling of the additional thermal 
mass.  
108 
8 Experimental Results: Propane in Secondary Loop Systems 
Chapter ?8 shows experimental results which provide a basis for 
understanding the performance of Propane (R290) in mobile air-conditioning 
systems (MACs) as an alternative to R152a. As a low-GWP refrigerant with 
excellent thermophysical characteristics, Propane has been gaining attention in 
residential and commercial applications in recent years. Due to its high 
flammability, the use of an indirect system, such as the secondary loop system, 
is imperative for its use in automotive air-conditioning.  
The normal boiling point of R290, as shown in Table ?6.1, is significantly 
lower compared to both, R134a and R152a. As a result, R290 operates at higher 
pressures. R290 shows a significant increase in compressor power consumption, 
as well as a significantly increased volumetric capacity. 
Figure ?8.1 shows relative drop-in performance of R290 in the direct 
expansion system. Both at idling and highway driving, R290 showed an increase 
in power consumption by 40% above R134a results. Cooling capacity was 
increased by about 22%. In contrast, R152a showed a power consumption which 
was decreased by 12% - 18%, while cooling capacity was similar to R134a. As a 
result, COP of R290 was observed to be between 10% and 15% decreased 
compared to R134a, while COP of R152a was increased by about the same 
amount.  
As a result of increase of capacity and the decrease of COP during drop-in 
tests, subsequent tests with Propane in the secondary loop system were 
109 
performed with reduced compressor speeds. The compressor rpm for R290 2LP 
tests was modified to result in similar cooling capacities and transient cooling 
performance compared to R134a DX. R134a and R152a were operated at 
speeds of 2,500 and 900 for highway driving and idling. For R290, compressor 
speeds were reduced to 1,550 and 650 rpm for highway driving and idling, 
respectively. In an automobile this could be achieved by either using a 
compressor with smaller displacement volume, or by changing pulley size at the 
engine to adjust the gear ratio between engine and compressor.  
 
Figure ?8.1: R290 Drop-in Performance in the Direct Expansion System 
 
8.1 Charge Optimization: Propane in Secondary Loop Systems 
Due to lower liquid density, Propane was found to have a decreased 
optimum charge compared to R134a and R152a. Table ?8.1 shows charge 
-60
 -40
 -20
 0
 20
 40
 60
 COP   
(idle)
 COP   
(hwy)
 Q (cool) 
(idle)
 Q (cool) 
(hwy)
 Power 
(idle)
 Power 
(hwy)
 Ca
 pa
 cit
 y [
 %]
  ; C
 OP
  [%
 ]
 Performance Metrics
 R134a DX R152a DX R290 DX
 Performance Comparison (T35RH40, % R134a BSL)
110 
optimization results for the R290 secondary loop system in comparison to the 
R134a baseline and the R152a secondary loop system. A charge reduction of 60% 
was observed when changing the R134a DX system to a R290 2LP system. The 
decrease in charge is roughly twice the amount observed for the R152a 2LP 
system. This amounts to a 44% decrease in charge when replacing R152a with 
R290 in the secondary loop system. 
Table ?8.1: Charge Optimization Results: R290 2LP 
Refrigerant 
/System 
Optimized 
Charge [g] 
Charge 
(% R134a DX) 
R134a DX 1,730 0% 
R152a 2LP 1,240 -28% 
R290 2LP 700 -60% 
 
Figure ?8.2 shows the estimated charge savings, considering realistic pipe 
lengths in an automobile. For this estimation, vapor line lengths remained the 
same as in the 2LP test facility, but liquid line length was varied within 
reasonable boundaries, as shown in Figure ?8.2. Additionally, the volume of the 
mass flow rate meter was omitted, and the volumes of filter-drier and receiver in 
the liquid line were reduced to fit commercially available products. Based on 
above assumptions, Propane charge could be reduced to as little as 380 g for 
baseline systems, and 320 g for secondary loop systems. It is estimated that a 
realistic charge of Propane in secondary loop systems could stay below 350 g. 
On average, this could result in a charge reduction of roughly 43% when 
replacing R152a in secondary loop systems.  
111 
 
Figure ?8.2: Estimated Charge Savings for R290 Secondary Loop Systems 
 
8.2 Steady-state Performance: Propane in Secondary Loop Systems 
Chapter ?8.2 shows relative steady-state performance results of R290 in 
the 2LP system, as compared to R134a DX and R152a 2LP results. System 
performance,  compressor efficiencies, and sensible/latent capacity are provided.  
Figure ?8.3 shows temperature and pressure at several locations in the 
refrigeration cycle when operating at highway driving speed and ambient 
conditions of 35?C and 40% relative humidity. Temperature is shown on the first 
y-axis, while pressure is shown on the second y-axis. Propane operates at 
pressures well above R134a pressures. Suction pressure increased to about 500 
kPa, while high side pressure increased to about 1,750 kPa. Suction temperature 
112 
was very similar to suction temperature of R134a DX and R152a 2LP, while 
discharge temperature was significantly reduced (8 K) with respect to R134a DX. 
 
Figure ?8.3: System Temperature and Pressure State Points (R290 2LP) 
 
Figure ?8.4 shows R290 2LP performance metrics, compressor efficiencies, 
and sensible/latent performance relative to R134a DX. For reference, absolute 
steady-state data for the R134a DX system is provided in ?Appendix I. Figure ?8.4 
a) provides experimental results for cooling capacity, power consumption, and 
COP during highway driving and idling conditions. Since R290 2LP compressor 
rpm was adjusted, as described in Chapter ?8, performance results are not drop-in 
results. R290 2LP capacity was adjusted to show similar transient performance 
as the R134a DX system. Steady-state cooling capacity was observed to be 
reduced by about 10% for highway driving speed and about 20% for idling, while 
R152a 2LP cooling capacity stayed within -5% for both driving conditions. Due to 
0
 200
 400
 600
 800
 1000
 1200
 1400
 1600
 1800
 2000
 0
 10
 20
 30
 40
 50
 60
 70
 80
 90
 100
 Compressor 
Suction
 Compressor 
Discharge
 Condenser 
Outlet
 TXV Inlet
 Pr
 es
 su
 re 
[kP
 a]
 Te
 mp
 era
 tur
 e [
 ?C
 ]
 System Location
 R134a DX (T) R152a 2LP (T) R290 2LP (T)
 R134a DX (P) R152a 2LP (P) R290 2LP (P)
 Refrigerant Temperature/Pressure (T35RH40, 2500rpm)
113 
the reduction in rpm for R290 2LP, power consumption decreased accordingly, 
resulting in a decrease of roughly 15% and 5% for highway and idling conditions, 
respectively. At the same time, R152a 2LP showed a 5% smaller decrease at 
highway speed conditions and a significantly higher decrease of 15% at idling 
conditions. Consequently, COP for R290 was increased by more than 5% at 
highway driving condition. During idling, R290 2LP COP was observed to be 
decreased by about 15%, though measurement uncertainty was increased when 
operating at idling conditions. Figure ?8.4 b) shows relative changes in isentropic 
and volumetric efficiency. R290 2LP displayed a 10% increase in both, isentropic, 
as well as volumetric efficiency with respect to R134a DX. However, during idling 
a significant decrease of 20% and 25% was observed for volumetric and 
isentropic efficiency, respectively. It was determined that if compressor 
efficiencies at idling were similar to R134a DX efficiencies, due to optimization of 
the compressor for R290, idling COP would be increased 15% above R134a DX. 
Figure ?8.4 c) shows a comparison of sensible and latent cooling 
performance for idling and highway speed driving conditions. R290 2PL sensible 
capacity was similar to R134a DX and R152a 2LP sensible capacity during idling, 
while it was decreased by about 8% under highway driving conditions, much 
similar to R152a 2LP. Latent capacity for R290 2LP was measured to be zero 
during idling, while latent capacity of R134a DX and R152a 2LP were measured 
to be 280 W and 230 W, respectively. However, latent capacity during idling was 
associated with high measurement uncertainties, leaving the R290 2LP latent 
performance comparison without a definite conclusion.  
114 
 
Figure ?8.4: Steady-state Performance Results (R290 2LP) 
-30
 -20
 -10
 0
 10
 20
 30
 COP   
(idle)
 COP   
(hwy)
 Q (cool) 
(idle)
 Q (cool) 
(hwy)
 Power 
(idle)
 Power 
(hwy)
 C
 a
 p
 a
 c
 it
 y
  [
 %
 ]
  ;
  C
 O
 P
  [
 %
 ]
 Performance Metrics
 R134a DX R152a 2LP R290 2LP
 Performance Comparison (T35RH40, % R134a DX)
 -40
 -30
 -20
 -10
 0
 10
 20
 30
 40
 ? vol (idle) ? vol (hwy) ? isen (idle) ? isen  (hwy)
 E
 ff
 ic
 ie
 n
 c
 y
  [
 %
 ]
 Efficiency Metrics
 R134a DX R152a 2LP R290 2LP
 Efficiency Comparison (T35RH40, % R134a DX)
 -30
 -20
 -10
 0
 10
 20
 30
 Q sens 
(idle)
 Q sens 
(hwy)
 Q lat   
(idle)
 Q lat   
(hwy)
 SHF    
(idle)
 SHF   
(hwy)
 E
 ff
 ic
 ie
 n
 c
 y
  [
 %
 ]
 Efficiency Metrics
 R134a DX R152a 2LP R290 2LP
 Sensible/Latent Comparison (T35RH40, % R134a DX)
 a)
 b)
 c)
115 
The sensible heat factor of R290 2LP was similar to R134a DX SHF when 
operating at highway conditions and increased by roughly 10% during idling. 
Based on the assumption that compressor design could be changed to 
achieve idling efficiencies similar to R134a DX efficiencies, a theoretical estimate 
on possible R290 2LP performance was performed using Engineering Equation 
Solver. It was furthermore assumed that suction state properties, as well as 
discharge pressure would not change from the previously measured values. With 
isentropic and volumetric efficiencies similar to the R134a DX experimental 
values, idling capacity would change significantly, while power consumption 
would stay nearly the same. As a result, R290 2LP idling COP would increase to 
about 15% above R134a DX COP, as observed in Figure ?8.5, which would also 
translate into enhanced transient drive cycle performance.  
 
Figure ?8.5: Theoretical R290 2LP Performance Results with Increased Idling 
Compressor Efficiencies 
-30
 -20
 -10
 0
 10
 20
 30
 COP   
(idle)
 COP   
(hwy)
 Q (cool) 
(idle)
 Q (cool) 
(hwy)
 Power 
(idle)
 Power 
(hwy)
 Ca
 pa
 cit
 y [
 %
 ] ;
  C
 O
 P 
[%
 ]
 Performance Metrics
 R134a DX R152a 2LP R290 2LP
 Performance Comparison (T35RH40, % R134a DX)
116 
8.3 Transient Performance: Propane in Secondary Loop Systems 
Chapter ?8.3 provides transient experimental results for the R290 
secondary loop system in comparison to the R134a DX and the R152a 2LP 
systems. Pull-down performance at several ambient conditions and compressor 
speeds, as well as NEDC drive cycle performance, are discussed.  
 
8.3.1 Pull-down Performance: Propane 2LP 
Figure ?8.6 shows time to comfort information for R290 2LP, relative to the 
R134a DX system. At highway speed, time to comfort was increased by 30% and 
20% for 31?C ambient temperature and 41?C ambient temperature, respectively. 
At both conditions, the R290 2LP time to comfort was increased by about 10% 
with respect to R152a 2LP measurements. At 41?C idling, the air-conditioning 
system was not able to pull the cabin down to comfort setpoint, resulting in the 
abortion of the test for all systems after 60 minutes. Therefore, no difference in 
time to comfort can be observed at 41?C idling conditions. At 30?C idling, the 
R290 2LP system showed an increase of time to comfort above 60%. This was a 
result of the poor steady-state performance at low and moderate temperatures 
during idling. However, pull-down results may change, depending on how 
compressor rpm, compressor displacement, or pulley ratio are adjusted. The 
adjustment of compressor rpm for R290 2LP added to the increase in time to 
comfort. Therefore, time to comfort should not be taken as final result for the 
R290 2LP system, but rather as a parameter which can be adjusted. 
117 
 
Figure ?8.6: Pull-down Time to Cabin Comfort (R290 2LP) 
 
Figure ?8.7 shows relative results for accumulated performance metrics. 
Accumulated cooling capacity, power consumption, and transient performance 
factor results are shown. Due to accumulated capacity and power consumption 
being heavily influenced by time to comfort (i.e. length of the test), the rpm 
adjustment for R290 2LP affects capacity and power consumption. However, 
since both are affected in the same way, it can be expected that transient 
performance factor results retain their validity when rpm is adjusted within 
reasonable boundaries.  Figure ?8.7 a) shows relative results of energy available 
for cooling cabin air over the time of a pull-down test. At highway driving 
conditions, the R290 2LP system showed a 20% and 15% increased capacity for 
30?C and 40?C ambient temperature, respectively. In both cases, cooling energy 
was elevated above R152a 2LP values. Accumulated cooling capacity trends 
118 
followed the trends for time to comfort. At idling conditions, accumulated cooling 
capacity was decreased by 16% at 41?C. As test time is constant for the three 
systems, trends resemble steady-state results. However, at 30?C, R290 2LP 
accumulated capacity was increased by more than 40%, due to the long testing 
time. Figure ?8.7 b) shows relative results for a comparison of power consumption. 
At highway speeds, the R290 2LP system showed an increase in power 
consumption of about 25%. While the power consumption matched R152a 2LP 
power consumption at moderate temperatures, it increased by 10% over the 
R152a 2LP power consumption at 41?C ambient temperature. At idling conditions, 
R290 2LP power consumption was 15% higher compared to R134a DX for 41?C 
ambient temperature, while R152a 2LP consumption was about 9% increased. At 
30?C ambient temperature, R209 2LP power consumption is increased by about 
100% compared to R134a DX consumption. Figure ?8.7 c) shows a relative 
comparison of air-side transient performance factor. At highway speed, R290 
2LP was performing as well (41?C) or better (30?C) as R134a DX and R152a DX. 
However, at idling speed transient performance was reduced by about 10% to 
15%.  
Figure ?8.8 shows time-normalized results at highway driving conditions for 
ambient temperatures of 30?C and 41?C with respect to R134a DX performance. 
When time-normalized, accumulated cooling capacities were reduced by 9% 
(30?C) and 7% (41?C). The time-normalized R290 2LP results for cooling 
capacity were matching R152a 2LP results. Time-normalized energy 
consumption was decreased by 11% (30?C) and 6% (41?C).   
119 
 
Figure ?8.7: Pull-down Performance Results (R290 2LP) 
-50
 -40
 -30
 -20
 -10
 0
 10
 20
 30
 40
 50
 T30 idle T41 idle T30 hwy T41 hwy
 E
 n
 e
 rg
 y
  [
 %
 ]
 Ambient Temp. [?C] ; Compressor Speed [rpm]
 R134a DX R152a 2LP R290 2LP
 Accumulated Cooling Capacity (air, % R134a DX)
 -40
 -30
 -20
 -10
 0
 10
 2
 30
 40
 T30 idle T41 idle T30 hwy T41 hwy
 E
 n
 e
 rg
 y
  [
 %
 ]
 Ambient Temp. [?C] ; Compressor Speed [rpm]
 R134a DX R152a 2LP R290 2LP
 Accumulated Power Consumption (% R134a DX)
 -30
 -20
 -10
 0
 1
 20
 30
 T30 idle T41 idle T30 hwy T41 hwy
 T
 P
 F
  [
 %
 ]
 Ambient Temp. [?C] ; Compressor Speed [rpm]
 R134a DX R152a 2LP R290 2LP
 Transient Performance Factor (air, % R134a DX)a)
 b)
 c)
120 
In the case of 30?C ambient temperature, R290 2LP time-normalized 
energy consumption was 7% lower compared to R152a 2LP. The time-
 normalized transient performance factor of R290 2LP was decreased by about 
20% for both ambient temperatures. In the case of 41?C, R290 2LP time-
 normalized TPF was about 6% lower compared to R152a 2LP. 
 
Figure ?8.8: Pull-down Time-normalized Performance Results (R290 2LP) 
 
-30
 -20
 -10
 0
 10
 20
 30
 Q (air) E in TPF (air)
 En
 er
 gy
 ; T
 P
 F 
[%
 ]
 Ambient Temp. [?C] ; Compressor Speed [rpm]
 R134a DX R152a 2LP R290 2LP
 Time-Norm. Performance (T30 hwy, % R134a DX)
 -30
 -20
 -10
 0
 10
 20
 30
 Q (air) E in TPF (air)
 En
 er
 gy
 ; T
 P
 F 
[%
 ]
 Ambient Temp. [?C] ; Compressor Speed [rpm]
 R134a DX R152a 2LP R290 2LP
 Time-Norm. Performance (T41 hwy, % R134a DX)
 a)
 b)
121 
8.3.2 New European Drive Cycle Performance - Propane 2LP 
Figure ?8.9 shows transient trends for cabin supply temperature. The 
supply temperature profile of R290 2LP showed the same trends as the R152a 
2LP profile, but was elevated by about 1 K to 1.5 K. This was due to the adjusted 
compressor rpm for R290 2LP tests. Both for R152a 2LP and for R290 2LP the 
supply temperature profile showed less fluctuations compared to the R134a DX 
profile, due to the thermal mass of the coolant in the secondary loop. At the end 
of the drive cycle, supply temperature of R290 2LP was within 1 K of the supply 
temperature of the R134a DX profile. 
 
Figure ?8.9: NEDC Supply Temperature Trend (R290 2LP) 
 
Figure ?8.10 shows a comparison of relative performance metrics results. 
As can be observed in Figure ?8.10 a), the cooling capacity of the R290 2LP 
system follows the same profile as the R152a 2LP system. R290 2LP, as well as 
0
 5
 10
 15
 20
 25
 30
 35
 40
 0 5 10 15 20
 Tem
 per
 atu
 re 
[?C
 ]
 Time [min]
 R134a DX R152a 2LP R290 2LP
 NEDC Supply Temperature
122 
R152a 2LP air-side cooling capacity was decreased significantly during the first 
minutes of the cycle, due to the thermal mass of the secondary loop system. 
R290 2LP power consumption, shown in Figure ?8.10 b), followed the same 
profile as R134a DX power consumption. R290 2LP power consumption was 
lower than R152a 2LP power consumption during the high rpm sections of the 
cycle. The R290 2LP dehumidification profile, shown in Figure ?8.10 c), showed 
the same trend as the R152a 2LP profile, though with a slightly increased 
dehumidification throughout the drive cycle. Again, the differences to the direct 
expansion dehumidification trend, specifically reduced fluctuation due to a 
decrease in re-evaporation, become apparent.  
Figure ?8.11 introduces relative accumulated NEDC performance results 
for the R290 2LP system. Air-side capacity was reduced by 10%, similar to 
R152a 2LP air-side capacity. However, the power consumption was reduced by 
5%, resulting in a decrease of TPF of 5% with respect to R134a DX. This is a 
small improvement, +8%, with respect to R152a 2LP. As discussed in previous 
chapters, the refrigerant-side TPF was increased compared to R134a, while the 
air-side TPF was decreased. The largest contributors to this phenomenon are the 
thermal mass of the secondary loop, which reduces the accumulated air-side 
capacity, and the pump power, which steadily increases the accumulated total 
power consumption of the system. 
123 
 
Figure ?8.10: NEDC Performance Metrics Trends (R290 2LP) 
0
 1000
 2000
 3000
 4000
 5000
 6000
 0 5 10 15 20
 C
 a
 p
 a
 c
 it
 y
  [
 W
 ]
 Time [min]
 R134a DX R152a 2LP R290 2LP
 NEDC Cooling Capacity (air)
 0
 500
 1000
 1500
 2000
 2500
 3000
 3500
 4000
 4500
 0 5 10 15 20
 P
 o
 w
 e
 r 
[W
 ]
 Time [min]
 R134a DX R152a 2LP R290 2LP
 NEDC Total Power
 -10
 -8
 -6
 -4
 -2
 0
 2
 4
 6
 8
 10
 0 5 10 15 20
 D
 e
 h
 u
 m
 id
 if
 ic
 a
 ti
 o
 n
  [
 g
  H
 2
 O
 /
 k
 g
  A
 ir
 ]
 Time [min]
 R134a DX R152a 2LP R290 2LP
 NEDC Dehumidification
 b)
 c)
124 
 
Figure ?8.11: NEDC Accumulated Performance Metrics Results (R290 2LP) 
 
8.4 Summary: Propane as Refrigerant in Secondary Loop Systems 
? Using R290 as a drop-in with components optimized for R134a under 
the same operating conditions resulted in increased cooling capacities 
and increased power consumption. R290 2LP COP during drop-in 
tests was reduced by more than 10% compared to R134a DX. 
Therefore, R290 tests were carried out with reduced compressor rpm 
-40
 -30
 -20
 -10
 0
 10
 20
 30
 40
 Dehumid. Q sensible Q latent
 D
 eh
 u
 m
 id
 . [
 %
 ],
  E
 n
 er
 gy
  [%
 ]
 Performance Metrics
 R134a DX R152a 2LP R290 2LP
 Sensible/Latent Energy (NEDC, %R134a DX)
 a)
 b)
 -4
 -3
 -
 -1
 10
 20
 30
 40
 Q (air) Q (ref) E in TPF (air) TPF (ref)
 En
 er
 gy
 ; T
 P
 F 
[%
 ]
 Performance Metrics
 R134a DX R152a 2LP R290 2LP
 Acc. Performance Parameters (NEDC, % R134a DX)
125 
to simulate a smaller compressor and decrease cooling capacity to 
comparable levels. 
? Charge optimization for R290 2LP reduced refrigerant charge by 60% 
in comparison to R134a DX. A theoretical estimation showed that 
R290 charge may be about 60% smaller compared to R134a, based 
on liquid density. In an air-conditioning system for a small to mid-size 
vehicle, charge may be reduced by an additional 20% for secondary 
loop systems in comparison to direct expansion systems.  
? Steady-state performance of R290 in secondary loop systems changed 
significantly with compressor speed. At highway driving conditions, 
R290 2LP COP was about 8% higher than R134a DX COP, and 
therefore similar to R152a 2LP COP. At idling conditions, R290 2LP 
COP was reduced by 15% from R134 DX COP. This was due to 
volumetric and isentropic compressor efficiencies, which were reduced 
by up to 25% compared to R134a DX. If compressor efficiencies could 
be increased to R134a DX levels by using a compressor which is 
optimized for R290, COP would be increased 14% above R134a DX. 
Sensible and latent cooling capacity were decreased slightly from 
R152a 2LP results, due to a decrease in overall cooling capacity 
caused by the reduction in compressor rpm. 
? Transient pull-down results showed that time to comfort increased by 
20% to 30% for highway driving in comparison to R134a DX. This was 
due to the choice of reduced rpm and the subsequent availability of 
cooling capacity. Therefore, experimental results for pull down 
accumulated cooling capacity will vary, based on which rpm (or in a car: 
which compressor size) is chosen. However, TPF results show that 
transient performance of R290 2LP was similar or better than R134a 
DX during highway driving conditions. At idling conditions, TPF was 
decreased by 10% to 15%. A compressor designed specifically for 
R290 can significantly increase TPF at low compressor speeds. 
126 
? NEDC results show that R290 2LP transient performance was 
decreased by 5% with respect to R134a DX, as opposed to nearly 15% 
for R152a 2LP. Sensible capacity was about 14% decreased, which 
was influenced by the reduction in rpm. Latent capacity was about the 
same as R134a DX results, although measurement uncertainty was 
significant.  
? It was shown that R290 2LP steady-state COP, as well as transient 
TPF could be similar or higher, compared to R134a DX. However, 
reduced COP at low compressor speeds prevent the R290 2LP NEDC 
TPF to increase above R134a DX TPF. If components were used, 
which were optimized for use with R290, steady-state, as well as 
transient performance could improve well above R134a DX 
performance. 
Based on measurements of the components used in the experimental test 
facility, as well as a reasonable estimate of refrigerant and coolant line lengths in 
an mid-size sedan, the mass of the secondary loop system was assumed to be 
increased by about 24%, compared to a DX system. The increase of mass is due 
to the adding of coolant, coolant tubing, and secondary loop components, such 
as the coolant pump and an intermediate heat exchanger. Fischer [118] 
estimated an annual increase in fuel consumption of about 3.4 gallons (12.7 liters) 
due to the mass of a regular air-conditioning system. Due to the estimated 
increase in mass by 24%, when switching to a secondary loop system (as 
discussed above), an additional 0.82 gallons (3.1 liters) of fuel consumption 
might be expected per year.  
 
127 
Green-Mac-LCCP [119], a software which was peer reviewed and globally 
accepted as a credible method of comparing the climate performance of mobile 
air conditioning systems, was used to estimate the impact of secondary loop 
systems on the Life Cycle The model takes into account the change of mass of 
the system, the change of charge due to different refrigerants and system layouts, 
as well as other characteristics associated with the change of life cycle climate 
performance when using different refrigerants. Figure ?8.12 shows a comparison 
of CO2 equivalent emissions per lifetime per vehicle for a range of U.S cities in 
colder, as well as warmer climate. A secondary loop system using R152a or 
R290 as refrigerant will reduce the CO2 equivalent emissions per lifetime per 
vehicle by roughly 12% to 20% (R152a) or 13% to 22% (R290) as compared to a 
R134a DX system. The decrease in CO2 equivalent emissions is mainly due to 
decrease of direct contributions (i.e. due to the decrease in negative impact the 
refrigerant itself has on the environment). Inputs to the Green-MAC-LCCP model 
were based on experimentally measured cooling capacity and COP. A secondary 
loop system which is optimized for R290 as primary refrigerant would possibly 
reduce CO2 equivalent emissions even beyond the results shown in Figure ?8.12. 
A theoretical comparison of refrigerants in Chapter ?6 showed that the 
cycle efficiency, or COP, of R290 should be expected to be 0% to 5% decreased 
compared to an R134a system. However, experimental steady-state results 
showed an increase of 10% during highway driving conditions and a 15% 
decrease during idling conditions. During transient NEDC operation, where 
compressor speeds are much higher than idling speeds for a majority of the time, 
128 
the R290 2LP system faired similar to the R134a DX system despite capacity 
disadvantage due to increased thermal mass. This discrepancy can be explained 
by taking compressor efficiencies into account. Alefeld [115] distinguished 
irreversibilities into two categories: irreversibilities caused by fluid properties and 
by system design. System design irreversibilities are partially reflected in the 
isentropic compressor efficiency term in Equations (22) and (23) in Chapter ?6. As 
observed in Chapter ?8, R290 2LP isentropic efficiency during highway driving is 
10% increased compared to R134a DX. If this increase would have been 
included in the term gc in Equations (22) and (23), theoretical analysis and 
experiment would agree within 2%.   
 
Figure ?8.12: Impact of Secondary Loop Systems on Life Cycle Climate 
Performance of a Mid-size Passenger Car 
 
0
 1000
 2000
 3000
 4000
 5000
 6000
 7000
  9
 36
  9
 14
  9
 03
  8
 98
  6
 08
  2
 ,7
 48
  8
 97
  In
 di
 re
 ct
  D
 ire
 ct
 -17.31%
 -19.07%
 -12.96%
 -19.21%
 -22.12%
 -18.24%
 -12.18%
 -13.40%
 -13.87%
 -15.00%
 -19.86%
 -15.42%
 MiamiSan DiegoSan FranciscoBostonChicagoFargo
  
 
LC
 C
 P 
C
 O
 2 Eq.
  e
 m
 iss
 io
 ns
  p
 er
  li
 fe
 tim
 e/
 ve
 hi
 cl
 e 
(k
 g)
 US cities
  Baseline HFC-134a
  2LP HFC-152a
  2LP HC-290
 Cold Hot
129 
9 Thermal Storage During Off-Cycle Test Results 
Chapters ?7 and ?8 introduced transient performance results of the 
secondary loop system. Thermal lag during pull-down tests and NEDC drive 
cycle tests, due to an increased thermal mass, were characterized and discussed. 
While this can be a disadvantage for a fast cabin pull-down and consequently 
thermal comfort, it can be turned into an advantage during off-cycle operation. If 
the automobile idles and the compressor either runs with reduced rpm, or is 
turned off for short periods of time, cooling capacity is reduced. The thermal 
mass of the secondary loop system can preserve capacity and therefore thermal 
comfort during off-cycle periods.  
Chapter ?9 provides experimental results to quantify the effect of thermal 
storage of a secondary loop system during off-cycle periods. Two experimental 
test procedures were developed, namely the Start/Stop drive cycle and the off-
 cycle cabin warm-up after pull-down. The Start/Stop drive cycle is a modified 
version of the NEDC. The idling periods were replaced by off periods, simulating 
the start/stop operation of many modern cars. A further modification of the 
Start/Stop cycle is the Start/Stop 3x Idling drive cycle (SS3xI). All idling periods, 
or for that matter off-cycle periods, were tripled in length to determine how 
changes in off-cycle time affect the performance of secondary loop systems with 
respect to the performance of direct expansion systems. The second test 
procedure developed for thermal storage characterization is the cabin warm-up 
test. For this test, a regular pull-down test is performed. When the cabin 
temperature reaches are pre-determined comfort setpoint, the compressor is 
130 
turned off and the cabin is allowed to warm up. Cabin air and supply air 
temperature, as well as relative humidity, are measured during the warm-up 
period. ?Appendix E provides more detailed information about the thermal storage 
test procedures.  
In addition to the secondary loops thermal storage, ice storage was 
investigated as part of advanced thermal storage options. To this end, an ice 
storage heat exchanger (IS), introduced in Chapter ?3.3, was built and integrated 
into the secondary loop test facility. Chapter 3.3 also details the modifications 
made to the facility and resulting operating modes, such as IS charging, IS only 
cooling, and 2LP+IS cooling mode. Chapter ?9.2 shows experimental results for 
pull-down with ice storage, and thermal storage cabin warm-up tests. 
 
9.1 Secondary Loop Thermal Storage Performance 
Chapter ?9.1 provides experimental test results, which characterize the 
thermal storage effect of secondary loop systems. Start/Stop drive cycle results, 
as well as cabin warm-up after pull-down results are discussed.  Thermal storage 
tests were performed with 5.5 kg of ethylene glycol as coolant, as detailed in ?3.2. 
In a commercial vehicle, results may vary based on the amount of coolant. A 
comparison of thermal storage performance based on coolant mass fraction and 
based on secondary loop coolant volume is provided in Chapter ?11 as part of the 
transient simulation effort in Modelica.  
 
131 
9.1.1 Secondary Loop Start/Stop Drive Cycle Performance 
Start/Stop drive cycle tests are a modified version of the New European 
Drive Cycle (NEDC). Due to turning off the compressor, instead of running at 
idling speed, fluctuations in temperature and relative humidity during the city 
portion of the cycle are more pronounced. Over the course of the start/stop cycle, 
less energy is consumed compared to the NEDC, which was the reason why 
start/stop operation was introduced to commercial vehicles in the first place 
(though rather for fuel saving than for A/C energy saving). At the same time, less 
energy is available for cooling the cabin, due to the off-times of the compressor.  
Figure ?9.1 shows a comparison of cabin supply air temperature and 
relative humidity profiles for R134a DX, R152a 2LP, and R290 2LP. Similar to the 
NEDC results, the secondary loop lagged behind in cooling down the supply air 
during the first minutes. The capacity, and subsequently the supply temperature, 
of the R290 2LP system were determined by the reduced rpm, which was chosen 
in order to give a similar transient performance to R134a DX systems. Although 
the supply temperature of R290 was shown to be 1 K - 1.5 K higher compared to 
R152a, this can be adjusted by choosing an appropriate compressor 
displacement or compressor speed (for internal combustion engine vehicles: 
engine/compressor gear ratio). Supply air temperature of the secondary loop was 
up to 5K higher during the course of the drive cycle. However, the thermal mass 
of the secondary loop stabilized supply air temperature, as well as relative 
humidity in comparison to the direct expansion (DX) system. In Figure ?9.1 b), it 
can be observed that R134a DX supply relative humidity increased rapidly during 
132 
off-cycle periods, due to re-evaporation of condensate. At the end of the city 
portion of the cycle, supply air may be saturated or close to saturation, which 
might decrease the thermal comfort level of passengers and increase the threat 
of windshield fogging under certain operating conditions. Due to its thermal 
storage potential, the secondary loop could retain cooling capacity during off-
 cycle periods, which prevented excessive re-evaporation and stabilized supply 
relative humidity.  
Figure ?9.2 shows a comparison of accumulated performance results for 
the start/stop drive cycle with respect to R134a DX performance. Accumulated 
cooling capacity, as well as accumulated power consumption, and transient 
performance factor (TPF), are shown in Figure ?9.2 a). While secondary loop 
system results, R152a 2LP and R290 2LP, show a 8% decrease in energy 
available for cooling, cooling energy increased 2% to 3% compared to NEDC 
results, discussed in Chapter ?8.3.2. This is due to the secondary loop system 
retaining cooling capacity during off-cycle periods. Based on the energy input 
required for R152a 2LP and R290 2LP, TPF decreased by roughly 10% (R152a 
2LP), or was similar to R134a DX (R290 2LP). Although the secondary loop 
energy consumption was increased, due to the coolant pump operation during 
off-cycle periods, the cooling capacity advantage due to thermal storage resulted 
in a higher TPF, compared to NEDC operation. Figure ?9.2 b) shows accumulated 
sensible and latent performance during the start/stop cycle. Energy available for 
sensibly cooling the air was up to 10% reduced (R290 2LP) as compared to 
R134a DX, while latent energy available to dehumidify the air was reduced by up 
133 
to 30% (R152a 2LP). Measurement uncertainty for latent performance was 
significantly increased compared to uncertainty of sensible performance and may 
distort the results for latent cooling and dehumidification.  
 
Figure ?9.1: Start/Stop Drive Cycle Supply Air Profiles 
 
0
 5
 10
 15
 20
 25
 30
 35
 40
 0 5 10 15 20
 Te
 m
 pe
 ra
 tu
 re
  [
 ?C
 ]
 Time [min]
 S/S R134a DX S/S R152a 2LP S/S R290 2LP
 Start/Stop Supply Temperature
 20
 30
 40
 50
 60
 70
 80
 90
 100
 0 5 10 15 20
 R
 el
 at
 iv
 e 
H
 um
 id
 it
 y 
[%
 ]
 Time [min]
 S/S R134a DX S/S R152a 2LP S/S R290 2LP
 Start/Stop Supply Relative Humidity
 a)
 b)
134 
 
Figure ?9.2: Start/Stop Drive Cycle Accumulated Performance Results 
 
The start/stop drive cycle was modified by increasing the length of off-
 cycle periods. NEDC and start/stop drive cycles feature idling periods of roughly 
20 seconds length. For the Start/Stop 3x Idling (SS3xI) drive cycle, the off-cycle 
-40
 -30
 -20
 -10
 0
 10
 20
 30
 40
 Dehumid. Q sensible Q latent
 D
 eh
 um
 id
 . [
 %
 ] 
En
 er
 gy
  [%
 ]
 Performance Metrics
 R134a DX R152a 2LP R290 2LP
 Sensible/Latent Energy (S/S, %R134a BSL)
 -30
 -20
 -10
 0
 10
 20
 30
 Q (air) E in TPF (air)
 Re
 la
 ti
 ve
  E
 ne
 rg
 y;
  T
 PF
  [%
 ]
 Performance Metrics
 R134a DX R152a 2LP R290 2LP
 Acc. Performance Parameters (S/S, % R134a BSL)
 a)
 b)
135 
periods were tripled, featuring a length of roughly 60 seconds. This was felt to be 
still within a realistic boundary for traffic light stops during city driving.  
   Figure ?9.3 shows a comparison of transient supply temperature and 
relative humidity profiles for the SS3xI drive cycle operation. It can be observed 
that the fluctuations in supply temperature and relative humidity are more 
pronounced compared to regular start/stop operation. For a direct expansion 
system, increase in supply temperature during off-cycle periods could be as high 
as 10 K, while the secondary loop system was able to limit the fluctuations in 
supply temperature to about 2 K. The average increase of R152a 2LP or R290 
2LP supply temperature compared to R134a DX supply temperature was not as 
clear cut as observed in the NEDC or Start/Stop results. Figure ?9.3 b) shows 
R134a DX supply humidity to increase well beyond the saturation limit during the 
second half of the city portion of the cycle. Although there was a significant 
increase for R152a 2LP and R290 2LP supply humidity as well, relative humidity 
never reached 100% throughout the SS3xI drive cycle.  
A comparison of accumulated performance parameters for the Start/Stop 
3x Idling drive cycle, as well as accumulated sensible and latent performance, 
are shown in Figure ?9.4. As can be observed in Figure ?9.4 a), the energy 
available for cooling the air was on par with R134a DX for R290 2LP and was 
increased by about 4% for R152a 2LP. In comparison to NEDC and Start/Stop 
drive cycle, the long off-cycle periods of the SS3xI allowed the secondary loop to 
catch up in air-side cooling energy, despite the disadvantage of having to cool 
136 
down more thermal mass. Consequently, R290 2LP TPF increased by 10% 
above R134a DX, and R152a 2LP TPF was measured to be similar to R134a DX. 
 
Figure ?9.3: Start/Stop 3x Idling Drive Cycle Supply Air Profiles 
 
Sensible cooling performance for both secondary loop refrigerants, shown 
in Figure ?9.4 b), equaled R134a DX sensible cooling performance. Latent cooling 
0
 5
 10
 15
 20
 25
 30
 35
 40
 0 5 10 15 20 25 30
 Te
 m
 pe
 ra
 tu
 re
  [
 ?C
 ]
 Time [min]
 SS3xI R134a DX SS3xI R152a 2LP SS3xI R290 2LP
 SS3xI Supply Temperature
 20
 30
 40
 50
 60
 70
 80
 90
 100
 0 5 10 15 20 25 30
 R
 el
 at
 iv
 e 
H
 um
 id
 it
 y 
[%
 ]
 Time [min]
 SS3xI R134a DX SS3xI R152a 2LP SS3xI R290 2LP
 SS3xI Supply Relative Humidity
 a)
 b)
137 
was still reduced compared to the direct expansion system. R152a 2LP and 
R290 2LP latent cooling were reduced by about 40% and 25%, respectively. 
 
Figure ?9.4: Start/Stop 3x Idling Drive Cycle Accumulated Performance Results 
 
A comparison of the three cycles may show the difference between direct 
expansion and secondary loop systems more conclusively. Figure ?9.5 shows an 
-30
 -20
 -10
 0
 10
 20
 30
 Q (air) E in TPF (air)
 Re
 la
 ti
 ve
  E
 ne
 rg
 y;
  T
 PF
  [%
 ]
 Performance Metrics
 R134a DX R152a 2LP R290 2LP
 Acc. Performance Parameters (SS3xI, % R134a DX)
 -80
 -60
 -40
 -20
 0
 20
 40
 60
 80
 Dehumid. Q sensible Q latent
 D
 eh
 um
 id
 . [
 %
 ] 
En
 er
 gy
  [%
 ]
 Performance Metrics
 R134a DX R152a 2LP R290 2LP
 Sensible/Latent Energy (SS3xI, %R134a DX)
 a)
 b)
138 
accumulated performance comparison between all three systems for both, 
Start/Stop with respect to NEDC performance, and SS3xI with respect to NEDC 
performance. It can be observed that the R134a DX system "loses" 12% of air-
 side cooling energy when moving from regular NEDC operation to Start/Stop 
operation, while the secondary loop system accumulated cooling capacity 
decreases by 10%. The difference is well within measurement uncertainty. As 
off-cycle periods increase in length, as is the case for the SS3xI drive cycle, 
R134a DX "loses" nearly 40% of air-side cooling energy. At the same time, 
cooling energy of the secondary loop system decreased by only 33%. The 
reduction in energy consumption when switching from NEDC to Start/Stop 
operation is about 5% for both, direct expansion and secondary loop system. The 
reduction in energy for 2LP systems when switching to SS3xI is less compared to 
direct expansion system, due to the increased impact of pump work with 
increasing off-cycle periods. Nevertheless, decrease in TPF favors secondary 
loop systems, especially for longer off-cycle periods. It can be concluded that city 
driving with start/stop operation favors secondary loop systems with increasing 
off-cycle periods in terms of energy consumption and thermal comfort. 
 
139 
 
Figure ?9.5: Drive Cycle Accumulated Performance Results Comparison 
 
9.1.2 Secondary Loop Off-Cycle Cabin Warm-up Performance 
Cabin warm-up tests evaluate the development of cabin temperature and 
relative humidity during extended off-cycle periods. To this end, the cabin was 
pulled down to comfort condition, at which point the compressor was turned off 
and the cabin was allowed to warm back up. The airflow over the evaporator (DX) 
or cooler (2LP) was kept constant and the coolant pump was kept running for the 
secondary loop system. 
Figure ?9.6 shows a comparison of cabin and supply temperatures during 
an extended off-cycle period. Trends for the increase of cabin temperature are 
shown in Figure ?9.6 a). While the cabin temperature increased immediately for 
the direct expansion system, 2LP cabin temperature stayed at comfort setting 
(24?C) for a period of about 2 - 2.5 minutes. After the first minute of compressor 
-40
 -35
 -30
 -25
 -20
 -15
 -10
 -5
 0
 Q (air) E in TPF (air)
 En
 erg
 y [%
 ] ; T
 PF 
[%
 ]
 Performance Metrics
 R134a DX (S/S) R152a 2LP (S/S) R290 2LP (S/S)
 R134a DX (SS3xI) R152a 2LP (SS3xI) R290 2LP (SS3xI)
 Drive Cycle Comparison (% NEDC Performance)
140 
shut down, the difference in cabin temperature between the direct expansion 
system and the secondary loop system was about 1 K. Further along the cabin 
warm-up, the temperature difference increased to about 2.5 K and 3 K at 5 
minutes and 15 minutes, respectively. The temperature difference between direct 
expansion and secondary loop system seemed to stay fairly constant thereafter 
for an extended period of  time. The difference in cabin temperature can be 
traced back to the difference in cabin supply temperature, shown in Figure ?9.6 b). 
While the cabin supply temperature started to increase immediately after 
compressor shut down, due to loss of cooling capacity, the rate of increase was 
far higher for the direct expansion system. This was especially true for the first 
one to two minutes of the cabin warm-up process. The steep increase in supply 
temperature might partially be a result of condensation of refrigerant vapor inside 
the evaporator. While condensation might happen at the intermediate heat 
exchanger of the secondary loop system as well, the thermal mass of the coolant 
is able to compensate. During the initial minutes of the warm-up process, 2LP 
supply temperature was about 7 K lower compared to R134a DX supply 
temperature. During later stages of the warm-up process the temperature 
difference decreased to roughly 5 K.    
Another aspect of extended off-cycle periods is the re-evaporation of 
condensate off the outer surface of the evaporator (DX system) or cooler (2LP 
system). Figure ?9.7 shows supply relative humidity profiles, as well as 
dehumidification at the heat exchanger. DX supply relative humidity rose fast and 
reached saturated conditions after about 30 seconds, while it took about 2 
141 
minutes for 2LP supply relative humidity to achieve saturation. It was observed 
that direct expansion re-evaporation humidified the air with approximately 5 gram 
of water per kg of moist air within the first three minutes of the cabin warm-up 
process. 
 
Figure ?9.6: Secondary Loop Off-cycle Cabin Warm-up Profiles 
 
0
 5
 10
 15
 20
 25
 30
 35
 0 5 10 15 20
 Te
 m
 pe
 ra
 tu
 re
  [?
 C]
 Time [min]
 R134a DX R152a 2LP R290 2LP
 Supply Temperature in Off-Cycle Mode (T30, hwy)
 23
 4
 25
 26
 27
 28
 29
 30
 31
 32
 33
 0 5 10 15 20
 Te
 m
 pe
 ra
 tu
 re
  [?
 C]
 Time [min]
 R134a DX R152a 2LP R290 2LP
 Cabin Heat Up in Off-Cycle Mode (T30, hwy)
 a)
 b)
142 
Secondary loop thermal storage decelerated this process. A rate of 4 
gram of water per kg of moist air was reached after about 3 minutes. Starting 
from 6 minutes, the humidification for both the DX and 2LP system was reduced 
to about 2 g/kg and reduced slowly thereafter. The cabin was humidified at a 
slow rate for an extend period of time, mostly through the remaining condensate 
in the drain bin and drain pipe of the facility. Humidification during off-cycle 
periods should be avoided as much as possible, since the moisture results in an 
increased latent load when the compressor turns back on. At the same time, 
excessive humidity combined with rising temperatures might lead to reduced 
passenger comfort. The secondary loop system has a clear advantage compared 
to direct expansion systems in terms of decelerating re-evaporation and cabin 
humidification. However, even when employing a secondary loop system, off-
 cycle periods longer than a few minutes will lead to undesirable humidification.  
143 
 
Figure ?9.7: Secondary Loop Off-cycle Relative Humidity Profiles 
 
9.2 Ice Storage Thermal Storage Performance 
Advanced thermal storage options, such as phase change materials 
(PCMs), can be used to extend passenger comfort, as well as partially prevent 
excessive humidification of the cabin air during off-cycle periods. Depending on 
charging and releasing strategies, phase change materials might also assist the 
-6
 -4
 -2
 0
 2
 4
 6
 0 2 4 6 8 10
 De
 hu
 m
 id
 ifi
 ca
 tio
 n 
[g
 /kg
 ]
 Time [min]
 R134a DX R152a 2LP R290 2LP
 Dehumidification in Off-Cycle Mode (T30, hwy)
 60
 65
 70
 75
 80
 85
 90
 95
 100
 0.0 0.5 1.0 1.5 2.0 2.5 3.0
 Re
 la
 tiv
 e H
 um
 id
 ity
  [%
 ]
 Time [min]
 R134a DX R152a 2LP R290 2LP
 Supply Relative Humidity in Off-Cycle Mode (T30, hwy)
 a)
 b)
144 
vapor compression cycle during cabin pull-down, essentially reducing time to 
comfort. One of the least expensive and widely available phase change materials, 
operating within the acceptable temperature range, is water, H2O. While water in 
frozen form can assist the air-conditioning system during the cooling period in the 
summer, it may also assist as a low temperature reference point for automotive 
heat pump systems in the winter to increase cycle efficiency when outdoor 
temperatures are well below 0?C. In the present research, ice storage was used 
for cooling purposes as part of a secondary loop automotive air-conditioning 
system. A custom ice storage box, introduced in Chapter ?3.3, was built and 
integrated into the coolant loop. In the present research it was assumed that the 
ice storage box would be used in an electric vehicle, which allows the charging of 
ice at reduced electricity rates during the night. The ice would subsequently be 
used during the day to decrease A/C power consumption and reduce the effect of 
air-conditioning on the range of electric vehicles. Several options of how to 
charge the ice box exist. In an actual vehicle, ice could be produced by a 
dedicated facility outside of the vehicle, or while the ice storage box is installed in 
the vehicle. In the present research the ice was produced by the R290 2LP 
system, while the storage box was installed in the secondary loop. 
Studies on driving patterns show that an average commute in the USA 
lasts about 25 minutes, while 80% of commutes fall within 34 minutes of travel 
time [120]. In European countries this might reduce significantly. In France, an 
average commute lasts less than 15 minutes, while about 95% of commutes fall 
within a 30 minute time frame [121]. Since a usual pull-down from soaked cabin 
145 
condition to comfort temperature lasts about 15 minutes, even at highway speed, 
a large part of the commute will be spent at an uncomfortably high cabin 
temperature. The ice storage box was built with experimental pull-down results 
from previous R134a DX pull-down tests in mind. The energy consumption of the 
A/C system during the first 20 minutes of a pull-down was measured to be  
between 4,500 and 5,000 kJ. The mass of ice in the ice storage box was chosen 
based on the assumption that the ice storage box might be able to either cool the 
cabin without the vapor compression cycle (VCC), or at least help increase 
thermal comfort and reduce VCC power consumption significantly during pull-
 down. Consequently, a mass of 15 kg of ice was chosen, since this covered the 
energy needed for cooling during a 20 to 25 minute commute, as seen in 
Figure ?9.8.  
 
Figure ?9.8: Enthalpy of Fusion Based on Mass of Ice 
 
146 
While adding such a large mass to a vehicle will undoubtedly increase 
vehicle fuel consumption, the overall impact on fuel consumption might not be so 
clear cut when taking fuel consumption due to A/C power consumption into 
account. Reynolds [122] showed that the change in fuel consumption per 100kg 
for internal combustion engine vehicles (ICEVs) is about 0.7 l/100km, while it is 
0.4 l/100km for hybrid electric vehicles (HEVs). For the 15 kg ice storage box, 
this results in an increase in fuel consumption of roughly 0.1 l/100km for ICEVs 
and 0.06 l/100km for HEVs. For a ICEV which could drive 15 km per liter of fuel, 
this would result in a reduction of 0.22 km/l. At the same time, Clodic [121] and 
Johnson [123] determined that an A/C system typically uses 2,000 W to 3,000 W 
during steady-state cooling operation (peak power can be double). Farrington 
[124] shows that each vehicle experiences a fuel consumption penalty of roughly 
1 less mile per gallon per 400 W auxiliary load. For 2,400 W A/C load, this results 
in a 6 mpg fuel consumption penalty (2.55 km/l). If the ice storage box can cover 
only a third of the A/C load and gets used roughly a third of the year, it will have 
more than made up for the additional fuel consumption due to its mass, not 
speaking of the benefit of enhanced passenger comfort. 
 
9.2.1 Ice Storage Charging Performance 
Figure ?9.9 details the performance of the vapor compression system when 
charging the ice storage box. To form ice, the cooler of the secondary loop was 
bypassed and the vapor compression cycle cooled part of the secondary loop to 
sub zero temperatures. The water-glycol mixture was routed through the ice 
147 
storage heat exchanger charge the ice storage box. Figure ?9.9 a) shows 
charging time with respect to initial average temperature in the ice storage box. 
Six thermocouples were placed in the storage box to evaluate average 
temperature during the charging and the cooling processes. It can be observed 
that charging time for the ice storage box was about 60 minutes if the initial 
average temperature was above 10?C. Sensibly cooling the ice took a small 
portion of the overall time, so that charging time did not increase significantly at 
higher initial temperatures. At initial temperatures close to 0?C, the storage box 
may still hold ice, while the mass fraction of liquid water to frozen ice cannot be 
determined with certainty. Thus any correlation of charging time over initial 
temperature must be very weak at low initial temperatures. Figure ?9.9 b) shows 
energy used for charging the ice storage box with respect to initial average 
storage box temperature. It was observed that about 1,500 Wh were needed to 
freeze the 15 L of water and subcool the ice to -6?C average storage box 
temperature. Transient performance factor, shown in Figure ?9.9 c) did not 
change significantly with initial average temperature. On average, ice charging 
TPF was between 0.55 and 0.7. Ice charging TPF and charging time may be 
improved by optimizing the ice storage heat exchanger and designing the VCC 
for the conditions which are prevalent during the charging process.  
148 
 
Figure ?9.9: Ice Storage Charging Performance 
0
 10
 20
 30
 40
 50
 60
 70
 80
 0 5 10 15 20 25 30
 C
 h
 a
 rg
 in
 g
  T
 im
 e
  [
 m
 in
 ]
 Initial Temperature [?C]
 IS Charging Time vs. Initial Avg. Temperature
 0
 200
 400
 600
 800
 1000
 1200
 1400
 1600
 1800
 2000
 0 5 10 15 20 25 30
 En
 er
 gy
  In
 p
 u
 t [
 W
 h
 ]
 Initial Temperature [?C]
 IS Charging Energy vs. Initial Avg. Temperature
 0.0
 0.1
 0.2
 0.3
 0.4
 0.5
 0.6
 0.7
 0.8
 0.9
 1.0
 0 5 10 15 20 25 30
 C
 h
 ar
 gi
 n
 g 
TP
 F 
[-
 ]
 Initial Temperature [?C]
 IS Charging TPF vs. Initial Avg. Temperature
 b)
 c)
149 
9.2.2 Ice Storage Pull-down Performance 
Chapter ?9.2.2 ?9.2.2provides an example of using ice storage during cabin 
pull-down to assist the secondary loop system. In this operation mode, the water-
 glycol mixture is cooled by the VCC in the intermediate HX, subsequently routed 
through the ice storage box to be cooled down further, and finally cools down the 
cabin supply air at the cooler.  
Figure ?9.10 through Figure ?9.12 show a performance comparison of a 
cabin pull-down test with the R290 2LP system, labeled 2LP, and a pull-down 
test assisted by the ice storage box, labeled 2LP+IS. The test detailed in this 
chapter is a pull-down test at highway driving condition and ambient conditions of 
30?C/50% relative humidity. At the beginning of the test, the cabin was soaked to 
42?C, which is indicated by the label (S).  
Figure ?9.10 shows the cabin and supply temperature profiles for the pull-
 down comparison between 2LP and 2LP+IS system operation. It was observed 
that ice storage has the greatest effect during the first minutes of the pull-down 
test. Right from the start, the supply air temperature of the 2LP+IS system was 
significantly lower compared to the 2LP system supply air temperature. After 
about 8 minutes, the supply air temperatures of both systems approached a 
similar value. This is reflected in the cabin temperature, which was decreased by 
about 1.5 K in the initial minute and remained 2 K below 2LP cabin temperature 
throughout the remainder of the test. As a result, more than 2 minutes could be 
shaved off time to comfort by using the ice storage system.  
150 
 
Figure ?9.10: Ice Storage Pull-down Cabin Air Profiles 
 
Figure ?9.11 shows a comparison of transient profiles of cooling capacity 
and power consumption during a pull-down test. For the 2LP system, air-side 
cooling capacity is shown in Figure ?9.11 a), while for the 2LP+IS system total air-
 side cooling capacity, as well as ice storage heat exchanger cooling capacity are 
24
 26
 28
 30
 32
 34
 36
 38
 40
 42
 44
 0 2 4 6 8 10 12 14 16
 Te
 m
 p
 e
 ra
 tu
 re
  [
 ?C
 ]
 Time [min]
 2LP (S) 2LP+IS (S)
 Cabin Temperature Profile (T30, hwy)
 0
 5
 10
 15
 20
 25
 30
 35
 4
 45
 50
 55
 0 2 4 6 8 10 12 14 16
 R
 e
 la
 ti
 ve
  H
 u
 m
 id
 it
 y 
[%
 ]
 Time [min]
 2LP (S) 2LP+IS (S)
 Cabin Relative Humidity Profile (T30, hwy)
 a)
 b)
151 
shown. The trends for cooling capacity support the conclusions drawn from the 
supply air and cabin air temperature profiles. In the initial minutes, the ice storage 
heat exchanger added additional cooling capacity to the total capacity available 
on the air-side. However, after about 4 minutes the cooling capacities for 2LP 
and 2LP+IS became identical and 2LP+IS capacity continued to decrease below 
the cooling capacity of the 2LP only system. On one hand this was due to a 
faster decreasing temperature difference for the 2LP+IS system. On the other 
hand, it was also due to the evaporation temperature reaching low enough 
values, resulting in the ice storage system hindering a further rapid pull-down 
(due to the water/ice mixture remaining at 0?C). This was observed at an 
advanced stage after 10 minutes, when the value for ice storage cooling capacity 
becomes negative, indicating that the ice storage box was charged by the 2LP 
cycle, instead of assisting in cooling cabin supply air. Figure ?9.11 b) shows a 
comparison of power consumption profiles for both 2LP and 2LP+IS. While 
power consumption was highest for the 2LP only system during the initial 
minutes of the test, when the compressor and the pump were started, power 
consumption of the 2LP+IS system was lowest during the start-up phase. After 
about 4 minutes, the trends of power consumption of both systems were the 
same, with the power consumption of the 2LP+IS system being slightly reduced 
by about 100 W. Overall, the ice storage system allowed for energy saving during 
cabin pull-down when assisting a vapor compression cycle, but energy savings 
did not add up to a significant amount.  
 
152 
 
Figure ?9.11: Ice Storage Pull-down Transient Performance Profiles 
 
Figure ?9.12 shows the accumulated performance results of the 2LP+IS 
system with respect to the performance of the 2LP system. It can be observed 
that time to comfort was decreased by about 16%. Consequently, energy 
-2000
 0
 2000
 4000
 6000
 8000
 10000
 0 2 4 6 8 10 12 14 16
 C
 o
 o
 li
 n
 g 
C
 ap
 ac
 it
 y 
[W
 ]
 Time [min]
 Q_air 2LP (S) Q_air 2LP+IS (S) Q_ICE 2LP+IS (S)
 Pull-down Cooling Capacity Profile (T30, hwy)
 0
 1000
 2000
 3000
 4000
 5000
 6000
 0 2 4 6 8 10 12 14 16
 P
 o
 w
 e
 r 
C
 o
 n
 su
 m
 p
 ti
 o
 n
  [
 W
 ]
 Time [min]
 P_tot 2LP (S) P_tot 2LP+IS (S)
 Pull-down Power Consumption Profile (T30, hwy)
 a)
 b)
153 
available to cool cabin supply air was reduced by about 10%, due to the reduced 
testing time. The 2LP+IS system was able to save about 21% of power 
consumption with respect to the power consumption of the 2LP system. The 
overall TPF increased by 15%.  
 
Figure ?9.12: Ice Storage Pull-down Accumulated Performance Results 
 
9.2.3 Ice Storage Off-cycle Cabin Warm-up Performance 
Chapter ?9.2.3 ?9.2.3discusses thermal storage performance of the ice 
storage system during a cabin warm-up test. Temperature, as well as relative 
humidity results for the R290 2LP+IS system are compared to R290 2LP results 
and R134a DX results. The test was performed at highway driving speed and 
ambient conditions of 30?C/50% relative humidity. The cabin was soaked to 42?C 
-25
 -20
 -15
 -10
 -5
 0
 5
 10
 15
 20
 25
 Time Q_air E tot TPF
 Tim
 e [
 %]
 , En
 erg
 y [
 %]
 , TP
 F [
 %]
 Performance Metrics 
2LP (S) 2LP+IS (S)
 Acc. Pull-down Performance Parameters (% 2LP)
154 
at the start of the pull-down, which preceded the cabin warm-up portion of the 
test.  
Figure ?9.13 shows cabin, as well as cabin supply temperature trends 
during the cabin warm-up test. Throughout the pull-down procedure enough ice 
remained frozen in the ice storage box to provide cooling capacity to keep the 
cabin temperature at comfort set point or slightly lower for an extended amount of 
time. Cabin temperature increased to 28?C and 31.5?C after 15 minutes for R290 
2LP and R134a 2LP, respectively. At the same time, R290 2LP+IS cabin 
temperature remained below 23.5?C at 15 minutes into the test. This was 
reflected in the cabin supply temperature profile, which did not increase above 
15?C 15 minutes into the test, remaining roughly 10 K and 15 K below R290 2LP 
and R134 DX supply temperature, respectively.  
The differences in thermal storage performance are illustrated by 
Figure ?9.14, which shows a comparison of accumulated sensible capacity over 
the duration of a cabin warm-up test. Energy available for cooling was the same 
for the R134a DX, the R290 2LP, and the R290 2LP+IS system within the first 30 
seconds. After the first half minute the capacity of the direct expansion system 
started to decrease significantly compared to secondary loop system capacity, 
resulting in a decrease of slope of the R134a DX cooling energy, as shown in 
Figure ?9.14. For the first three minutes the ice storage system (2LP+IS) showed 
no advantage in comparison to the regular secondary loop system. However, 
after three minutes the secondary loop system cooling capacity decreased, while 
the cooling capacity of the ice storage system stayed constant (resulting in a 
155 
steady increase of cooling energy with time). Consequently, the ice storage 
system had a significant advantage over the unmodified secondary loop systems 
in terms of thermal comfort for extended off-cycle periods.  
Figure ?9.13 shows supply relative humidity trends, as well as 
dehumidification during off-cycle cabin warm-up.  
 
Figure ?9.13: Ice Storage Off-cycle Cabin Warm-up Temperature Profiles 
0
 5
 10
 15
 20
 25
 30
 35
 0 5 10 15 20
 Te
 m
 pe
 ra
 tu
 re
  [?
 C]
 Time [min]
 R134a DX R290 2LP R290 2LP+IS
 Supply Temperature in Off-Cycle Mode (T30 hwy, soak)
 22
 24
 26
 28
 30
 32
 34
 0 5 10 15 20
 Te
 m
 pe
 ra
 tu
 re
  [?
 C]
 Time [min]
 R134a DX R290 2LP R290 2LP+IS
 Cabin Heat Up in Off-Cycle Mode (T30 hwy, soak)
 a)
 b)
156 
It was observed that the ice storage system retained enough latent 
capacity to reduce re-evaporation at the cooler to a degree which allowed the 
supply air to stay below 95% relative humidity for an extended period of time. 
This was supported by the dehumidification profile, which showed a severe 
decrease in humidification of cabin supply air during the first 7 minutes of the 
cabin warm-up test, compared to R290 2LP and R134a DX. The secondary loop, 
with the assistance of the ice storage box, was able to preserve passenger 
thermal comfort, as well as prevent excessive cabin air humidification for a period 
longer than 20 minutes after compressor shut down.  
 
Figure ?9.14: Cabin Warm-up Accumulated Sensible Capacity Comparison 
0
 0.1
 0.2
 0.3
 0.4
 0.5
 0.6
 0 2 4 6 8 10 12 14 16 18 20
 Ene
 rgy
  [kW
 h]
 Time [min]
 R134a DX R290 2LP R290 2LP+IS
 Accumulated Sensible Capacity (Cabin Warm-up)
157 
 
Figure ?9.15: Ice Storage Off-cycle Cabin Warm-up Humidity Profiles 
 
9.3 Summary: Thermal Storage in Secondary Loop Systems 
? Several test procedures were designed to test the effect of thermal 
storage with respect to extension of thermal comfort and transient 
performance improvement during off-cycle operation. 
60
 65
 70
 75
 80
 85
 90
 95
 100
 0 5 10 15 20
 Re
 la
 tiv
 e 
H
 um
 id
 ity
  [%
 ]
 Time [min]
 R134a DX R152a 2LP R290 2LP+IS
 Supply Relative Humidity in Off-Cycle Mode (T30 hwy)
 -6
 -4
 -2
 0
 2
 4
 6
 0 5 10 15 20
 D
 eh
 um
 id
 ifi
 ca
 tio
 n 
[g
 /kg
 ]
 Time [min]
 R134a DX R290 2LP R290 2LP+IS
 Dehumidification in Off-Cycle Mode (T30 hwy)
 a)
 b)
158 
? During Start/Stop operation, a regular direct expansion system 
experiences large fluctuations in cabin supply temperature and relative 
humidity. Fluctuations in humidity are mostly due to re-evaporation of 
condensate from the evaporator (DX) or cooler (2LP) surface, once the 
compressor is turned off. The thermal mass of the secondary loop 
decreases fluctuations in supply temperature and humidity and 
therefore improves stability of supply and cabin conditions, which can 
be related to an increase in thermal comfort. 
? During Start/Stop operation, the energy available for cooling the supply 
air was reduced due to the larger thermal mass of the secondary loop. 
However, cooling energy was reduced by less than 10% in comparison 
to R134a DX, while R290 2LP TPF was equal to R134a TPF. 
? When the Start/Stop drive cycle was modified by increasing off-cycle 
periods by three times, the secondary loop system outperformed the 
direct expansion system in terms of dehumidification, as well as 
efficiency. With increasing idling (or off-cycle) periods, the direct 
expansion system was prone to re-humidifying the cabin while the 
compressor was turned off. The secondary loop system prevented 
excessive cabin humidification and kept cabin temperature stable 
during off-cycle periods. The energy available to cool the cabin supply 
air was the same compared to the R134a direct expansion system. 
Therefore, R290 2LP TPF increased to 9% above R134a DX TPF. The 
sensible energy available was the same compared to the R134a direct 
expansion system, while the latent energy was smaller, due to the 
increased dehumidification needed by the direct expansion system. 
When switching from a car without start/stop to a car with start/stop 
automatic operation, less cooling energy will be available over the 
course of the drive cycle. Due to thermal storage, secondary loop 
systems face less of a penalty compared to direct expansion systems.  
159 
?  If the compressor is turned off after pull down to comfort temperature, 
the secondary loop reduces the heating up of the cabin. During an 
extended time of 20 minutes, secondary loop cabin temperature was 
consistently 2.5 to 3 K below direct expansion cabin temperature 
during cabin warm-up. At the same time, the secondary loop could not 
prevent excessive humidification of the cabin after the compressor was 
turned off, but it was able to delay it by about 1.5 minutes longer than 
the direct expansion system. Taking cabin humidification into 
consideration, the secondary loop can be useful for short off-cycle 
periods of a few minutes.  
?  An ice storage box with an ice to water-glycol heat exchanger was 
designed and built. 15 L of water were used in the ice storage box, 
which resulted in a charging time of about one hour and a charging 
energy of about 1.5 kWh. Since the regular vapor compression system 
was used for charging, the unusually low temperature conditions 
resulted in an average TPF of 0.65 during charging. 
? Ice storage assisted in pulling down the cabin to comfort temperature 
by decreasing power consumption and decreasing time to comfort. 
However, ice storage was only advantageous during the first minutes 
of the pull-down test. With increasing time the ice storage system 
becomes less profitable, as evaporating temperature decreases and 
the ice storage system inhibits higher cooling capacity. TPF was 
increased by 15%. The biggest advantage of the ice storage system 
comes to bear during off-cycle cabin warm-up, where the ice storage 
system could keep the cabin at comfort temperature for periods of 20 
minutes, while preventing excessive cabin humidification.  
160 
10 Experimental Results - A/C Controls and Cabin 
Preconditioning 
Chapter ?10 provides experimental results of research done in the area of 
air-conditioning systems control and cabin preconditioning in combination with 
secondary loop systems and ice storage. In previous chapters, pull-down tests 
were used to compare time to comfort and cooling energy available on the air-
 side. The present chapter focuses on A/C energy consumption during tests of 60 
minutes, including the initial pull-down. The A/C control tests strive to determine 
energy savings when different control strategies for the compressor and coolant 
pump are used. Cabin preconditioning tests focus on energy savings which may 
be achieved when the cabin is prevented from soaking by pre-conditioning cabin 
air to either ambient temperature or comfort temperature. While the results of A/C 
control experiments and cabin preconditioning are universal to all types of cars, 
they have their biggest impact in electric vehicles (EVs). On the one hand, cabin 
preconditioning may be more conveniently realized in EVs, while some of the 
discussed compressor control strategies are applicable only to electric 
compressors. On the other hand, results for cabin preconditioning and 
compressor control strategies explicitly focus on reducing A/C power 
consumption during vehicle operation and therefore have an immediate effect on 
driving range of EVs. It will be discussed how cabin preconditioning and thermal 
storage (here ice storage) can work together with alternative control strategies to 
reduce overall A/C power consumption, including heater (or PTC element) power 
consumption. 
161 
Reducing A/C energy consumption is especially important for electric 
vehicles, as shown by Farrington and Rugh [124]. While an internal combustion 
vehicle ICEV) may have a fuel penalty of 5 mpg for 2,000 W auxiliary load -a very 
conservative estimate for A/C systems-, fuel consumption of an advanced 
electric hybrid vehicle could see a penalty of as much as 34 mpg and range 
could reduce by 27%. This analysis did not include the PTC heater or other 
heating device for dehumidifying and series reheating the air. As Clodic [121] 
showed, a PTC heater alone can reduce EV range to below 90% at 35?C 
ambient condition. At high ambient conditions of 45?C driving range could reduce 
to roughly 75%. Umezu and Noyama [125] gave some information on the air-
 conditioning system and controls for the electric vehicle Mitsubishi i-MiEV in 2010. 
The i-MiEV uses a dedicated coolant cycle with a 5.0 kW PTC heating element to 
reheat the air after the evaporator. With A/C and heater on, A/C power 
consumption may increase by 30% to 40%, while cruising range may decrease 
from 160 km to 90 km (A/C and heater on, 35?C ambient temperature). The next 
important steps for Mitsubishi at the time were to look into cabin preconditioning 
and improving PTC control to reduce PTC energy usage.  
While the above paragraph clearly stated the importance of reducing 
energy consumption for EVs, the question remains how to do so. This chapter 
introduces two possibilities: alternative A/C compressor and heater controls, as 
well as cabin preconditioning. Both solutions have been evaluated under different 
circumstances by other researchers before. Forrest and Bhatti [126] looked into 
control of the recirculation ratio (or mixing ratio with fresh air), as well as series 
162 
reheat reduction. Instead of cycling the compressor to prevent frost at the 
evaporator, an externally controlled variable displacement compressor was used 
to decrease capacity when it was not needed. As a result, cabin relative humidity 
increased from about 30% steady-state value at comfort condition to 45% to 50%. 
It was estimated that the reheat reduction potential could reduce 28% of cooling 
capacity, which would not have to be delivered by the compressor. Experimental 
tests with the externally controlled compressor and series reheat reduction found 
a 60% (city) and 52% (highway) A/C fuel consumption reduction. However, the 
authors stated that due to the complexity of the test vehicles, tests were difficult 
to reproduce and there were high uncertainties due to an abundance in variables 
for vehicle conditioning. Furthermore, the study was focused on the reduction of 
ICEV fuel consumption, which does not translate well into energy savings an 
range preservation for EVs. Therefore, the current research focuses on a 
comprehensive investigation of energy saving potential due to alternative 
compressor and heater controls in a laboratory environment with high accuracy 
measurement capabilities.  
Cabin preconditioning was investigated by Roy et al. in 2003 and 2005. In 
a simulation study [127] and an experimental study [128] the influence of cabin 
preconditioning strategies on the soaking profile and cabin temperature profile 
during cool down was evaluated. The study focused on comparing different pre-
 conditioning strategies, such as forced ventilation, vacuum panels, or special 
glazing for windows. It was found that combinations of various technologies could 
reduce soak and achieve energy savings of about 15% to 20%. The study did not 
163 
elaborate on how the energy savings were determined. Instead of focusing on 
temperature profiles and technologies to achieve preconditioning, the current 
research focuses on energy savings achievable during driving, given a certain 
preconditioning. Experimental results are focused on a combination of cabin 
preconditioning and alternative A/C controls to gain significant reductions in 
compressor and heater energy consumption for EVs and providing insight into 
the circumstances which make cabin preconditioning beneficial.  
 
10.1 A/C Control Performance and Benefits 
The A/C controls discussed in this chapter focus on the reduction of 
compressor power consumption and the reduction of re-heating during extended 
driving. The different control strategies are based on the assumption that a fixed 
displacement compressor, such as the compressor used in this research, can 
either be cycled, or can be speed-controlled externally. The latter assumption is 
true only for compressors which are not belt-driven by an internal combustion 
engine, and therefore is geared towards hybrid electric vehicles and full electric 
vehicles.  
For each control test the cabin is initially pulled down to comfort 
temperature and the different control strategies evolve around keeping cabin 
temperature at comfort set point during an extended amount of time. Traditionally, 
this was done by leaving the compressor turned on, while the air-side evaporator 
outlet temperature decreased below the temperature needed to preserve thermal 
164 
comfort. Consequently, air leaving the evaporator was reheated. The compressor 
is cycled off only to prevent frost at the evaporator. Hence, frost cycling (F cyc) 
tests serve as a reference point for the alternative control strategies discussed in 
this chapter, namely the relative humidity cycling (RH cyc) and relative humidity 
continuous control (RH ctn) tests. A more detailed introduction to the test 
procedures is given in ?Appendix E. 
Figure ?10.1 shows a comparison of power consumption, cabin 
temperature, and cabin supply relative humidity for the R290 2LP system, 
utilizing the three above mentioned control strategies. Tests were conducted with 
the R290 2LP system at ambient conditions of 30?C ambient temperature and 50% 
relative humidity at highway driving conditions. Power consumption profiles for F 
cyc, RH cyc, and RH ctn control are shown in Figure ?10.1 a). It was observed 
that F cyc had the overall highest power consumption. Once cabin temperature 
reached the comfort set point, the compressor remained running at full speed 
and soon thereafter heater power had to be used to keep the cabin at thermal 
comfort. Several minutes thereafter, the compressor started to cycle to prevent 
frost at the evaporator. The cycles occurred in intervals of 3-4 minutes, with the 
compressor on time being significantly longer than the compressor off times. RH 
cyc power consumption was significantly reduced compared to F cyc power 
consumption. The compressor cycled faster, since it controlled cabin/supply 
humidity, rather than preventing frost. Compressor off times were longer than 
compressor on times. Additionally, heater power consumption was significantly 
reduced compared to F cyc control. RH ctn control had the overall lowest power 
165 
consumption. Once thermal comfort temperature was achieved, the compressor 
speed, as well as the coolant pump speed were controlled by a PID controller. 
The compressor and the pump kept running continuously, though the PID 
controller reduced compressor speed significantly. Figure ?10.1 b) shows the 
cabin temperature for the three control tests. It was observed that all three 
strategies were able to keep cabin temperature within a very narrow range (< 0.3 
K) of comfort set point. F cyc supply relative humidity, shown in Figure ?10.1 c), 
averaged at about 50% after thermal comfort was achieved. RH cyc, as well as 
RH ctn supply relative humidties were measured to be about 80% on average. At 
the same time, cabin relative humidty averaged at about 50%. While RH ctn 
supply relative humidity was more stable, compressor cycling during the RH cyc 
test introduced large fluctuations (< ?20%).  
Figure ?10.2 shows a comparison of accumulated performance results for 
three different control strategies. While Figure ?10.2 a) compares results for R290 
2LP only, Figure ?10.2 b) compares accumulated performance results of the ice 
storage system. For 2LP only, A/C power consumption could be reduced about 
40% and 60% for RH cyc and RH ctn control, respectively. This includes 
compressor, pump, and heater power (E in C+P+H). When including only 
compressor and coolant pump power (E in C+P), power consumption was 
reduced by about 30% and 55% for RH cyc and RH ctn control strategies, 
respectively. Due to reduced capacity delivered by the compressor and a 
reduced need for reheating, the energy available to cool the cabin supply air 
decreased by about 35%. This resulted in an overall increase in TPF of about 55% 
166 
for RH ctn, while RH cyc TPF decreased by about 50%. When RH cyc and RH 
ctn strategies using the ice storage system (2LP+IS) were compared with frost 
cycling using only the secondary loop system (2LP), power consumption was 
reduced significantly only for the RH cyc control strategy. Through the use of ice 
storage, power consumption during RH cyc control decreased to a similar level 
as power consumption druing RH ctn control (55%). 
167 
 
Figure ?10.1: A/C Controls Power Consumption and Cabin Air Profiles Comparison 
0
 500
 1000
 1500
 2000
 2500
 3000
 3500
 4000
 4500
 5000
 0 10 20 30 40 50 60
 P
 o
 w
 e
 r
  C
 o
 n
 s
 u
 m
 p
 t
 io
 n
  [
 W
 ]
 Time [min]
 F cyc (2LP) RH cyc (2LP) RH ctn (2LP)
 Total Power Consumption (2LP, soaked)
 20
 25
 30
 35
 40
 45
 0 10 20 30 40 50 60
 T
 e
 m
 p
 e
 r
 a
 t
 u
 r
 e
  [
 ?C
 ]
 Time [min]
 F cyc (2LP) RH cyc (2LP) RH ctn (2LP)
 Cabin Temperature (2LP, soaked)
 0
 10
 20
 30
 40
 50
 60
 70
 80
 90
 100
 0 10 20 30 40 50 60
 R
 e
 la
 t
 iv
 e
  H
 u
 m
 id
 it
 y
  [
 %
 ]
 Time [min]
 F cyc (2LP) RH cyc (2LP) RH ctn (2LP)
 Supply Relative Humidity (2LP, soaked)
 a)
 b)
 c)
168 
Power consumption of the RH ctn control strategy did not benefit from the 
ice storage system. At the same time, ice storage adds cooling capacity, which 
improves TPF for both, RH cyc and RH ctn control strategies. It can be 
concluded that RH ctn control strategy has the potential to reduce A/C energy 
consumption by as much as 60% at the given ambient and driving conditions. RH 
cyc control was able to reduce power consumption by about 40%. However, the 
increased start up currents for electric compressors when cycling were not taken 
into account, since the compressor used in the present research was a belt drive 
compressor. Therefore, results for compressor power consumption may vary for 
RH cyc operation in electric vehicles. Using thermal storage systems can benefit 
cyclic control, while it is of little benefit for the power consumption of a continuous, 
reduced speed compressor control. In the extreme case, there would be no need 
for a PTC element or other form of electrical reheating of the cabin supply air. 
This could be achieved by using the condenser heat from an indirectly cooled 
A/C condenser for reheating. This would be possible, since the condenser 
transferred -on average- 2.0 - 2.5 kW of heat during the time when re-heating 
was needed, based on experimental measurements. A theoretical evaluation of 
this possibility was performed, assuming a heater power consumption of zero 
(free heating) for the RH cyc and RH ctn cases. If using an indirectly cooled 
condenser, this modification would yield another 5% (RH cyc) and 7.5% (RH ctn) 
reduction of total energy consumption compared to the F cyc control baseline. 
This would add up to a 46% (RH cyc) and 64% (RH ctn) decrease in total energy 
169 
consumption. TPF would consequently increase to 26% (RH cyc) and 82% (RH 
ctn) above F cyc baseline TPF, as seen in Figure ?10.3. 
 
Figure ?10.2: A/C Controls Accumulated Performance Results Comparison 
 
-80
 -60
 -40
 -20
 0
 20
 40
 60
 80
 Q air E in (C+P) E in 
(C+P+H)
 TPF TPF 
(C+P+H)
 En
 er
 gy
  [%
 ];
  T
 P
 F 
[%
 ]
 Performance Parameters
 F cyc (2LP) RH cyc (2LP) RH ctn (2LP)
 Performance - Control Mode (2LP, soak, %F cyc)
 -80
 -60
 -40
 -20
 0
 20
 40
 60
 80
 Q air E in (C+P) E in 
(C+P+H)
 TPF TPF 
(C+P+H)
 En
 er
 gy
  [%
 ];
  T
 P
 F 
[%
 ]
 Performance Parameters
 F cyc (2LP) RH cyc (2LP+IS) RH ctn (2LP+IS)
 Performance - Control Mode (2LP, soak, %F cyc)
 a)
 b)
170 
 
Figure ?10.3: Theoretical Change in A/C Ctrl Accumulated Performance when Using 
an Indirectly Cooled Condenser for Supply Air Reheating 
 
10.2 Cabin Preconditioning Performance and Benefits 
Chapter ?10.2 introduces a comparison of performance results for 
extended driving after pull-down when the passenger cabin was preconditioned 
to a certain temperature. Three initial cabin temperatures were compared: 
soaked condition (S) at 42?C, ambient condition (A) at 31?C, and comfort 
conditions (C) at 25?C. The initial cabin temperature was 1 K above ideal 
preconditioning temperature at ambient (A) and at comfort preconditioning (C). 
Cabin preconditioning can be achieved with a variety of techniques. 
Preconditioning close to ambient temperature may be achieved by simply parking 
in a garage, in some cases by forced convection, or by having an advanced 
glazing to reduce soak by reducing transmittance of the windows. 
Preconditioning close to comfort conditions may be achieved either by parking in 
-100
 -80
 -60
 -40
 -20
 0
 20
 40
 60
 80
 100
 Q air E in (C+P) E in 
(C+P+H)
 TPF TPF 
(C+P+H)
 En
 er
 gy
  [%
 ]; 
TP
 F [
 %
 ]
 Performance Parameters
 F cyc (2LP) RH cyc (2LP) RH ctn (2LP)
 Performance - Control Mode (2LP, soak, %F cyc)
171 
an air-conditioned or well ventilated parking garage, by intermittently running the 
vehicle air-conditioning system (when the EV is plugged in), or by using a 
charged thermal storage device while parking in the shade or in a garage.  
A comparison of accumulated performance parameters for different cabin 
preconditioning settings is shown in Figure ?10.4. The tests were performed at 
30?C/50% ambient conditions, 950 W solar load, highway driving conditions, and 
a test time of 60 minutes. Figure ?10.4 a) shows a comparison for Frost cycling 
control, while Figure ?10.4 b) shows a comparison for RH cycling control. It was 
observed that cabin preconditioning did not decrease overall power consumption, 
when utilizing frost cycling control. While preconditioning decreased compressor 
power consumption by about 5% (E in C+P), it increased heater power 
consumption significantly, since the heater had to be employed much earlier in 
the test. This resulted in an overall increase of power consumption (E in C+P+H) 
of about 5% and 10% for ambient preconditioning and comfort preconditioning, 
respectively. This trends can be observed in Figure ?10.5.  It is assumed here that 
heating is not free, as it would be in the case of utilizing engine coolant from an 
internal combustion engine, which is at the same time the primary engine for the 
vehicle. The increase in heater energy over the course of a test resulted in a 
decrease of TPF of 4% for ambient preconditioning and 8% for comfort 
preconditioning. 
However, when utilizing control strategies which reduce the need for 
reheating and allow the compressor to cycle or reduce its capacity, cabin 
preconditioning can achieve significant energy savings. As shown in Figure ?10.4 
172 
b), compressor and pump power consumption could be reduced by as much as 
30% for ambient preconditioning and 40% for comfort preconditioning. At the 
same time, overall power consumption, including heater power, was reduced by 
25% - 30%. Overall TPF was similar to the unconditioned cabin, as the available 
energy for cooling was reduced by reducing compressor capacity and heater 
power. 
Figure ?10.5 shows transient profiles of system power consumption for 
cabin preconditioning scenarios while using frost cycling control in the R290 2LP 
system (no ice storage). The comparison of profiles explains the fact that cabin 
preconditioning does not help reduce power consumption when compressor and 
heater are used continuously, as is the case with frost cycling and reheating 
control. While initial compressor power decreases slightly when preconditioning 
the cabin to ambient and comfort temperature, the heater has to kick in earlier, 
since the cabin reaches thermal comfort condition faster. 
The heater power necessary to reheat the supply air could reach up to 1.4 
kW for comfort preconditioning. This resulted in a significant increase in overall 
power consumption during the initial 20 minutes of the test, until the compressor 
started to cycle to prevent frosting. At this point, thermal conditions were similar 
between tests and power consumption for the remainder of the cycle was alike.   
Figure ?10.6 a) shows a comparison of the cabin temperature profiles, as 
well as power consumption profiles, when only the charged ice storage box 
without the vapor compression cycle was used for air-conditioning. For a soaked 
173 
cabin, a solar load of 950 W and the load of one passenger, the ice storage box 
alone was unable to pull the cabin down to comfort set point. 
 
Figure ?10.4: Preconditioning - Accumulated Performance Results Comparison 
 
-15
 -10
 -5
 0
 5
 10
 15
 Q air E in (C+P) E in 
(C+P+H)
 TPF TPF 
(C+P+H)
 En
 er
 gy
  [%
 ];
  T
 P
 F 
[%
 ]
 Performance Parameters
 soak ambient comfort
 Performance - Pre-Conditioning (F cyc, 2LP, %soak)
 -60
 -40
 -20
 0
 20
 40
 60
 Q air E in (C+P) E in 
(C+P+H)
 TPF TPF 
(C+P+H)
 En
 er
 gy
  [%
 ];
  T
 P
 F 
[%
 ]
 Performance Parameters
 soak ambient comfort
 Performance - Pre-Conditioning (RH cyc, 2LP+IS, %soak)
 a)
 b)
174 
A similar result was observed when preconditioning the cabin close to 
ambient temperature. However, when preconditioning the cabin close to comfort 
temperature, the ice storage box was able to pull the cabin down to comfort set 
point and retain thermal comfort for an extended period (> 60 minutes). 
 
Figure ?10.5: Preconditioning - Power Consumption Transient Profile Comparison 
 
Figure ?10.6 b) shows a comparison of power consumption profiles. The 
coolant pump was continuously on when the cabin was preconditioned to soaked 
and ambient temperature. When pre-conditioned to comfort temperature, cooling 
capacity was controlled by on/off control, as well as PID control during the on-
 times. It can be observed that the ice storage box was losing capacity towards 
the end of the test when PID control increased pump speed to retain thermal 
comfort in the passenger cabin.  
0
 500
 1000
 1500
 2000
 2500
 3000
 3500
 4000
 4500
 5000
 0 10 20 30 40 50 60
 Po
 we
 r C
 on
 su
 mp
 tio
 n [
 W]
 Time [min]
 2LP (S) 2LP (A) 2LP (C)
 Total Power Consumption (2LP, F cyc)
 S
 A
 C
175 
This result has some interesting implications. First, there is no vapor 
compression system needed to keep the cabin at comfort temperature when the 
cabin is preconditioned. Second, although the ice storage box was not able to 
pull down the cabin from ambient preconditioning or higher initial starting 
temperatures, this is only a problem of heat exchanger optimization. The ice 
storage box provides enough energy for a cabin pull down in terms of fusion 
energy of the stored ice, but the heat transfer rate has to be increased by 
optimizing the ice storage heat exchanger for this application. Clodic [121] 
showed that the average commute in the US lasts about 25 minutes. A 
comparison of energy consumption relative to R290 2LP Frost cycle operation 
with soaked cabin preconditioning for a 25 minute commute is shown in 
Table ?10.1. Approximately 15 minutes are spent on pull-down with the 
compressor permanently running, while the last 10 minutes are spent at comfort 
temperature with the compressor and heater controlling cabin temperature (and 
humidity in the RH cyc cases). As can be observed in Table ?10.1, using 
alternative A/C controls (RH cyc) can reduce energy consumption by about 17%. 
Preconditioning the cabin at the same time to comfort condition results in 
combined energy savings of about 78% during the 25 minute commute. When 
using only ice storage and controlling cabin temperature by modulating coolant 
pump speed, energy savings of about 98% can be realized, which may preserve 
EV range to nearly the full extend. Essentially, nearly all of the A/C energy 
needed during a regular commute can be saved by preconditioning the cabin and 
176 
using latent storage without a VCC. The cabin relative humidity was fluctuating 
between 55% and 65%, which is at the high end of thermal comfort requirements.     
Table ?10.1: Comparison of Energy Consumption for a 25 Minute Commute Relative 
to 2LP F cyc (S) 
2LP, F cyc (S) 2LP, RH cyc (S) 2LP, RH cyc (C) IS (C) 
0% -16.8% -77.7% -97.6% 
 
 
Figure ?10.6: Preconditioning Ice Storage Performance Comparison 
0
 10
 20
 30
 40
 50
 60
 70
 80
 0 10 20 30 40 50 60
 P
 o
 w
 er
  [
 W
 ]
 Time [min]
 IS (S) IS (A) IS (C)
 Pre-Conditioning: Power Consumption (T30, highw)
 22
 24
 26
 28
 30
 32
 34
 36
 38
 40
 42
 44
 0 10 20 30 40 50 60
 Te
 m
 pe
 ra
 tu
 re
  [
 ?C
 ]
 Time [min]
 IS (S) IS (A) IS (C)
 Pre-Conditioning: Cabin Temperature (T30, hwy)
 a)
 b)
177 
10.3 Summary: Benefits of A/C Controls and Cabin Pre-Conditioning 
? Compressor and pump controls to reduce A/C power consumption 
were implemented. The control of  humidity by the compressor, as well 
as the preconditioning of the passenger cabin were investigated as 
possible strategies to reduce power consumption. 
? All three control strategies, Frost cycling, Relative Humidity cycling, 
and Relative Humidity continuous operation, were able to keep the 
cabin temperature at comfort set point after pull down. Relative 
Humidity cycling and Relative Humidity continuous operation kept the 
cabin supply relative humidity at 80% and the cabin relative humidity at 
50% throughout the test.  
?  Energy consumption of the R290 secondary loop system, including 
heater power, could be reduced by as much as 40% for relative 
humidity cycling, and as much as 60% for relative humidity continuous 
control when compared to Frost cycling control. When using ice 
storage, energy consumption of relative humidity continuous did not 
decrease further. However, relative humidity cycling reduced the 
energy consumption by about 55%. 
? Preconditioning the passenger cabin to ambient temperature or 
comfort temperature (instead of soaked condition) does not reduce 
overall power consumption (including heater power) for control 
strategies that rely on the compressor and heater being turned on for 
most of the time. However, when employing the relative humidity 
cycling control strategy, cabin preconditioning could reduce power 
consumption by 25% (A) and 30% (C). 
? When operating only the charged ice storage box and the coolant 
pump without the vapor compression cycle, the cabin could not be 
pulled down from an initial hot soaked condition. However, when pre-
 conditioned to comfort temperature, the ice storage box was able to 
178 
keep the cabin at comfort temperature for one hour. Latent storage 
may therefore be used in combination with cabin preconditioning to 
either replace or significantly reduce the use of the vapor compression 
system during commutes.  
179 
11 Transient Simulation of Secondary Loop Systems 
Modelica? is an object-oriented, equation based language used to model a 
variety of physical systems. Modelica? is heavily used in the automotive industry 
to design energy efficient vehicles, as well as design thermal management and 
air-conditioning systems. The programming environment Dymola was used to 
develop the secondary loop system, as well as perform transient simulation. 
 
11.1 Cabin Model 
A lumped model of a passenger cabin was developed in Modelica? to 
allow transient simulation of automotive air-conditioning systems. The cabin 
model developed by Gado [12] was used as a starting point and modified for the 
present research. The structure of the passenger cabin model, as well as the 
equations and modifications applied to Gado's model are discussed in the 
following chapter.  
 
11.1.1 Cabin Model Equations and Structure 
The heart of the model are a set of energy, mass, and moisture balance 
equations, which compute the psychrometric processes in a passenger cabin. 
The relevant equations are explained in the following paragraph. Table ?11.1 
provides an overview of the indices used in the cabin model equations.   
180 
Table ?11.1: Cabin Model Equations - Index Explanation 
 
 
 
 
 
The set of equations contains a sensible sub set, as well as a subset for 
latent energy and mass balance. Equations (24) through (26) relate to the 
sensible portion, while Equations (27) through (29) related to the latent portion. 
Equation (24) describes the relation of cabin air and cabin interior thermal 
storage with the various loads on the cabin, such as solar insolation, sensible 
passenger load, and heat transfer with ambient. In comparison with Gado [12], 
the term which described the load on the evaporator, due to introduction of 
ambient air was left out, as it did not relate to load on the cabin itself. Equation 
(25) describes the change in interior temperature based on convective heat 
transfer with cabin air. Equation (26) describes the relation between the cabin air 
and the return air which leaves the cabin. The equation was modified from 
Gado's version by removing the influence of mixture with ambient air. In the 
present model, mixing with ambient air occurs after the air leaves the cabin and 
before it returns upstream to the evaporator (DX) or cooler (2LP). Energy, mass, 
and moisture balance for mixing are therefore not part of the cabin model. At the 
same time, an enthalpy balance was used due to a higher accuracy.  
Index Explanation 
c Cabin air 
i Cabin interior 
s Supply air 
r Return air 
m Mixed air 
a Ambient air 
sol Solar (insolation) 
ps Passenger, sensible 
pl Passenger, latent 
181 
      
   
  
     
   
  
                                         (24) 
    
   
  
                    (25) 
          (26)  
Equation (27) describes the moisture balance in the cabin, taking into 
account the storage of moisture in the cabin air, the introduction of moisture by 
the supply air, as well as the passengers, and the decrease of moisture due to 
the leaving return air. Equation (28) describes the relation between the humidity 
ratio of cabin air to the humidity ratio of return air leaving the cabin. As in 
previous equations, the influence of mixing with ambient air was omitted. 
     
   
  
                   (27)  
          (28) 
Parameters Xs and Xr are mass fractions of water vapor to total mass of air. 
There relation to humidity ratio, W,  is given by Equation (30). 
  
 
   
     (29) 
Finally, since there is no mixing with ambient air inside of the cabin, the 
mass flow rate of return air equals the mass flow rate of supply air, as shown in 
Equation (30). 
          (30)  
The air leaving the cabin is allowed to mix with ambient air, based on the 
recirculation ratio. To this end, an air splitter, as well as an air mixer were used. 
The air splitter splits up the air stream leaving the cabin into two streams, based 
182 
on a predefined recirculation ratio. One of the partial streams is mixed with 
ambient air, while the other is dumped to ambient. This is necessary to preserve 
mass balance with the added ambient air stream. Equations (31) through (33) 
determine the mass flow rate, energy flow rate, and moisture flow rate of the air 
stream which later mixes with ambient air and returning to the evaporator (out,2). 
Equations (34) through (36) shows the overall mass, energy and moisture 
conservation for the air splitter.  
                     (31) 
                           (32) 
                           (33) 
 
                    (34) 
                              (35) 
                              (36) 
The air mixer adiabatically mixes fresh, ambient air with the partial return 
air stream leaving the cabin, based on the recirculation ratio. The mass flow rate 
of ambient air is determined by Equation (37).  Equations (38) through (40) show 
the mass, energy and moisture balance for the mixing process. Following the 
adiabatic mixing process, the air is recirculated upstream of the evaporator.  
                     (37) 
                  (38) 
                          (39) 
183 
                          (40) 
Figure ?11.1 shows the top level structure of the cabin model in Modelica?. 
Inlets for supply air, ambient air, solar insulation, as well as ambient temperature 
(for heat transfer) are present. Two air sinks serve as outlets in the bench test 
case. When combined with a direct expansion cycle, the supply air inlet would be 
connected to the outlet of the evaporator (DX) or cooler (2LP), while one of the 
outlets would be connected to the evaporator or cooler inlet. 
 
Figure ?11.1: Modelica Passenger Cabin Model 
 
184 
11.1.2 Cabin Model Validation 
The cabin model was validated, based on the process outlined in Gado 
[12]. A supply air temperature profile from Huang [15] was used as input to the 
cabin model. Based on the physical properties of the cabin used for validation, 
shown in Table ?11.2, the cabin model calculates the state of the cabin air. The 
resulting cabin air profile is compared to the experimental values from Huang [15], 
as well as to the cabin model validation profile from Gado [12]. Figure ?11.2 
shows the comparison of cabin temperature profiles. It can be observed that the 
current cabin model (Eisele 2012) coincides with the validation results from Gado, 
despite the subtle differences introduced to the  equations. Following the 
validation, the cabin model was used in the transient simulation of the direct 
expansion, as well as the secondary loop system. 
Table ?11.2: Cabin Model Validation Physical Parameters 
Parameter Magnitude Source 
Ambient Temperature [?C] 43.3 Huang (1998) 
Ambient Relative Humidity [%] 65 Huang (1998) 
Solar Insolation [W] 950 Huang (1998) 
Degree of Soak [K] 16.7 Huang (1998) 
Cabin Surface Area 30 Huang (1998) 
Overall Heat Transfer Coefficient [W/m2K] 2.5 Tuned 
Internal Cabin Volume [m3] 8 Huang (1998) 
Interior Mass [kg] 200 Huang (1998) 
Interior Specific Heat [J/kg K] 400 Huang (1998) 
Interior Surface Area [m2] 3 Typical Value 
Convection Coefficient (interior) [W/m2 K] 100 Tuned 
Number of Passengers 0 Huang (1998) 
Fresh Air Ratio 0 Huang (1998) 
Supply Air Temperature [?C] profile Huang (1998) 
Supply Air Relative Humidity [%] profile Typical Value 
 
185 
 
Figure ?11.2: Cabin Model Validation Results 
 
11.2 Direct Expansion System 
As a first step towards transient simulation of secondary loop systems, an 
R134a direct expansion (DX) cycle was modeled and validated for steady-state, 
as well as transient operation. The direct expansion system was modeled after 
the experimental DX system, using microchannel refrigerant to air heat 
exchangers, a variable speed fixed displacement compressor, as well as a PID 
controlled electronic expansion valve. The refrigerant components were modeled 
by Qiao [129], a member of the Center for Environmental Energy Engineering 
(CEEE). For the present research, the components were combined to a direct 
expansion cycle and tuned to match experimental R134 DX system performance.   
20
 25
 30
 35
 40
 45
 50
 55
 60
 0 2000 4000 6000 8000
 Tem
 per
 atu
 re [
 ?C]
 Time [s]
 Dymola Cabin Model Verification - Temperature 
Eisele (2012) Gado (2006) Huang (1998)
186 
11.2.1 Direct Expansion Model and Components 
Figure ?11.4 shows the structure of a R134a DX stationary (steady-state) 
model in Dymola. The stationary model consists of a variable speed fixed 
displacement compressor (1), a microchannel condenser (2), a PID controlled 
electronic expansion valve (3), and a two pass microchannel evaporator (4). 
Refrigerant lines (5) permit heat transfer with ambient to take heat loss into 
account. Two types of boundaries, an air sink (6), and an air inlet boundary (7) 
are used to fix air-side conditions in the stationary model.  The components of 
the direct expansion system were modeled by Qiao [129]. A brief description and 
summary of governing equations of the main direct expansion system component 
models is given in ?Appendix H.  
187 
 
Figure ?11.3: Direct Expansion Steady-state System Model 
 
1
 2
 3
 4
 5
 6
 7
  
# Component 
1 Compressor, fixed displacement 
2 Condenser, microchannel 
3 Expansion Valve, PID controlled 
4 Evaporator, microchannel 
5 Refrigerant Piping 
6 Air Sink 
7 Air Boundary 
188 
Figure ?11.4 shows the structure of the R134a DX transient system model. 
For the purpose of transient simulation, the DX model was combined with the 
automotive cabin model. The cabin model (6) replaces the air sink and inlet 
boundary of the stationary model evaporator, shown in Figure ?11.3. The cabin 
model receives supply air from the second evaporator bank outlet. Ambient air 
for mixing is provided by an air inlet boundary (7), while heating by solar 
insolation is added through a heat transfer boundary (9). With this configuration, 
the R134a DX model is able to simulate transient pull-down, as well as drive 
cycle tests. While pull-down tests are performed at constant compressor speed 
and condenser air flow rate in the present research, drive cycle tests feature a 
predefined compressor speed and condenser air flow rate profile. Speed and air 
flow rate profiles are supplied to the compressor and the condenser model by 
time table modules, as shown in Figure ?11.4. 
The following chapter discusses validation results of the R134a DX 
system model for steady-state, as well as transient operation.  
189 
 
Figure ?11.4: R134a Direct Expansion Transient System Model 
1
 2
 3
 4 5
 6
 7
 8
 9
  
# Component 
1 Compressor, fixed displacement 
2 Condenser, microchannel 
3 Expansion Valve, PID controlled 
4 Evaporator, microchannel 
5 Refrigerant Piping 
6 Cabin Passenger Model 
7 Air Boundary 
8 Air Sink 
9 Heat Flow Boundary 
190 
11.2.2 R134a Direct Expansion Model Validation 
The R134a DX model was validated with experimental data for steady-
 state, as well as transient operation. The inner geometry of heat exchangers, 
expansion valve, and compressor (flow channels) were unknown. Steady-state 
experimental data was used to tune the component geometries, as well as the 
refrigerant charge of the system. Initialization of the model requires initial 
temperature, pressure, as well as initial enthalpy in the refrigerant components to 
be known. While the pressure can be assumed to be saturation pressure at the 
initial temperature if the system is charged correctly, initial enthalpy requires the 
knowledge of system volume and refrigerant charge to determine initial density. 
Initial volume can be estimated by measuring the combined volume of all 
components in the experimental system. However, estimation of refrigerant 
charge for transient simulation poses a challenge, although the charge is known 
for the experimental system. Refrigerant charge while the compressor is running 
(i.e. while there is a mass flow rate through the components) varies from initial 
charge, due to the behavior of the mass derivatives in the component models. 
Therefore, all system models need to be tuned by adjusting initial enthalpy 
(which is a function of initial refrigerant charge).  
Table ?11.3 shows the results of the steady state R134 DX model 
validation for two different ambient conditions, 25?C and 50% relative humidity, 
as well as 35?C and 40% relative humidity. At 35?C ambient temperature, the 
model was validated for idling conditions, as well as highway driving conditions. 
From Table ?11.3 it can be observed that high side pressure is slightly 
191 
overpredicted at highway driving conditions, while low side pressure is slightly 
overpredicted at idling conditions. Discharge pressure is consistently 
underpredicted by as much as 7 K. However, cooling capacity is reasonably well 
predicted with an absolute deviation of as much as 50 W, which compares to a 1% 
relative difference compared to experimental data. Compressor power 
consumption was predicted by the model with a deviation of less than 1% 
throughout the three cases that were validated. As a result, system COP was 
predicted with less than a 1% deviation. Due to the excellent match of cooling 
capacity, compressor power consumption and consequently system COP, the 
tuned components were subsequently used in the transient simulation of pull-
 down and NEDC tests.  
Table ?11.3: R134a Direct Expansion Steady-state Model Validation Results 
Parameter Experiment Simulation Deviation [abs] 
25?C / 50% ; highway    
High Side Pressure [kPa] 945.5 983.0 37.5 
Low Side Pressure [kPa] 241.3 244.1 2.8 
Discharge Temperature [?C] 65.2 58.1 -7.1 
Cooling Capacity [W] 4724.2 4673.8 -50.4 
Compressor Power [W] 2578.5 2558.7 -19.8 
System COP [-] 1.83 1.83 0.0 
35?C / 40% ; idle    
High Side Pressure [kPa] 1235.6 1233.1 -2.5 
Low Side Pressure [kPa] 415.7 453.4 37.7 
Discharge Temperature [?C] 65.0 61.0 -4.0 
Cooling Capacity [W] 2965.6 3000.4 34.8 
Compressor Power [W] 1167.6 1171.9 4.3 
System COP [-] 2.55 2.56 0.4 
35?C / 40% ; highway    
High Side Pressure [kPa] 1341.8 1382.8 41 
Low Side Pressure [kPa] 306.9 310.0 3.1 
Discharge Temperature [?C] 78.6 71.3 -7.3 
Cooling Capacity [W] 5510.4 5456.3 -54.1 
Compressor Power [W] 3397.2 3379.6 -17.6 
System COP [-] 1.62 1.61 0.6 
 
192 
Figure ?11.5 shows a comparison of simulation and experimental results of 
an NEDC drive cycle test, performed at 30?C/50% ambient condition and an 
initial soak of 5.6 K. Profiles of cabin temperature, cooling capacity, and power 
consumption are presented.  
Figure ?11.5 a) shows a comparison of cabin temperature profiles. During 
model validation, it was observed that the model would constantly underpredict 
the cabin temperature at the beginning of a pull-down (or NEDC) test, while the 
curves would run parallel during the mid to late section of the comparison. This 
can be explained by the thermal storage of the evaporator (DX)/cooler (2LP) 
windtunnel. The thermal mass of the windtunnel  sensibly cools or heats the air 
between the point where it leaves the evaporator/cooler and the point where the 
thermocouple grid measures air-side outlet temperature. This thermal mass is 
unaccounted for in the Modelica? model. To adjust for this discrepancy, the cabin 
model physical parameters were tuned by adjusting the specific heat capacity of 
the interior, as well as the convection coefficient between the cabin air and the 
interior to reflect the behavior of the experimental system. From Figure ?11.5 a) it 
can be concluded that after the tuning of the thermal mass of the interior, the 
model accurately predicts the cabin temperature profile during the NEDC test. 
Cabin temperature at the beginning of the cycle is still slightly lower compared to 
experimental results, but the overall profile is coherent. 
Figure ?11.5 b) shows a comparison of cooling capacity profiles. The 
overall trends for cooling capacity match well between the simulation and the 
experimental profiles. However, it can be observed that during the later parts of 
193 
the city cycle, significant deviations between the experimental and the simulation 
trends occur. The specific parts of the cycle were highlighted with a dotted circle 
in Figure ?11.5 b). The deviations can be explained by the lack re-evaporation of 
condensate off the evaporator/cooler fins in the model. The experimentally 
observed cooling capacity during idling at later parts of the NEDC cycle includes 
the effect of re-evaporation and subsequent changes in the latent capacity. The 
model is not able to predict re-evaporation and consequently determines an 
increased total cooling capacity during idling periods. 
  Figure ?11.5 c) shows a comparison of NEDC power consumption profiles. 
The overall trend of power consumption is well observed by the model. Small 
deviations occur at the points of load shifting at the end of an acceleration or 
deceleration phase. The model does not take the effects of inertia in the 
compressor, as well as in the electric motor which drives the compressor, into 
account.  
Table ?11.4 shows a comparison of accumulated performance parameters. 
Based on the comparison, the model is able to predict the total energy available 
for cooling the air within a deviation of 6.7%. The total energy put into the direct 
expansion system in the form of compressor power consumption was predicted 
with a deviation of 4.2%. The difference between the simulated and the 
experimental transient performance factor was observed to be 2.0%. 
194 
 
Figure ?11.5: R134a Direct Expansion NEDC Validation Results 
15
 20
 25
 30
 35
 40
 0 5 10 15 20
 T
 e
 m
 p
 e
 r
 a
 t
 u
 r
 e
  [
 ?
 C
 ]
 Time [min]
 Cabin Temperature
 T cabin (Exp) T cabin (Sim)
 0
 1000
 2000
 3000
 4000
 5000
 6000
 0 5 10 15 20
 C
 a
 p
 a
 c
 i
 t
 y
  [
 W
 ]
 Time [min]
 Cooling Capacity (air)
 Q air (Exp) Q air (Sim)
 0
 1000
 2000
 3000
 4000
 0 5 10 15 20
 P
 o
 w
 e
 r
  [
 W
 ]
 Time [min]
 Compressor Power
 P compr (Exp) P compr (Sim)
 a)
 b)
 c)
195 
Table ?11.4: R134a Direct Expansion NEDC Accumulated Performance Validation 
Parameter Experiment Simulation Deviation [%] 
NEDC, 35?C/50%, 5.6 K soak     
Accumulated Capacity (air) [kWh] 1.04 1.11 6.7 
Accumulated Compr. Power [kWh] 0.72 0.75 4.2 
Transient Performance Factor [-] 1.45 1.48 2.0 
 
 
Figure ?11.6 shows a transient comparison of cabin temperature, cooling 
capacity, and power consumption for a pull-down test at 30?C/50% ambient 
condition and 2,500 rpm compressor speed, with an initial degree of soak of 12 K. 
With the tuning of components and cabin model, described in above paragraph, 
the model is able to predict pull-down performance reasonably well.  
Figure ?11.6 a) shows a comparison of cabin temperature profiles. The 
model predicted cabin temperature within 0.5 K, except during the initial minutes 
of the pull-down test.  Figure ?11.6 b) compares the experimental and simulated 
cooling capacity profiles. The initial transients in capacity are not well captured, 
possibly due to a difference in electronic expansion valve (EXV) behavior. During 
the remainder of the pull-down test cooling capacity is slightly overpredicted by 
the model. Experimental compressor power, shown in Figure ?11.6 c), is well 
matched by the model. Again, initial transients are not captured, and the model 
slightly underpredicts compressor power consumption during the pulldown 
process.  
196 
 
Figure ?11.6: R134a Direct Expansion Pull-down Validation Results 
20
 25
 30
 35
 40
 45
 0 2 4 6 8 10 12
 T
 e
 m
 p
 e
 r
 a
 t
 u
 r
 e
  [
 ?
 C
 ]
 Time [min]
 Cabin Temperature
 T cabin (Exp) T cabin (Sim)
 0
 1000
 2000
 3000
 4000
 5000
 6000
 7000
 8000
 0 2 4 6 8 10 12
 C
 a
 p
 a
 c
 i
 t
 y
  [
 W
 ]
 Time [min]
 Cooling Capacity
 Q air (Exp) Q air (Sim)
 0
 1000
 2000
 3000
 4000
 5000
 0 2 4 6 8 10 12
 P
 o
 w
 e
 r
   
[
 W
 ]
 Time [min]
 Compressor Power
 P comp (Exp) P comp (Sim)
 a)
 b)
 c)
197 
Accumulated performance results are compared in Table ?11.5. The model 
predicted accumulated capacity and accumulated power consumption to within 
6%. Since cooling capacity was overpredicted and power consumption was 
underpredicted, the simulation TPF increased by 10.5% over experimentally 
measured TPF. 
Table ?11.5: R134a Direct Expansion Pull-down Acc. Performance Validation 
Parameter Experiment Simulation Deviation [%] 
Plldn, 30?C/50%, 12K soak, 1550rpm     
Accumulated Capacity (air) [kWh] 0.744 0.788 5.9 
Accumulated Compr. Power [kWh] 0.559 0.536 -4.1 
Transient Performance Factor [-] 1.33 1.47 10.5 
 
11.3 Secondary Loop System 
To model the secondary loop system, several components had to be 
introduced in addition to the direct expansion cycle components. As part of this 
research, a generic glycol media package has been developed, as well as a 
coolant pump, coolant tubes, an intermediate heat exchanger, as well as a 
coolant to air heat exchanger. The following chapters introduce the coolant  
components and the validation of the secondary loop system model.  
 
11.3.1 Coolant Cycle Model and Components 
Figure ?11.7 shows the secondary loop steady-state system model. 
Components previously described for the direct expansion system, such as 
compressor (1), condenser (2), expansion valve (3), and refrigerant lines (5) don't 
change. The length of the refrigerant lines was changed according to the 
198 
changes observed when the experimental test facility was switched from direct 
expansion to a secondary loop system layout. The evaporator was replaced by a 
refrigerant to coolant plate type heat exchanger (4). A coolant to air heat 
exchanger (6) was used to cool down the supply air. The coolant was circulated 
through the secondary loop by a coolant pump (7). Coolant lines (8) were used to 
connect the secondary loop components.  
The glycol media package for aqueous solutions of ethylene glycol was 
developed based on equations from M. Conde Engineering [109]. Property 
equations and coefficients used for the present research are restated 
in ?Appendix G.  
Governing equations for the coolant tubing are given by Equations (41), 
(42), (43), (44) and (45). Equation (41) provides information about the 
determination of the forced convection heat transfer coefficient for liquid coolant. 
The heat transfer coefficient for each segment i is a function of the inner diameter 
of the tube, thermal conductivity, dynamic viscosity, mass flow rate, and specific 
heat capacity. Equation (42) provides information about the pressure drop 
correlation used for the coolant tubing. The pressure drop for each segment in 
the Blasius type solution is a function internal tube diameter, segment length, 
mass flow rate, density and dynamic viscosity.  
                                       (41) [130] 
               
  
    
               (42) [131] 
199 
Forced convection heat transfer in each tube segment was determined by 
Equation (43), where As was surface area, n the number of segments, Tw was 
wall temperature and T was the temperature of the coolant. 
       
  
    
              (43) 
 Mass and energy conservation equations were used for the coolant tube 
control volume, as shown in Equations (44) and (45). Equation (45) allows for 
thermal storage in the secondary loop components. 
               (44) 
      
   
  
                    (45) 
The above equations used for the coolant tube control volume were 
employed for the coolant-side control volumes of the intermediate heat 
exchanger and the air cooler as well. Flow splitters and mixers were added 
before and after the control volume to model microchannel and plate type heat 
exchanger behavior.  
A quasi-steady state, efficiency based model was used for the coolant 
pump. Equation (46) shows the energy equation for a generic pump, where p is 
pressure,???is?density,?v?is?velocity,?g?is?gravitational?acceleration?and?ht?is?height.? 
   
        
 
 
    
     
 
 
                   (46) 
The electrical power consumed by the pump, depending on volumetric, 
hydraulic, and motor efficiency, is given by Equation (47). It was assumed that 
200 
half of the heat produced due to motor inefficiencies was introduced as heat gain 
into the fluid.    
   
      
               
     (47) 
To simulate transient operation, the secondary loop model was combined 
with the cabin model (9) to enable pull-down and drive cycle testing, as shown in 
Figure ?11.8. The cabin model receives supply air from the cooler air-side outlet, 
while the mixed air is returned back to the inlet of the cooler. Similar to the direct 
expansion cycle transient model, ambient air is introduced to the cabin model by 
an air inlet boundary (10).  
The following chapter discusses the validation of the secondary loop 
model with experimental data for steady-state, as well as transient operation. 
 
 
 
201 
 
Figure ?11.7: Secondary Loop Steady-state System Model 
1
 2
 3
 4
 5
 6 7 8
 9
 10
  
# Component 
1 Compressor, fixed displacement 
2 Condenser, microchannel 
3 Expansion Valve, PID controlled 
4 Intermediate Heat Exchanger, plate type 
5 Refrigerant Piping 
6 Air Cooler, microchannel 
7 Coolant Pump 
8 Coolant Tubing 
9 Air Boundary 
10 Air Sink 
202 
 
Figure ?11.8: Secondary Loop Transient System Model 
1
 2
 3
 4 5
 6 7 8
 10
 9
 11
  
# Component 
1 Compressor, fixed displacement 
2 Condenser, microchannel 
3 Expansion Valve, PID controlled 
4 Intermediate Heat Exchanger, plate type 
5 Refrigerant Piping 
6 Air Cooler, microchannel 
7 Coolant Pump 
8 Coolant Tubing 
9 Cabin Model 
10 Air Sink 
11 Air Boundary 
12 Heat Flow Boundary 
203 
11.3.2 R290 Secondary Loop Model Validation 
The operation of the secondary loop system model was validated for 
steady-state, as well as transient operation. Table ?11.6 shows the validation 
results for steady-state operation at different ambient temperature and idling, as 
well as highway driving conditions. As for the direct expansion system model, the 
secondary loop components were tuned, since the interior geometry of the 
intermediate heat exchanger and the air cooler were unknown. Across different 
ambient conditions, as well as driving conditions, high side pressure in the vapor 
compression cycle is very well predicted. However, low side pressure is 
overpredicted at highway driving conditions, while it is slightly underpredicted at 
idling conditions. Discharge pressure is within 2 K of the experimentally 
measured value during highway driving conditions, while it is underpredicted by 
about 5.66 K at idling conditions. Cooling capacity is underpredicted by less than 
6% for highway driving, while it is overpredicted by about 6% for idling conditions. 
Simulated results for compressor power consumption are within 2% of the 
experimentally measured values for the examined operating conditions. Air-side 
COP was captured within deviations of 0.2 absolute, which corresponds to a 7.5% 
relative deviation at 25?C highway driving. Air-side cooler outlet temperature 
matches with measured results within less than 2 K. Following steady-state 
validation, the secondary loop model was used in conjunction with the cabin 
model to simulate transient pull-down and drive cycle operation. 
 
204 
Table ?11.6: Secondary Loop Stationary System Validation Results 
Parameter Experiment Simulation Deviation [abs] 
25?C / 50% ; highway    
High Side Pressure [kPa] 1326.8 1334.7 7.9 
Low Side Pressure [kPa] 412.2 449.1 36.9 
Discharge Temperature [?C] 58.8 57.4 -1.4 
Cooling Capacity [W] 4425.7 4181 -244.7 
Compressor Power [W] 2161.1 2143 -18.1 
System COP [-] 1.98 1.83 -0.2 
Cooler Outlet Temp. (col) [?C] 7.5 4.6 -2.9 
Cooler Outlet Temp. (air) [?C] 7.7 7.7 0.0 
30?C / 40% ; idle    
High Side Pressure [kPa] 1571.9 1572.6 0.7 
Low Side Pressure [kPa] 720.2 700.2 -20.0 
Discharge Temperature [?C] 66.3 60.8 -5.6 
Cooling Capacity [W] 2390.6 2541.5 150.9 
Compressor Power [W] 1032.5 1019.9 -12.6 
System COP [-] 2.17 2.29 0.1 
Cooler Outlet Temp. (col) [?C] 20.1 21.3 1.2 
Cooler Outlet Temp. (air) [?C] 20.0 18.5 -1.5 
30?C / 40% ; highway    
High Side Pressure [kPa] 1732.5 1744.7 12.2 
Low Side Pressure [kPa] 506.7 574.0 67.3 
Discharge Temperature [?C] 71.5 69.8 -1.7 
Cooling Capacity [W] 5022.8 4889.2 -133.6 
Compressor Power [W] 2809.4 2766.0 -43.4 
System COP [-] 1.7 1.7 0.0 
Cooler Outlet Temp. (col) [?C] 14.3 13.2 -1.1 
Cooler Outlet Temp. (air) [?C] 14.9 15.2 0.3 
 
Figure ?11.9 shows a comparison of NEDC performance between the R290 
2LP Modelica model and experimental results. Figure ?11.9 a) shows the 
validation of the cabin temperature profile over time. The model underpredicts 
cabin temperature by about 1 K throughout the NEDC cycle, but captures the 
general trends very well. The validation results for cooling capacity, shown in 
Figure ?11.9 b) present a similar conclusion as the R134a DX validation results. 
While the initial transients within the first minute were not captured well, the 
simulated results stay fairly true to experimentally measure results throughout the 
205 
latter stages of the cycle. Towards the middle of the NEDC, a slight 
overprediction of cooling capacity was observed, since the model cannot capture 
the re-evaporation of condensate during idling periods. Compressor power 
consumption, shown in Figure ?11.9 c), followed the trend of the experimental 
data. However, power consumption was slightly overpredicted throughout the 
cycle, which was especially prominent during the idling periods.  
A comparison of the total energy available for air-side cooling, the total 
input to the system, and the transient performance factor during the NEDC is 
shown in Table ?11.7. Deviations of accumulated capacity and accumulated 
compressor power are within 6%. Due to an overprediction of capacity, as well as 
compressor power, the simulated TPF shows only a very small deviation (-0.7%) 
from experimental TPF.  
Table ?11.7: Secondary Loop NEDC Acc. Performance Validation 
Parameter Experiment Simulation Deviation [%] 
NEDC, 35?C/50%, 5.6 K soak    
Accumulated Capacity (air) [kWh] 0.946 0.989 4.5 
Accumulated Compr. Power [kWh] 0.689 0.727 5.5 
Transient Performance Factor [-] 1.37 1.36 -0.7 
 
Figure ?11.10 shows a comparison of performance parameters during a 
pull-down test at 30?C/50% ambient condition and an initial degree of soak of 12 
K. The cabin temperature profiles during pull-down are compared in Figure ?11.10 
a). During the initial minutes, the Modelica model predicts a cabin temperature, 
which is decreased by about 1 K from the experimentally measured value. 
206 
 
Figure ?11.9: Secondary Loop NEDC Validation Results 
15
 20
 25
 30
 35
 40
 0 5 10 15 20
 T
 e
 m
 p
 e
 r
 a
 t
 u
 r
 e
  [
 ?
 C
 ]
 Time [min]
 Cabin Temperature
 T cabin (Exp) T cabin (Sim)
 0
 1000
 2000
 3000
 4000
 5000
 0 5 10 15 20
 C
 a
 p
 a
 c
 i
 t
 y
  [
 W
 ]
 Time [min]
 Cooling Capacity (air)
 Q air (Exp) Q air (Sim)
 0
 1000
 2000
 3000
 4000
 0 5 10 15 20
 P
 o
 w
 e
 r
  [
 W
 ]
 Time [min]
 Compressor Power
 P compr (Exp) P compr (Sim)
 a)
 b)
 c)
207 
This can be attributed to the thermal storage associated with the 
windtunnel in the experimental measurements. During the later part of the pull-
 down, simulated cabin temperature follows experimental cabin temperature very 
closely. A comparison of air-side cooling capacity is shown in Figure ?11.10 b). As 
observed in the direct expansion system results, capturing the initial transients 
due to the electronic expansion valve adjustments is challenging for the model. 
However, the general trends throughout the pull-down tests are observed, with a 
slight overprediction of cooling capacity by the model. Figure ?11.10 c) compares 
power consumption profiles for experimental and simulated compressor data. 
Much similar to the cooling capacity results, initial transients during the first 
minute of the pull-down test are not captured precisely. However, power 
consumption during the remainder of the test follows experimentally measured 
power consumption closely. 
Table ?11.8 shows the a comparison of accumulated pull-down 
performance. Total energy available for cooling cabin supply air was predicted 
with a deviation of 1.8% by the model. Energy consumed by the compressor over 
the duration of the test was captured by the model with a deviation of 2.4%. 
Simulated TPF decreased by 0.6% compared to experimental TPF. 
Table ?11.8: Secondary Loop Pull-down Acc. Performance Validation 
Parameter Experiment Simulation Deviation [%] 
Plldn, 30?C/50%, 12 K soak, 1550rpm     
Accumulated Capacity (air) [kWh] 0.892 0.876 1.8 
Accumulated Compr. Power [kWh] 0.631 0.616 2.4 
Transient Performance Factor [-] 1.41 1.42 -0.6 
 
208 
 
Figure ?11.10: Secondary Loop Pull-down Validation Results 
20
 25
 30
 35
 40
 45
 0 2 4 6 8 10 12 14 16
 T
 e
 m
 p
 e
 r
 a
 t
 u
 r
 e
  [
 ?
 C
 ]
 Time [min]
 Cabin Temperature
 T cabin (Exp) T cabin (Sim)
 0
 1000
 2000
 3000
 4000
 5000
 6000
 0 2 4 6 8 10 12 14 16
 C
 a
 p
 a
 c
 i
 t
 y
  [
 W
 ]
 Time [min]
 Cooling Capacity
 Q air (Exp) Q air (Sim)
 0
 500
 1000
 1500
 2000
 2500
 3000
 3500
 0 2 4 6 8 10 12 14 16
 P
 o
 w
 e
 r
   
[
 W
 ]
 Time [min]
 Compressor Power
 P comp (Exp) P comp (Sim)
 a)
 b)
 c)
209 
11.4 Influence of Coolant Volume on Transient Performance 
The current experimental system had a coolant charge of about 5 liters. 
The coolant charge was heavily dependent on the facility layout and the 
components and space available. It is to be expected that a smaller charge may 
result in faster cool down of the cabin and a better system efficiency. On the 
other hand, a larger coolant charge may be beneficial for thermal comfort during 
off-cycle periods. The R290 2LP Modelica? model, introduced in Chapter ?11.3, 
was used to perform a parametric study investigating the influence of coolant 
volume in the secondary loop on cabin pull-down and cabin warm-up 
performance. To this end, the length of coolant tubing in the model was varied. 
This approach was chosen, due to the possibility of secondary loop systems 
being used for electronics and battery cooling, as well as seat cooling/heating or 
radiative surfaces. A water-ethylene glycol mixture of 32wt% ethylene glycol was 
used, which equals the coolant used in the experimental study. The pull-down 
test was simulated using an ambient condition of 30?C and 50% relative humidity. 
At the start of the simulation the cabin was soaked to 42?C. The R290 2LP 
highway compressor speed of 1,550 rpm was chosen, as well as the R209 2LP 
coolant mass flow rate of 250 g/s. After reaching comfort temperature (24?C), the 
compressor was turned off and the cabin was allowed to warm up, while the 
coolant pump was still running to employ the thermal storage of the secondary 
loop for cooling during the off-cycle period. 
 Figure ?11.11 shows the influence of coolant volume on cabin pull-down 
performance. Figure ?11.11 a) shows supply temperature profiles for different 
210 
coolant charges. As expected, increasing coolant charge added thermal mass 
and increased supply temperature. Consequently, time to comfort increased with 
increasing coolant volume. Throughout most of the pull-down period, an increase 
in one liter of coolant resulted in roughly a 0.6 K increase in cabin supply 
temperature. Cooling capacity profiles are compared in Figure ?11.11 b). It was 
observed that the influence of coolant volume on capacity is largest during the 
initial minutes of the pull-down test. At the point of largest divergence, every 
additional liter of coolant decreases capacity by roughly 150 W. However, the 
significance of coolant charge diminishes after the initial minutes. Figure ?11.11 c) 
shows the system power (compressor + pump). Pump power consumption is 
dependent on coolant density and viscosity. Despite the increase in pipe length, 
it was found that pump power slightly decreased with increasing coolant charge, 
since the coolant stayed at a higher temperature and viscosity was reduced. The 
effects were negligible when compared to the overall  system power consumption. 
From Figure ?11.11 c) it was observed that changes in system power 
consumption resulting from variation of coolant charge were small.  
Figure ?11.12 shows the influence of coolant volume on cabin warm-up 
behavior during off-cycle periods. During the initial minutes, the addition of one 
liter of coolant may decrease the cabin supply temperature by about 0.7 K. The 
influence of coolant volume decreases with increasing warm-up time. After 15 
minutes, the effect of an additional liter of coolant volume on the supply 
temperature became negligible. Figure ?11.12 b) shows a comparison of cooling 
capacity, which is preserved by the thermal mass of the secondary loop system.  
211 
 
Figure ?11.11: Influence of Coolant Volume on Cabin Pull-down  
(R290 2LP, T30/RH50, 12 K soak, 1550 rpm, 250 g/s) 
0
 4
 8
 12
 16
 20
 24
 28
 32
 36
 40
 0 2 4 6 8 10 12 14 16
 Te
 m
 p
 e
 ra
 tu
 re
  [
 ?C
 ]
 Time [min]
 Supply Temperature
 2.0 L
 3.0 L
 4.0 L
 5.0 L
 6.0 L
 7.0 L
 8.0 L
 2,000
 2,500
 3,000
 3,500
 4,000
 4,500
 5, 00
 0 2 4 6 8 10 12 14 16
 C
 a
 p
 a
 ci
 ty
  [
 W
 ]
 Time [min]
 Cooling Capacity
 2.0 L
 3.0 L
 4.0 L
 5.0 L
 6.0 L
 7.0 L
 8.0 L
 2,000
 2,200
 2,400
 2,600
 2,800
 3,000
 0 2 4 6 8 10 12 14 16
 P
 o
 w
 e
 r 
[W
 ]
 Time [min]
 System Power
 2.0 L
 3.0 L
 4.0 L
 5.0 L
 6.0 L
 7.0 L
 8.0 L
 a)
 b)
 c)
212 
The influence of coolant volume on the system power during cabin warm-
 up is presented in Figure ?11.12 c). Since the compressor is turned off, only the 
coolant pump is contributing to power consumption. A larger coolant volume 
results in longer tubing, as well as a lower coolant temperature during cabin 
warm-up. Consequently, power consumption during off-cycle periods increases 
with increasing coolant volume. The relative increase can be as much as 3%-4% 
per additional liter of coolant, but the absolute increase is small. 
Figure ?11.13 shows the influence of coolant charge on typical pull-down 
performance metrics, such as time to comfort, energy consumption during the 
course of the cabin pull-down, and transient performance factor. It can be 
observed that time to comfort increases by about 22 seconds per additional liter 
of coolant. At the same time, energy consumption increases by about 17 Wh per 
additional liter of coolant, as observed in Figure ?11.13 b). Figure ?11.13 c) shows 
the influence of coolant volume on transient performance factor. It was found that 
each additional liter of coolant decreases TPF by about 0.03.  
While some of the above mentioned results may vary based on system 
configuration and operating conditions, this parametric study provided a 
qualitative answer to the question of influence of coolant volume on transient 
system performance. In summary, it can be stated that small variations in coolant 
charge below 1 to 2 liters may not result in significant changes of thermal 
performance and passenger comfort.  
213 
 
Figure ?11.12: Influence of Coolant Volume on Cabin Warm-up  
(R290 2LP, T30RH50, 12 K soak, 1550 rpm, 250 g/s) 
0
 3
 6
 9
 12
 15
 18
 21
 24
 27
 30
 0 5 10 15
 Te
 m
 p
 e
 ra
 tu
 re
  [
 ?C
 ]
 Time [min]
 Supply Temperature - Warm-up
 2.0 L
 3.0 L
 4.0 L
 5.0 L
 6.0 L
 7.0 L
 8.0 L
 0
 500
 1,000
 1,500
 2,000
 2,500
 3,000
 3, 00
 0 5 10 15
 C
 a
 p
 a
 ci
 ty
  [
 W
 ]
 Time [min]
 Cooling Capacity - Warm-up
 2.0 L
 3.0 L
 4.0 L
 5.0 L
 6.0 L
 7.0 L
 8.0 L
 40
 60
 80
 100
 120
 140
 160
 180
 0 5 10 15
 P
 o
 w
 e
 r 
[W
 ]
 Time [min]
 System Power - Warm-up
 2.0 L
 3.0 L
 4.0 L
 5.0 L
 6.0 L
 7.0 L
 8.0 L
 a)
 b)
 c)
214 
 
Figure ?11.13: Influence of Coolant Volume on Performance Metrics  
(R290 2LP, T35RH50, 12 K soak, 1550 rpm, 250 g/s) 
13.0
 13.5
 14.0
 14.5
 15.0
 15.5
 16.0
 1 2 3 4 5 6 7 8 9
 T
 im
 e
  [
 m
 in
 ]
 Coolant Volume [L]
 Time to Comfort
 580
 600
 620
 64
 660
 68
 700
 1 2 3 4 5 6 7 8 9
 E
 n
 e
 rg
 y
  [
 W
 h
 ]
 Coolant Volume [L]
 Energy Consumption
 1.25
 1.30
 1.35
 1.40
 1.45
 .5
 1 2 3 4 5 6 7 8 9
 T
 P
 F
  [
 -]
 Coolant Volume [L]
 Transient Performance Factor
 a)
 b)
 c)
215 
11.5 Influence of Coolant Concentration on Transient Performance 
Previous research on secondary loop systems by Ghodbane [104] used a 
coolant mass fraction of 32wt% ethylene glycol in water. It was accepted that the 
mixture could freeze in harsh winter conditions, but the pressure would stay 
below the burst pressure of the tubing. The research was performed under the 
assumption that the secondary loop system would be used for cabin cooling in 
the summer only. However, recent developments (Eilemann [132], Battista [133], 
Di Sciullo [134], Petitjean [135], Seccardini [136], etc.) show that secondary loop 
systems (also called indirect systems) may be used for general thermal 
management in future vehicles. Possibly, engine thermal management, thermal 
management of batteries and auxiliary systems, as well as passenger cabin 
thermal management could be combined into one compact, smart thermal 
management system. Since this system would need to be operated in low 
ambient winter conditions as well, a coolant mass fraction of 32wt% ethylene 
glycol may not be permissible. Based on the freezing temperature of water-
 ethylene glycol mixtures with varying concentration, shown in Figure ?11.14, a 
32wt% mixture will freeze at roughly -15?C. For countries with a strong winter, 
coolant concentration may need to be increased to 50wt% or above. In countries 
with a yearlong hot climate, it may be beneficial to reduce coolant concentration 
to 20wt% or below to take advantage of the better heat transfer properties of 
pure water. The R290 2LP Modelica? model was used to investigate the 
influence of coolant concentration on the transient performance of a secondary 
loop system during cabin pull-down and cabin warm-up. 
216 
 
Figure ?11.14: Change of Freezing Temperature of an Aqueous Ethylene Glycol 
Mixture based on Mass Fraction 
 
The simulation was performed under an ambient temperature of 30?C, 
ambient relative humidity of 50%, and an initial soak of 12 K. A compressor 
rpmof 1,550, as well as a coolant mass flow rate of 250 g/s were assumed. After 
the cabin reached a comfort temperature of 24?C, the compressor was turned off 
and the cabin was allowed to warm up. The following paragraphs discuss the 
results of the pull-down portion, as well as the cabin warm-up portion of the 
simulation.  
Figure ?11.15 shows the influence of water-ethylene glycol concentration 
on the cabin pull-down process. It can be observed that at 10% difference in 
mass fraction may result in a 0.8 K difference during the initial minutes of the 
pull-down process. At the end of the pull-down (after ~15 min) the difference in 
supply temperature per 10% change in mass fraction decreases to 0.25 K. 
-90
 -80
 -70
 -60
 -50
 -40
 -30
 -20
 -10
 0
 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
 Te
 m
 pe
 ra
 tu
 re
  [?
 C]
 Coolant Concentration [kg/kg]
 Freezing Temperature
217 
Cooling capacity, shown in Figure ?11.15 b), is not significantly affected by coolant 
mass fraction past the first two minutes into the pull-down process. However, 
system power consumption, shown in Figure ?11.15 c), changes significantly due 
to the change in coolant pump power consumption. As the glycol mass fraction 
increases, viscosity and density increase, resulting in an increase in pressure 
drop and pressure head. At the end of the pull-down process, instantaneous 
system power consumption may be increased by as much as 2.4% per 10% 
increase in coolant mass fraction.   
Figure ?11.16 presents the influence of coolant mass fraction on the warm-
 up process of the cabin during an off-cycle period. As can be observed in 
Figure ?11.16 a), the influence of coolant mass fraction on supply temperature 
during cabin warm-up is nearly constant throughout the first 15 minutes of the off-
 cycle period. Supply temperature increases by about 0.4 K for every 10% 
increase of glycol mass fraction. This is due to the additional capacity available 
for low mass fraction coolants, as observed in Figure ?11.16 b). While the change 
in mass fraction represents a significant change in cooling capacity during the 
later stages of the off-cycle period, when cooling capacity is low, the influence on 
power consumption is significant. Due to the increased pump power at high 
glycol mass fractions, instantaneous off-cycle period power consumption may 
increase -on average- by about 15 W per 10% increase in glycol mass fraction.  
218 
 
Figure ?11.15: Influence of Coolant Concentration on Cabin Pull-down  
(R290 2LP, T30RH50, 12 K soak, 1550 rpm, 250 g/s) 
900
 1,200
 1,500
 1,800
 2,100
 2,400
 2,700
 3,000
 3,300
 0 2 4 6 8 10 12 14 16
 P
 o
 w
 e
 r 
[W
 ]
 Time [min]
 System Power
 X 0.1
 X 0.2
 X 0.3
 X 0.4
 X 0.5
 X 0.6
 X 0.7
 X 0.8
 500
 1,000
 1,500
 2,000
 2,5
 3,000
 3,500
 4,000
 4,500
 5,000
 0 2 4 6 8 10 12 14 16
 C
 a
 p
 a
 ci
 ty
  [
 W
 ]
 Time [min]
 Cooling Capacity
 X 0.1
 X 0.2
 X 0.3
 X 0.4
 X 0.5
 X 0.6
 X 0.7
 X 0.8
 0
 4
 8
 12
 16
 2
 24
 28
 32
 36
 40
 0 2 4 6 8 10 12 14 16
 Te
 m
 p
 e
 ra
 tu
 re
  [
 ?C
 ]
 Time [min]
 Supply Temperature
 X 0.1
 X 0.2
 X 0.3
 X 0.4
  0.5
  0.6
 X 0.7
 X 0.8
 a)
 b)
 c)
219 
 
Figure ?11.16: Influence of Coolant Concentration on Cabin Warm-up  
(R290 2LP, T30RH50, 12 K soak, 1550 rpm, 250 g/s) 
0
 100
 200
 300
 400
 500
 600
 700
 800
 900
 1,000
 0 2 4 6 8 10 12 14 16
 P
 o
 w
 e
 r 
[W
 ]
 Time [min]
 System Power - Warm-up
 X 0.1
 X 0.2
 X 0.3
 X 0.4
 X 0.5
 X 0.6
 X 0.7
 X 0.8
 0
 500
 1, 0
 1,500
 2,000
 2,500
 3,000
 3,500
 0 2 4 6 8 10 12 14 16
 C
 a
 p
 a
 ci
 ty
  [
 W
 ]
 Time [min]
 Cooling Capacity - Warm-up
 X 0.1
 X 0.2
 X 0.3
 X 0.4
  .5
  .6
  .7
  .8
 2
 5
 8
 11
 14
 17
 20
 23
 26
 29
 32
 0 2 4 6 8 10 12 14 16
 Te
 m
 p
 e
 ra
 tu
 re
  [
 ?C
 ]
 Time [min]
 Supply Temperature - Warm-up
 X 0.1
 X 0.2
 X 0.3
 X 0.4
 X 0.5
 X 0.6
 X 0.7
 X 0.8
 a)
 b)
 c)
220 
Figure ?11.17 shows the effect of coolant concentration on typical 
performance metrics for a pull-down process. Time to comfort is nearly linearly 
dependent on the glycol mass fraction and increases by roughly 12 seconds per 
10% increase in mass fraction. As shown in Figure ?11.17 b), a 10% increase in 
coolant mass fraction result in a 18 Wh increase of energy consumption (~2.8% 
relative increase). TPF decreases by about 0.035 for a 10% increase in coolant 
concentration.  
The mass fraction range of interest is marked by a shaded area in 
Figure ?11.17 a) through c). It was found that when changing from a 30% glycol 
mass fraction to a 60% mass fraction time to comfort is expected to increase by 
about half a minute. Energy consumption during a pull-down would be increased 
by about 7.8%, while TPF is expected to decrease by about 0.1. 
It can be concluded that glycol mass fraction has only little influence on 
time to comfort, while it has a more significant influence on thermal comfort 
during off-cycle periods. To save energy and reduce pull-down-time, it would be 
advantageous to reduce glycol mass fraction, possibly below 30%. However, 
increasing mass fraction to 50% or 60% may result in an increase of energy 
consumption of about 8% or less and a small increase in time to comfort.   
221 
 
Figure ?11.17: Influence of Coolant Concentration on Performance Metrics  
(R290 2LP, T30RH50, 12 K soak, 1550 rpm, 250 g/s) 
222 
12 Summary and Future Work 
The present research focused on the performance evaluation of 
secondary loop systems for automotive air-conditioning. The use of efficient and 
inexpensive refrigerants, as well as the potential to reduce A/C energy 
consumption by applying thermal storage alternatives were investigated.  
 
12.1  Research Contributions 
The present research resulted in the following contributions: 
? Development and operation of a dynamic laboratory-scale test bench, 
which allows for testing of drive cycles in changing climates and 
enables the testing of A/C control algorithms 
? Conclusive characterization of transient performance of secondary 
loop systems during pull-down and drive cycle tests 
? Experimental evaluation of Propane in automotive secondary loop air-
 conditioning systems with respect to energy consumption and transient 
performance 
? Quantification of benefits from secondary loop thermal storage during 
short off-cycle periods in drive cycles and prolonged off-cycle cabin 
warm-up 
? Experimental characterization of the benefits of ice storage in 
secondary loop systems 
? Evaluation of alternative control strategies and cabin pre-conditioning 
in electric vehicles with regard to reducing A/C energy consumption for 
secondary loop systems with thermal storage 
223 
? Development of a validated model to perform transient simulation of 
secondary loop systems in Modelica 
 
12.2 Summary of Research Outcomes 
A summary of outcomes for the present research is given as follows: 
? A test facility was built to test direct expansion, as well as secondary 
loop systems. Controls were implemented to allow for drive cycle 
testing and investigation of A/C control strategies. A passenger cabin 
model, first developed in previous research, was modified to include 
typical A/C controls, as well as the ability to use ambient drive cycles 
(temperature, relative humidity, and solar insolation) 
? A secondary loop system using R152a as primary refrigerant was 
tested in steady-state, as well as transient operation. Using the same 
components as a baseline R134a direct expansion system, the R152a 
2LP system showed a reduced power consumption at about the same 
capacity. Steady-state COP was increased by up to 10%. During 
transient pull-down tests, the secondary loop system showed some 
thermal lag, as well as increased power consumption. Although time to 
comfort was increased, the transient performance factor was within 5% 
of R134a DX results. Due to increased thermal mass, the energy 
available to cool the air during an NEDC test decreased by about 10%. 
Due to a slight increase in power consumption (5%), TPF decreased 
by 10%.  
? R290 was investigated as possible refrigerant in the VCC of secondary 
loop systems. Regular R134a components were used in the VCC. 
Compressor speed was adjusted to simulate a smaller compressor and 
reduce high cooling capacity and compressor power consumption. The 
R290 2LP system was found to perform well with an increase in COP 
224 
of 8% in comparison to R134a DX during highway driving operation. 
When idling, COP decreased by 15% from R134a DX. The decrease in 
efficiency was attributed to low compressor isentropic and volumetric 
efficiencies at idling conditions. Due to the adjustment of compressor 
rpm, the comparison of transient accumulated capacity and power 
consumption was somewhat ambiguous. However, TPF for pull-down 
at highway driving conditions was similar or better than R134a DX TPF. 
For NEDC testing, R290 2LP TPF was decreased by 5% with respect 
to R134a DX, showing a better performance than R152a 2LP. It was 
concluded that if the cycle, primarily the compressor, would be 
optimized for R290, the R290 2LP system could achieve a 
performance significantly better than an R134a direct expansion 
system. 
? The thermal storage potential of secondary loop systems and it's 
benefits for thermal comfort and energy consumption were investigated. 
New test procedures were defined for this research. During Start/Stop 
operation R290 2LP TPF was equal to R134a DX TPF.  When tripling 
the off-cycle periods in the cycle (traffic light stops), R290 2LP TPF 
increased by roughly 10% over R134a DX TPF. While preserving 
thermal comfort, the secondary loop system also achieved a better 
thermal performance with increasing length of off-cycle periods. For 
cabin warm-up during extended off-cycle periods, the secondary loop 
system was able to keep the cabin 3 K lower, while cabin 
humidification due to re-evaporation was reduced. 
? The possibility of using ice storage for added thermal comfort and 
reduced power consumption during commutes was evaluated. The 
present research used a custom made heat exchanger in a 15 L ice 
storage tank, which was charged by the R290 vapor compression 
cycle. It was observed that ice storage could reduce time to comfort by 
more than 15%, while reducing the power consumption of the 
225 
compressor. However, for long periods a bypass should be used once 
evaporating temperature decreased to the point where ice storage 
prevents a faster pull-down. During off-cycle periods after a pull-down 
to comfort temperature, the present ice storage box was able to 
sustain thermal comfort and prevent excessive cabin humidification for 
more than 20 minutes.  
? The benefits of alternative compressor controls and cabin pre-
 conditioning with regards to reducing power consumption in electric 
vehicles were investigated. It was found that total energy consumption, 
including re-heating, can be reduced by as much as 60% when 
controlling cabin supply humidity with either compressor speed or by 
cycling. Using ice storage together with alternative control strategies 
showed the most benefits for cyclic operation, rather than continuous 
operation. Cabin pre-conditioning was found to be detrimental for the 
traditional concept of compressor cycling for frost prevention and re-
 heating for temperature control. In contrast, when employing 
alternative control strategies where the compressor cycles to control 
cabin supply humidity while cabin temperature is close to the comfort 
setpoint, energy savings of up to 30% were observed by pre-
 conditioning the cabin.  
? Models of the passenger cabin and the secondary loop system were 
developed in the Modelica language. Models of an automotive direct 
expansion system, the secondary loop system, and the passenger 
cabin were validated with experimental data for steady-state operation, 
as well as transient pull-down and drive cycle operation. The models 
were further used to evaluate the influence of coolant volume and 
coolant concentration on transient system performance.      
 
226 
12.3 Recommendations for Future Work 
Based on experience from the present research, current developments in 
the area of automotive air-conditioning, and insights from reviewed literature, the 
following recommendations are given for possible future work: 
? The automotive air-conditioning industry is moving towards an 
integrative approach, in which the thermal management system of the 
front end is partially integrated with the climate system and auxiliary 
thermal management tasks, such as battery or electronics cooling. It is 
recommended to focus future research on completely indirect systems, 
where both, the A/C condenser, as well as the cabin air cooler are 
cooled by a secondary (indirect) coolant loop. 
? With the alternative cooling strategies discussed in the present 
research, it was possible to reduce total power consumption in electric 
vehicles considerably. Another big step in reducing power consumption 
could be achieved by removing the need of some kind of heater power 
for re-heating. To this end, the indirect loop which cools the A/C 
condenser could be used for re-heating the cabin air. Re-heating for 
summer cooling operation is typically only needed 10+ minutes after 
the air-conditioning is turned on and enough heating capacity is 
available at the condenser. Energy savings and implications of this 
solution should be quantified by experiment.   
? In addition to cooling, heating for winter operation is a challenging task 
for electric vehicles. The use of secondary loop systems for heat 
pumping needs to be investigated. Thermal storage, possibly ice 
storage, could be used to provide a relatively high fixed temperature 
for the heat pump when operating in sub zero ambient conditions. 
Transient simulations, as well as experiments should be performed to 
quantify benefits and challenges 
227 
? Thermal storage may be a key technology in electric vehicles, both for 
summer time cooling, as well as winter time heating. In the present 
research, the use of ice storage was investigated, using a custom 
made heat exchanger. A second generation heat exchanger, optimized 
for the task at hand could improve the performance of such a system. 
While ice as a phase change material is conveniently available, well 
understood and ready to use, the present research should be extended 
to include other phase change materials, which can be integrated more 
seamlessly into the existing thermal management system.  
? The present cabin passenger model is a lumped model, working with 
energy and mass balances. A more sophisticated, multi-dimensional 
model of the passenger cabin would open the door to thermal comfort 
research. The use of indirect cooling loops in the summer, as well as in 
the winter would lend itself to the operation of radiant heat exchangers 
in the cabin, as well as seat cooling and heating. Numerous studies in 
the residential and commercial air-conditioning area show that the use 
of radiant hydronic systems increases energy efficiency and reduces 
the need for extremely low (or high) cabin air temperatures. A multi-
 dimensional model of the cabin in Dymola, as well as a thermal 
comfort model will be needed for this research. 
 
228 
Appendix A Specification of Test Facility Components 
 
Appendix A provides the specifications of the main components, used in 
the experimental test facility. Parameters and dimensions are given to the best 
knowledge of the author. Heat exchanger outer dimensions were measured 
where possible, and inner dimensions were estimated based on outer 
dimensions. Specifications of compressor, pump, and valve are given based on 
manufacturer's data. 
Table ?A.1 provides specifications for the automotive evaporator used only 
in the direct expansion baseline system. The evaporator was a plate and fin heat 
exchanger with louver fins and unknown inner geometry.  
Table ?A.1: Evaporator Specifications 
Parameter Magnitude Unit 
Number of passes 2 [-] 
Number of tubes per pass 20 [-] 
Number of ports per tube* 8 [-] 
Tube pitch 0.01358 [m] 
Tube length 0.2125 [m] 
Tube depth 0.035 [m] 
Tube height 0.00333 [m] 
Fin depth 0.035 [m] 
Fin density 12.9 [1/in] 
Fin thickness 0.00008 [m] 
Port hydraulic diameter* 0.001814 [m] 
Inner Volume 0.000711 [m3] 
*values estimated based on outer dimensions 
 
Table ?A.2 provides specifications for the automotive condenser used in all 
system configurations throughout this study. The condenser was a micro channel 
229 
heat exchanger with louver fins and unknown port geometry. It did not feature an 
integrated receiver-dryer, as some other models do.  
Table ?A.2: Condenser Specifications 
Parameter Magnitude Unit 
Number of passes 1 [-] 
Number of tubes per pass 35 [-] 
Number of ports per tube* 12 [-] 
Tube pitch 0.01095  [m] 
Tube length 0.66 [m] 
Tube depth 0.0185 [m] 
Tube height 0.002 [m] 
Fin depth 0.0185 [m] 
Fin density 17 [1/in] 
Fin thickness 0.00008 [m] 
Port hydraulic diameter* 0.00075 [m] 
Inner Volume 0.000292 [m3] 
*values estimated based on outside dimensions 
 
Table ?A.3 provides specifications for the intermediate heat exchanger 
used in the secondary loop system. The intermediate heat exchanger was a 
commercially available plate type heat exchanger with unknown inner geometry. 
Table ?A.3: Intermediate Heat Exchanger Specifications 
Parameter Magnitude Unit 
Rated capacity (R134a) 6.1 [kW] 
Rated pressure drop 7.6/3.7 [kPa] 
Number of channels 11/12 [-] 
Refrigerant-side volume 0.000935 [m3] 
Refrigerant-side volume 0.00102 [m3] 
 
Table ?A.4 provides specifications for the cooler used in the secondary loop 
system. The cooler was a micro channel heat exchanger with louver fins and 
unknown port geometry.  
230 
Table ?A.4: Cooler Specifications 
Parameter Magnitude Unit 
Number of passes 2 [-] 
Number of tubes per pass 36 [-] 
Number of ports per tube* 12 [-] 
Tube pitch 0.00675 [m] 
Tube length 0.216 [m] 
Tube depth 0.028 [m] 
Tube height 0.0015 [m] 
Fin depth 0.028 [m] 
Fin density 18 [1/in] 
Fin thickness 0.00009 [m] 
Port hydraulic diameter* 0.00123 [m] 
Inner Volume 0.000877 [m3] 
*values estimated based on outside dimensions 
 
Table ?A.5 provides specifications for the compressor used in all system 
configurations throughout this study. The compressor is a variable speed, fixed 
displacement compressor. The efficiency information given in the table was back 
calculated from experimentally measured data and is a rounded average over a 
several ambient conditions and compressor speeds. 
Table ?A.5: Compressor Specifications 
Parameter Magnitude Unit 
Displacement  75 [cm3] 
Maximum speed 6500 [rpm] 
Volumetric efficiency* 0.9 [-] 
Isentropic efficiency* 0.8  [-] 
*average values, determined from experimental measurement 
 
Table ?A.6 provides the specifications for the electronic expansion valve 
used in all system configurations throughout this study.  
231 
Table ?A.6: Expansion Valve Specifications 
Parameter Magnitude Unit 
Rated evaporating temperature  4.44 [?C] 
Rated differential pressure 650 [kPa] 
Rated Capacity (R134a) 7.53 [kW] 
Number of steps 500 [-] 
 
Table ?A.7 provides specifications for the coolant pump used in the 
secondary loop system. The coolant pump is a  gear pump.  
Table ?A.7: Coolant Pump Specifications 
Parameter Magnitude Unit 
Maximum differential pressure  690 [kPa] 
Motor rated power 0.33 [hp] 
Motor rated rpm 1725 [rpm] 
 
232 
Appendix B Control Flow Charts 
Appendix B provides flow charts for the controls of the facility and facility 
components. 
Figure ?B.1 shows the flow chart of the Cabin Model. The flow chart 
describes the various controls embedded in the Cabin Model, including the 
option to use pre-defined ambient drive cycles, the option to use a virtual heater 
core to artificially raise supply temperature, and the option to either turn off or 
enable the continuous measurement of evaporator/cooler air flow rate after the 
compressor is turned off (used during off-cycle testing). The Cabin Model is 
turned on and off by a switch, embedded in the LabVIEW GUI of the Main DAQ 
Ctrl. 
A more detailed description of the virtual heater core is given in the 
thermostat control flow chart, Figure ?B.2. The flow chart describes the operation 
of the virtual heater core, once thermostat control was activated. The heater is 
typically only used when the compressor is running, otherwise the original supply 
temperature is not altered. If heating functionality is needed to preserve comfort 
temperature, an artificially heightened supply temperature signal, controlled by a 
software PID control, is fed to the Cabin Model equations.  
233 
 
Figure ?B.1: Cabin Model Flow Chart 
234 
 
Figure ?B.2: Thermostat Control Flow Chart 
 
The drive cycle module, shown in Figure ?B.3, uses PID control to operate 
components of the test facility. Since the setpoint profile is known, setpoints of 
future time steps can be used to control components more accurately (feed 
forward control). The output signal can be confined to a minimum and maximum 
magnitude to stay within safe operating conditions (e.g.: to prevent a full closing 
of the EXV). The drive cycle module moves forward to the next entry in the 
respective drive cycle file, until the last entry of the profile. 
235 
 
Figure ?B.3: Drive Cycle Control Flow Chart 
 
Two electronic expansion valves were used in the test facility to control 
superheat. The EXV flow chart, shown in Figure ?B.4, provides information about 
control on the top level (one valve used vs. both valves used), and on the valve 
level. Valves can be controlled either manually, by emergency override, or by 
PID control, which is the usual operation mode. Manual control, as well as the 
236 
emergency override function are used mainly during tuning of the PID control to 
prevent erratic behavior, which could damage the compressor. To prevent loss of 
superheat for an extended time, a sub function, which reduces EXV opening in 
pre-defined increments, was added.  
 
Figure ?B.4: Electronic Expansion Valve Flow Chart 
237 
Figure ?B.5 shows the flow chart of the coolant pump control. The pump 
can either be controlled manually, or by PID control. In the latter case, the control 
variable can be either MFR, or cabin temperature. During temperature control 
mode, the pump can be cycled on and off to keep cabin temperature close to the 
comfort setting. Similarly, pump speed can be reduced using PID control to 
reduce power consumption as much as possible, while cabin temperature is 
close to or lower than comfort setpoint. 
Figure ?B.6 shows the flow chart for the control of compressor clutch and 
compressor rpm. The compressor can be controlled in seven ways, six of which 
are shown in Figure ?B.6. The seventh is drive cycle control, with a pre-
 determined compressor speed and clutch actuation profile. The first operation 
mode is manual operation, in which the compressor rpm and the clutch actuation 
can be chosen from the GUI of the Main DAQ Ctrl. The pump temperature 
control mode turns off the compressor clutch, precluding the VCC from cooling 
the cabin. Pull-down off-cycle mode turns off the compressor clutch, once a pre-
 defined comfort temperature is reached in the passenger cabin. The temperature 
in the cabin will subsequently rise, without the compressor turning back on. 
Compressor defrost cycling, relative humidity cycling, and continuous 
compressor control are modes used for air-conditioning controls research. All 
three modes of operation are used to keep the cabin at a stable temperature, 
once comfort conditions have been achieved. In relative humidity cycling and 
continuous mode, relative humidity is controlled in addition to cabin temperature. 
238 
 
Figure ?B.5: Coolant Pump Control Flow Chart 
239 
 
Figure ?B.6: Compressor Control Flow Chart 
 
The controls outlined in Figure ?B.6 are shown in detail in the following flow 
charts. 
Figure ?B.7 shows the manual control flow chart. In the Main Ctrl DAQ GUI, 
the user can set the desired rpm, as well as turn on and off the compressor 
clutch.  
240 
 
Figure ?B.7: Manual Compressor Control Flow Chart 
 
Figure ?B.8  shows the control chart for temperature control mode of the 
pump. This control mode is used when only ice storage is used to cool the cabin. 
Consequently, the compressor clutch remains turned off while in this mode and 
the thermostat function remains disabled. 
 
Figure ?B.8: Pump Temperature Control Flow Chart 
 
241 
The pull down control, shown in Figure ?B.9, enables the recording of off-
 cycle pull down data. The control turns off the compressor clutch, once comfort 
temperatures are achieved during a pull down test. A flag is set to prevent the 
compressor from turning on once the cabin temperature starts increasing during 
the off cycle. 
 
Figure ?B.9: Compressor Pulldown Control Flow Chart 
 
Figure ?B.10 shows the defrost cycling control chart. Once air-side 
evaporator outlet temperature decreases below a certain set point, the 
compressor shuts odd to prevent frosting. 
242 
 
Figure ?B.10: Defrost Cycling Control Flow Chart 
 
Figure ?B.11 shows the flow chart for relative humidity cycling control. If 
cabin temperature gets close to comfort temperature, the compressor starts 
controlling relative humidity of the cabin supply air stream, rather than prioritizing 
to pull the cabin temperature down further. The thermostat control is used to 
prevent cabin temperature from dropping below comfort setpoint. If the cabin 
temperature is close to comfort conditions and the relative humidity of the supply 
air is low enough, the compressor is turned off to save energy. 
243 
 
Figure ?B.11: Relative Humidity Cycling Control Flow Chart 
 
The continuous relative humidity control, shown in Figure ?B.12, controls 
the compressor speed and coolant pump speed, using a PID control, once cabin 
temperature is close to setpoint. Thermostat control is used to retain cabin 
temperature close to the comfort setting.  
244 
 
Figure ?B.12: Relative Humidity Continuous Control Flow Chart 
 
245 
Appendix C Drive Cycle Development 
?Appendix C provides more details how the conversion of the NEDC drive 
cycle (and similar drive cycles) from vehicle speed to compressor speed. In 
addition, the implementation of drive cycle into LabVIEW, and the development 
of a novel drive cycle to showcase the versatility of the test facility, are discussed.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        
development of drive cycles and drive cycle control in the data acquisition system, 
which was performed to gain the ability to run drive cycle, such as the NEDC and 
the I-95 cycle. 
The NEDC drive cycle, as well as other drive cycles, such as the USA 
urban or extra-urban cycle, are given by a vehicle speed versus time profile. In 
the current study MAC components are tested on a test bench, while the vehicle 
is replaced by a cabin model. This results in the need to convert the original drive 
cycle profiles to compressor rpm versus time profiles. Table ?C.1 shows 
specifications for a medium size, 2.2L engine 5-speed manual car. The shown 
parameters are used to convert vehicle speed to engine rpm, based on gear 
choice. 
246 
Table ?C.1: Automobile Parameters for Speed to RPM Conversion 
Gear 
Gear Ratio 
(5-speed manual) 
1. 3.58 
2. 2.02 
3. 1.35 
4. 0.98 
5. 0.69 
Final Drive Ratio 3.94 
Pulley Ratio 0.786 
  
Tire Specifications 
Tire Type P195/70R14 
Rolling Circumference [m] 1.9748 
 
Figure ?C.1 shows the conversion results in form of graphs. For each gear, 
a linear relationship between vehicle speed and A/C compressor rpm was 
determined. The conversion equations, shown next to the respective gear curves, 
were used in the conversion fo the NEDC drive cycle and the I-95 cycle. 
 
Figure ?C.1: Gear Equations for vehicle speed to compressor rpm conversion 
y = 154.31x - 3E-12
 y = 87.067x
 y = 58.188x + 2E-12
 y = 42.24x
 y = 29.741x
 0
 1000
 2000
 3000
 4000
 5000
 6000
 7000
 8000
 9000
 10000
 11000
 0 20 40 60 80 100 120 140
 A/
 C
  C
 om
 pr
 ess
 or
  S
 pee
 d 
[r
 pm
 ]
 Linear Speed [kph]
 Compressor Speed vs. Linear Speed, Chevy Cavalier 2004
 Gear 1
 Gear 2
 Gear 3
 Gear 4
 Gear 5
247 
To make above conversion equations useful, gear selection at every point 
during the drive cycle needs to be known. To this end, an urban drive cycle and 
an urban shift cycle, based on EPA data [110], were superimposed. Figure ?C.2 
shows gear selection, based on superposition of the two cycles, which allows the 
correlation of speed and gear choice. 
 
Figure ?C.2: Superimposed EPA FTP Drive Cycle and Shift Cycle 
 
However, the superposition reveals only a gear selection at a specific 
speed along the drive cycle and not the exact speed at which gears are shifted. 
Since gear shifts in practice vary between acceleration and deceleration, shift 
points need to be chosen according to driving situation. During acceleration, the 
driver usually shifts up at higher rpm, while during deceleration the driver usually 
shifts down at lower rpm. This was taken in consideration in Table ?C.2. 
Table ?C.2: Shift Point Table for Vehicle Acceleration and Deceleration 
Gear 
Gear Selection 
(EPA) [kph] 
Accel.Shift 
Pt. [kph] 
Decel. Shift 
Pt. [kph] 
1. 24 22.5 16 
2. 40 38.5 32 
3. 64 62.5 54 
4. 72 71 64 
5. - - - 
0
 10
 20
 30
 40
 50
 60
 0
 1
 2
 3
 4
 5
 6
 0 200 400 600 800 1000 1200 1400
 Targ
 et S
 pee
 d (m
 ph)
 Gea
 r
 Time [s]
 EPA - Urban Shift Cycle
248 
The resulting compressor rpm profile for the NEDC drive cycle is shown in 
Figure ?C.3, along with the original vehicle speed profile for comparison. Idle 
compressor speed is 850 rpm, while maximum compressor speed increases up 
to 3600 rpm. At shift points, changes in compressor rpm can be abrupt and result 
in severe transients. To prevent excessive load on the electric motor, which 
drives the compressor, the rpm profile was smoothed to prevent excessive 
transients.  
 
Figure ?C.3: NEDC Compressor Speed Profile 
 
To showcase the abilities of the test facility control (and initially for truck 
anti-idling research), a new drive cycle, the I-95 drive cycle was developed. The 
goal was to assess air-conditioning performance in a more realistic, but highly 
0
 700
 14 0
 210
 8 0
 350
 42 0
 0
 2
 40
 6
 80
 10
 20
 0 200 400 600 800 1000 1200
 Compressor Speed
  [rpm]Target Speed [kph]
 Time [s]
 New European Drive Cycle (NEDC)
 NEDC Vehicle Speed [kph]
 Compressor Spe d [rpm]
 0
 500
 1000
 1500
 2000
 2500
 3000
 3500
 4000
 0 200 400 600 800 1000 1200
 Comp
 resso
 r Spee
 d [rpm
 ]
 Time [s]
 NEDC - Compressor Speed (smoothed)
249 
variable setting, which includes multiple climate zones. The I-95 cycle simulates 
a car driving down the I-95 corridor from Maine to Florida on the United States 
east coast. Climate data along the route was acquired in form of TMY2 data from 
the National Renewable Energy Laboratory (NREL) [111]. As TMY2 data was 
available only for a limited number of cities, the respective cities were used to 
pinpoint locations along the route. Google maps was used to determine 
distances between locations and travel time during regular traffic hours. 
Table ?C.3 provides the I-95 route locations, as well as distances and average 
speed in between locations. 
Table ?C.3: I-95 Cycle - Route and Travel Information 
# City (State) 
Travel 
time [s] 
Distance 
[km] 
Avg. Speed 
[km/h] 
1 Caribou (ME) 18600 481.2 93.1 
2 Portland (ME) 7800 172.2 79.5 
3 Boston (MA) 3900 80.8 74.6 
4 Providence (RI) 7800 196.3 90.6 
5 Bridgeport (CT) 5400 97.2 64.8 
6 New York City (NY) 1800 16.7 33.5 
7 Newark (NJ) 6300 147.4 84.2 
8 Philadelphia (PA) 2100 46.5 79.7 
9 Wilmington (DE) 5400 112.3 74.9 
10 Baltimore (MD) 20400 495.7 87.5 
11 Raleigh (NC) 8100 210.8 93.7 
12 Wilmington (NC) 13200 331.5 90.4 
13 Columbia (SC) 10200 255.9 90.3 
14 Savannah (GA) 9300 223.7 86.6 
15 Jacksonville (FL) 6300 143.6 82.0 
16 Daytona Beach (FL) 12300 318.6 93.3 
17 West Palm Beach (FL) 4800 114.3 85.7 
18 Miami (FL) - - - 
 
For simplicity, it was assumed that the driver would start at 8:00am every 
morning and drive until 12:00pm. There would be a rest time of one hour during 
which the engine would be turned off. Following the rest time, the driver would 
250 
drive from 1:00pm through 9:00pm. The driver would then stop for the night to 
take 11 hours rest. The average speed can vary in between cities, based on 
speed limits and traffic situation that is usually encountered in the respective 
locations. 
Dry bulb temperature, relative humidity, and solar insolation profiles were 
determined using TMY2 weather data. TMY2 weather data is available as hourly 
data for one location. The data is available averaged over the last 30 years, for 
the chosen month at the specified day, hour, and location. As the data is 
available only for a limited number of locations along the route, the data needs to 
be interpolated in steps of one hour between locations. This creates transient 
profiles which change temporally and spatially. The developed profiles for dry 
bulb temperature, relative humidity, and solar insolation, are shown in Figure ?C.4 
a) through c). The profiles were used as drive cycle input to the cabin model, and 
the controls of the environmental chamber. 
During each start and stop of the car, the car was assumed to leave the I-
 95 to stop at a restaurant or inn. One set of the urban drive cycle portion of the 
NEDC was chosen to be performed during each start and stop operation. The 
compressor speed profile was developed with the method outlined above. The 
clutch profile follows the starts and stops for rest and night stops.  
Operation of one complete I-95 cycle lasted about 88 h, which is the time 
a car would need to drive from Caribou, Maine, to Miami, Florida, under the 
251 
assumptions previously mentioned. The test facility was controlled in time steps 
of 2 seconds. 
 
Figure ?C.4: I-95 Cycle TMY2 Interpolated Climate Data 
 
Figure ?C.5 shows performance parameters which were measured during 
the I-95 cycle test, such as cooling capacity and compressor power. It can be 
observed that cooling capacity is smallest, around 1.65 kW in the colder climate 
of Maine, during the beginning of the first day. The capacity remains nearly 
constant at about 1.9 kW during the remaining days of the I-95 test. Compressor 
power increases with ambient temperature. At the same time, compressor power 
follows the compressor speed profile. Compressor power in Maine was on 
0
 5
 10
 15
 20
 25
 30
 35
 0 20 40 60 80
 Temp
 eratu
 re [?C
 ]
 Time [h]
 I-95 - Dry Bulb Temperature Profile
 Dry Bulb Temp [?C]
 0
 10
 20
 30
 4
 50
 60
 7
 80
 9
 100
 0 20 40 60 80
 Rela
 tive H
 umid
 ity [%
 ]
 Time [h]
 I-95 - Relative Humidity Profile
 Relative Humidity [%]
 0
 100
 200
 300
 400
 500
 600
 700
 800
 900
 1000
 0 20 40 60 80
 Diffu
 se R
 adiat
 ion [
 Wh/m
 ^2]
 Time [h]
 I-95 - G obal Horizontal Radiation Profile
 Global Radiation [%]
 a) b)
 c)
252 
average about 0.75 kW, while the average compressor power was about 1.25 
kW in Florida. 
Figure ?C.6 shows the profile of cabin temperature during the entire I-95 
cycle. The ambient temperature, as well as the solar insolation, are shown for 
better explanation of the cabin temperature behavior. It can clearly be observed 
that the cabin model was able to control the cabin temperature at a comfort 
setting around 25?C for the entire on-cycle portion of the I-95 test. During 
afternoon stops and early evening stops, the cabin soaks due to solar insolation. 
This can most prominently be observed during the last afternoon stop in Florida, 
around 4,600 minutes into the test, where the strongest soaking was determined 
to be about 14 K. At night, when the A/C is turned off, the cabin temperature 
follows the outdoor temperature. This can be observed well at around 1,000 
minutes or 2,500 minutes.  
Cabin relative humidity is shown in Figure ?C.7. The relative humidity 
remains constant at about 32% after a short start-up period. At night the cabin 
relative humidity increases up to a relative humidity of 53%, due to the decrease 
in cabin temperature and the exchange with colder ambient air. During afternoon 
stops, the relative humidity decreases, due to the soaking of the cabin. 
From above results it can be concluded that the existing test facility is 
capable of controlling all parameters relevant to capture transient performance of 
a MAC system for a driving vehicle, including varying climate conditions and day 
and night cycles.  
253 
 
Figure ?C.5: I-95 Cycle - Cooling Capacity and Power Consumption 
 
 
Figure ?C.6: I-95 Cycle - Cabin Temperature 
 
 
Figure ?C.7: I-95 Cycle - Cabin Relative Humidity 
254 
Appendix D Steady-state Results Summary: Secondary Loop 
Versus Direct Expansion 
?Appendix D provides a summary of steady-state experimental results for 
the comparison of secondary loop system to direct expansion system, using 
R152a as refrigerant.  
Figure ?D.1 shows steady-state trends for relative magnitude of 
performance metrics over a range of ambient temperatures and relative 
humidities for idling and highway driving conditions, with respect to R134a DX 
magnitudes. The relative humidities which accompany respective ambient 
temperatures is shown in Table ?7.1.  
Figure ?D.1 a) shows variation of cooling capacity, while Figure ?D.1 b) 
shows variation of power consumption, and Figure ?D.1 c) shows variation of 
cycle COP.  As a general trend, cooling capacity of both, R152a DX and R152a 
2LP are similar to R134a DX at highway speed. During idling, cooling capacity 
decreases by about 5% to 10% for R152a 2LP systems. A comparison of power 
consumption shows that R152a DX consumes 10% to 15% less power compared 
to R134a DX at highway conditions. 2LP power consumption is increased, due to 
additional power consumption by the coolant pump. This is especially felt during 
idling at low temperatures, as the compressor consumes the least power in these 
conditions and pump power is a higher fraction of total power. However, power 
consumption does not rise above R134a DX power consumption even at low 
temperature idling. COP of R152a DX is generally higher than for R152a 2LP 
255 
systems. R152a 2LP COP is typically increased by 5% to 10% above R134a DX, 
except for idling at low ambient temperatures. 
Figure ?D.2 provides a summary of sensible and latent performance for a 
range of ambient conditions. From Figure ?D.2 a) it can be observed that sensible 
cooling capacity of the R152a DX and 2LP systems are close to the capacity of 
the R134a DX system. At idling, sensible capacity of the R152a 2LP system 
seems to be slightly increased to the R152a DX system. In comparison to the 
R134a DX system, the secondary loop system seems to have a similar sensible 
capacity at moderate temperatures, while it is reduced by more than 10% at low 
and high ambient temperature. Latent capacity, shown in Figure ?D.2 b), is less 
conclusive, as high measurement uncertainties apply. However, at highway 
speeds, the latent capacity of R152a 2LP is similar to R134a DX capacity. At 
idling, trends seem to be unclear, except of at high ambient temperature of 45?C, 
where latent capacity is zero for all systems. A summary of sensible heating 
factor (SHF) values, not commonly reported in the automotive area, is shown in 
Figure ?D.2 b). Both for R152a DX, as well as for R152a 2LP, the SHF stays 
within 5% of its value for R134a DX, except of during idling at low ambient 
temperatures.  
R152a compressor volumetric efficiencies, both for the DX and the 2LP 
system, shown in Figure ?D.3 a), are observed to be higher than R134a DX at 
highway speeds. At idling speeds, R152a DX volumetric efficiencies are the 
same as R134a DX efficiencies, while R152a 2LP efficiencies are up to 10% 
decreased at low ambient temperatures. 
256 
 
Figure ?D.1: Steady-state Performance Metrics Summary (2LP Versus DX) 
257 
 
Figure ?D.2: Steady-state Sensible/Latent Performance Summary (2LP Versus DX) 
-20
 -15
 -10
 -5
 0
 5
 10
 15
 20
 T15       
idle
 T25       
idle
 T35       
idle
 T45       
idle
 T15 
hwy
 T25 
hwy
 T35 
hwy
 T45 
hwy
 S
 H
 F
  [
 %
 ]
 Ambient Temperature [?C] ; Compressor Speed [rpm]
 R134a DX R152a DX R152a 2LP
 Sensible Heat Factor (% R134a DX)
 -60
 -40
 -20
 0
 20
 40
 60
 T15       
idle
 T25       
idle
 T35       
idle
 T45       
idle
 T15 
hwy
 T25 
hwy
 T35 
hwy
 T45 
hwy
 C
 a
 p
 a
 c
 it
 y
  [
 %
 ]
 Ambient Temperature [?C] ; Compressor Speed [rpm]
 R134a DX R152a DX R152a 2LP
 Latent Capacity Comparison (% R134a DX)
 -20
 -15
 -10
 -5
 0
 5
 10
 15
 20
 T15       
idle
 T25       
idle
 T35       
idle
 T45       
idle
 T15 
hwy
 T25 
hwy
 T35 
hwy
 T45 
hwy
 S
 H
 F
  [
 %
 ]
 Ambient Temperature [?C] ; Compressor Speed [rpm]
 R134a DX R152a DX R152a 2LP
 Sensible Heat Factor (% R134a DX)a)
 b)
 c)
258 
Isentropic efficiencies, shown in Figure ?D.3, are increased by about 10% 
when using R152a as drop-in in the direct expansion system during highway 
driving. An increase of about 20% can be observed for the R152a secondary 
loop system. During idling, the R152a DX system shows an increase of up to  
10% in isentropic efficiency, decreasing with ambient temperature. The R152a 
2LP system showed isentropic efficiencies similar to R134a DX at low ambient 
temperature and increases of about 10% at higher ambient temperature.   
 
Figure ?D.3: Steady-state Compressor Efficiency Summary (2LP Versus DX) 
-30
 -20
 -10
 0
 10
 20
 30
 T15       
idle
 T25       
idle
 T35       
idle
 T45       
idle
 T15 
hwy
 T25 
hwy
 T35 
hwy
 T45 
hwy
 Ef
 fi
 ci
 en
 cy
  [
 %
 ]
 Ambient Temperature [?C] ; Compressor Speed [rpm]
 R134a DX R152a DX R152a 2LP
 Isentropic Efficiency (% R134a DX) 
-30
 -20
 -10
 0
 10
 20
 30
 T15       
idle
 T25       
idle
 T35       
idle
 T45       
idle
 T15 
hwy
 T25 
hwy
 T35 
hwy
 T45 
hwy
 Ef
 fi
 ci
 en
 cy
  [
 %
 ]
 Ambient Temperature [?C] ; Compressor Speed [rpm]
 R134a DX R152a DX R152a 2LP
 Volumetric Efficiency (% R134a DX) 
a)
 b)
259 
Appendix E Transient Test Procedure Illustrations 
?Appendix E provides illustrations of all transient test procedures used in 
the present study. The illustrations show profiles of controlled parameters, such 
as compressor speed and heater core load, as well as their effect on cabin 
temperature and relative humidity. All illustrations show the progress of time over 
the length of a test on the x-axis. The magnitude of parameters is not identified at 
the axis, as the illustrations are meant to merely facilitate understanding of test 
procedures and resulting trends in cabin parameters.  
Figure ?E.1 shows the test procedure for a pull-down test. Pull-down tests 
are performed at fixed compressor speeds and fixed condenser air flow rates, 
corresponding to an automobile driving at a steady speed. The compressor is 
turned on at time zero, providing cooling capacity to pull down the cabin to a pre-
 defined comfort temperature. Once the cabin has reached comfort temperature, 
the test is stopped. 
260 
 
Figure ?E.1: Pull-down Test Procedure Illustration 
 
Figure ?E.2 provides an illustration of the New European Drive Cycle 
(NEDC) test procedure. The NEDC is a fixed profile of vehicle speed over time. 
The vehicle speed profile was converted into the compressor speed profile (rpm), 
shown in the figure. The condenser air flow rate still follows the original outline of 
the vehicle speed profile, since it is directly correlated with vehicle speed. The 
general trends of supply and cabin temperature and their fluctuation with 
compressor speed are shown.   
261 
 
Figure ?E.2: NEDC Test Procedure Illustration 
                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   
To test thermal storage characteristics of secondary loop systems, off-
 cycle test procedures were developed. The pull-down off-cycle test, shown in 
Figure ?E.3, is a variation of the original pull-down test. Once cabin temperature 
corresponds to the comfort setting, the compressor is turned off and the cabin is 
allowed to heat up again. Temperature and relative humidity trends are recorded 
to allow the comparison between secondary loop and direct expansion systems.  
 
262 
 
Figure ?E.3: Pull-down Off-cycle Test Procedure Illustration 
 
A second test procedure for off-cycle testing, the Start/Stop cycle, is 
shown in Figure ?E.4. The Start/Stop cycle is an NEDC, modified to the extent that 
idling sections were replaced by compressor off sections. This simulates the 
start/stop automatic mechanism, used in modern cars, which turns off the engine  
instead of idling when waiting for a  traffic light.  
263 
 
Figure ?E.4: Start/Stop Test Procedure Illustration 
 
The start/stop cycle was further modified to increase engine-off time. 
Figure ?E.5 shows the Start/Stop three times idling (SS3xI) test procedure. The 
engine-off times were tripled with respect to the Start/Stop engine-off times (and 
NEDC idling sections), thereby introducing greater fluctuations in supply 
temperature and relative humidity. The thermal storage advantage of secondary 
loop systems can be observed better with increasing engine-off times. 
264 
 
Figure ?E.5: Start/Stop 3x Idling Test Procedure Illustration 
 
Figure ?E.6 through Figure ?E.8 introduces three test procedures for A/C 
controls research. Figure ?E.6 shows the procedure for frost cycling tests. A 
regular pull-down procedure decreases cabin temperature to comfort setting. 
After comfort temperature is achieved, the compressor continues to run, while a 
heater core is employed to keep the cabin at comfort temperature. Since the 
compressor is still running, supply temperature will eventually decrease until 
evaporator outlet temperature is close to frost condition. At this point the 
265 
compressor is cycled to prevent frosting. As a result, power consumption of 
compressor and heater are high, while supply relative humidity is low. 
 
Figure ?E.6: A/C Ctrl - Frost Cycling Test Procedure Illustration 
 
266 
Figure ?E.7 shows the test procedure for relative humidity cycling control. 
The relative humidity cycling control procedure is similar to the frost cycling 
procedure, though the cycling trigger is changed. Instead of cycling at the onset 
of frost at the evaporator, the compressor is mainly turned off when the cabin 
temperature is close to comfort setting. As long as relative humidity of the supply 
air does not increase above a pre-defined setpoint, the compressor remains 
turned off. This allows to expend less heater power and reduces power 
compressor consumption, though fluctuations in supply temperature and relative 
humidity are increased. 
Figure ?E.8 introduces the continuous relative humidity control procedure. 
Similar to the relative humidity cycling, the compressor is used to control relative 
humidity. However, instead of cycling, compressor speed is controlled by a PID 
controller to keep supply relative humidity at a pre-defined setpoint, as long as 
cabin temperature does not warrant higher compressor speeds to gain additional 
cooling capacity.   
267 
 
Figure ?E.7: A/C Ctrl - RH Cycling Test Procedure Illustration 
 
268 
 
Figure ?E.8: A/C Ctrl - RH Continuous Test Procedure Illustration 
 
 
 
269 
Appendix F Steady-state Results Summary: Propane 2LP 
?Appendix F provides a summary of R290 2LP steady-state results over a 
range of ambient temperatures for idling, as well as highway driving conditions 
with respect to R134a DX. Figure ?F.1 shows am summary of cooling capacity, 
total power consumption, and transient performance factor (TPF), while 
Figure ?F.2 focuses on sensible versus latent performance, and Figure ?F.3 
provides information on compressor efficiencies. 
Figure ?F.1 a) shows the cooling capacity of R290 2LP with respect to 
R134a DX. When operating under highway driving conditions, the R290 2LP 
capacity was decreased by less than 10%. During idling, the capacity was about 
20% decreased. As discussed in Chapter ?8, R290 2LP compressor speed was 
reduced for both, idling and highway driving, which decreased cooling capacity. 
Therefore, cooling capacity results can be varied actively by adjusting 
compressor rpm. In an automobile, a compressor with a different displacement 
might be used. A comparison of power consumption, shown in Figure ?F.1 b), 
shows that power consumption was up to 16% reduced at high speed conditions. 
At idling conditions, power consumption was reduced up to 11%. COP of R290 
2LP, shown in Figure ?F.1 c), was increased by up to 9% at highway driving and 
moderate ambient temperatures. At idling conditions, power consumption was 
reduced between 10% and 15%. Although uncertainty during idling is increased,  
it can be observed that R290 2LP COP is generally decreased as compared to 
R152a 2LP COP. During highway driving, COP of R290 2LP is similar or higher 
than COP of R152a 2LP.   
270 
 
Figure ?F.1: Steady-state Performance Metrics Summary (Propane 2LP) 
271 
Figure ?F.2 shows a relative comparison of sensible and latent capacity, as 
well as the sensible heating factor (SHF). Figure ?F.2 a) shows a comparison of 
accumulated sensible capacity. It can be found that R290 2LP sensible capacity 
is similar to R152a sensible capacity. The impact factor for is the secondary loop 
and the need to sensibly cool the additional thermal mass. At low temperature 
idling conditions, sensible capacity is reduced by 25%, though measurement 
uncertainty prevents a clear cut conclusion. Figure ?F.2 b) shows a comparison of 
accumulated latent capacity. Latent capacity is associated with high 
measurement uncertainties, especially at idling conditions, since absolute latent 
capacity is very low. In general latent capacity at highway driving conditions is 
similar to latent capacity of R134a DX. A clear trend cannot be observed at idling 
conditions. Figure ?F.2 c) shows a comparison of SHF. At highway driving 
conditions, SHF of R134a DX, R152a 2LP and R290 2LP compare with each 
other within 5%. At idling speed, R290 2LP latent capacity is increased by 10% at 
moderate ambient temperatures, while it is slightly decreased at low ambient 
temperatures.  
Figure ?F.3 shows a comparison of volumetric and isentropic compressor 
efficiencies at different ambient and driving conditions. In general, it can be 
observed that R290 2LP compressor efficiencies are increased by about 10% at 
highway driving conditions. At idling conditions, both volumetric and isentropic 
efficiencies, are decreased by about 20% at moderate to high ambient 
temperatures. At low ambient temperatures, compressor efficiency decreases 
significantly by more than 30%.  
272 
 
Figure ?F.2: Steady-state Sensible/Latent Performance Summary (Propane 2LP) 
273 
 
Figure ?F.3: Steady-state Compressor Efficiency Summary (Propane 2LP) 
 
 
 
 
 
-50
 -40
 -30
 -20
 -10
 0
 10
 20
 30
 40
 50
 T15       
idle
 T25       
idle
 T35       
idle
 T45       
idle
 T15 
hwy
 T25 
hwy
 T35 
hwy
 T45 
hwy
 Ef
 fi
 ci
 en
 cy
  [
 %
 ]
 Ambient Temperature [?C] ; Compressor Speed [rpm]
 R134a DX R152a 2LP R290 2LP
 Volumetric Efficiency (% R134a DX) 
-50
 -40
 -30
 -20
 -10
 0
 10
 20
 30
 40
 50
 T15       
idle
 T25       
idle
 T35       
idle
 T45       
idle
 T15 
hwy
 T25 
hwy
 T35 
hwy
 T45 
hwy
 Ef
 fi
 ci
 en
 cy
  [
 %
 ]
 Ambient Temperature [?C] ; Compressor Speed [rpm]
 R134a DX R152a 2LP R290 2LP
 Isentropic Efficiency (% R134a DX) 
a)
 b)
274 
Appendix G Glycol Property Equations 
The properties of the secondary working fluid, ethylene glycol, were 
calculated using property equations from M. Conde Engineering [109]. The 
respective equations were used in LabView, as well as in the Modelica glycol 
media package and are restated below for completeness.  
The properties (P) density, thermal conductivity, and specific heat capacity 
can be determined for an aqueous ethylene glycol mixture by Equation (G1), 
while dynamic viscosity and Prandtl number can be calculated with the help of 
Equation (G2). T is the temperature in [?C] at which properties are evaluated, 
while x is the mass fraction of glycol in water. Coefficients used in the following 
equations are detailed in XXX. 
           
      
 
    
      
 
    
      
 
 
 
     (G1) 
               
      
 
    
      
 
    
      
 
 
 
     (G2) 
 
Table ?G.1: Ethylene Glycol Property Equation Coefficients 
Parameter ? Cp k ? Pr 
Order [kg/m3] [kJ/kg K] [W/m K] [Pa s] [-] 
0     1.0 
1 658.49825 5.36449 0.83818 -4.63024 -0.06982 
2 -54.81501 0.78863 -1.37620 -2.14817 -0.35780 
3 664.71643 -2.59001 -0.07629 -12.70106  
4 232.72605 -2.73187 1.07720 5.40536  
5 -322.61661 1.43759 -0.20174 10.98990  
 
 
275 
Appendix H Modelica Direct Expansion System Equations 
?Appendix H provides information about the general equations used in the 
modeling of the direct expansion system components. The modeling work for the 
direct expansion components was done by Hongtao Qiao. The following 
equations can be found together with a more detailed discussion in Qiao [129].  
The expansion cycle consists of four components, namely the compressor, 
expansion device, and the two heat exchangers.  
The compressor was modeled as a quasi-steady state, efficiency-based 
component, where refrigerant mass flow rate was computed as a function of 
displacement, revolutions per minute, suction density and volumetric efficiency, 
given by Equation (H1). Discharge enthalpy as a function of isentropic efficiency 
was computed according to Equation (H2). The power due to adiabatic 
compression work was determined by Equation (H3). 
                  (H1) 
   
       
   
     (H2) 
                (H3) 
Similar to the compressor model, the expansion valve was modeled as a 
quasi-steady state isenthalpic throttling process. Mass flow rate through the valve 
was determined by flow coefficient, flow area, inlet density and pressure drop 
across the valve, as shown in Equation (H4). 
                   (H4) 
276 
Heat exchanger models are composed of three control volumes, which are 
connected by heat transfer mechanisms. The control volumes are the refrigerant 
control volume, the finned wall, and the air-side control volume. The following 
conservation laws apply to the control volumes: 
Refrigerant-side 
   
   
   
 
 
   
  
  
   
   
 
 
   
  
                  (H5) 
     
   
   
 
 
   
   
  
      
   
   
 
 
    
   
  
                         (H6) 
                    (H7) 
Finned Wall 
     
   
  
             (H8) 
Air-side 
                      
      
      
      (H9) 
Mass and energy balance are calculated in control volumes, whereas the 
momentum balance is calculated between control volumes (through connectors). 
For the microchannel heat exchanger models, air and refrigerant splitter models, 
as well as air and refrigerant mixer models were added before and after the 
respective control volumes.  
277 
Appendix I R134a DX Steady-state Reference Data 
Chapters ?7, ?8, and ?9 show experimental results of the secondary loop 
system using various primary refrigerants relative to the results of the R134a DX 
system. ?Appendix I provides absolute R134a DX data as a reference point for 
these comparisons. 
Figure ?I.1 shows a summary of steady-state performance metrics data for 
a variety of ambient temperatures during idling and highway driving conditions. 
Cooling capacity, shown in Figure ?I.1 a), was lowest, about 2 kW, during idling at 
15?C ambient temperature. The highest cooling capacities, 5.5 kW, were 
measured at highway driving (high compressor rpm) and high ambient 
temperatures of 35?C and 45?C. Figure ?I.1 b) shows compressor power 
consumption  over a range of ambient temperatures for idle, as well as highway 
driving conditions. Power consumption  varied from as low as 0.6 kW during 
idling at 15?C ambient temperature to 4.3 kW during highway driving at 45?C 
ambient temperature. R134a DX coefficient of performance, shown in Figure ?I.1 c) 
decreases with increasing compressor rpm, as well as with increasing ambient 
temperature.  
A summary of sensible and latent performance data is shown in Figure ?I.2. 
Sensible capacity, shown in Figure ?I.2 a), increases with ambient temperature. 
The lowest sensible capacity, 1.5 kW, was measured at 15?C ambient 
temperature during idling. The highest sensible capacity was measured to be 4.6 
kW at 45?C ambient temperature during highway driving.   
278 
 
Figure ?I.1: Steady-state Performance Metrics Summary (R134a DX) 
0.0
 0.5
 1.0
 1.5
 2.0
 2.5
 3.0
 3.5
 4.0
 4.5
 5.0
 T15       
idle
 T25       
idle
 T35       
idle
 T45       
idle
 T15 
hwy
 T25 
hwy
 T35 
hwy
 T45 
hwy
 P
 o
 w
 e
 r
  [
 k
 W
 ]
  
Ambient Temperature [?C] ; Compressor Speed [rpm]
 Power  (%R134a DX)
 0
 1
 2
 3
 4
 5
 6
 7
 8
 T15       
idle
 T25       
idle
 T35       
idle
 T45       
idle
 T15 
hwy
 T25 
hwy
 T35 
hwy
 T45 
hwy
 C
 a
 p
 a
 c
 it
 y
  [
 k
 W
 ]
 Ambient Temperature [?C] ; Compressor Speed [rpm]
 Cooling Capacity (R134a DX)
 0
 1
 2
 3
 4
 5
 T15       
idle
 T25       
idle
 T35       
idle
 T45       
idle
 T15 
hwy
 T25 
hwy
 T35 
hwy
 T45 
hwy
 C
 O
 P
  [
 -]
 Ambient Temperature [?C] ; Compressor Speed [rpm]
 Coefficient of Performance (R134a DX)a)
 b)
 c)
279 
Figure ?I.2 b) presents latent capacity results. In general, latent capacity 
was decreased compared to sensible capacity. During idling, latent capacity 
decreased with increasing ambient temperature from 0.5 kW (15?C) to 0 kW 
(45?C). Latent capacity increased significantly during highway driving with the 
highest latent capacity measured to be 2.0 kW at 35?C. Figure ?I.2 c) presents the 
variation of sensible heat factor (SHF) with ambient temperature under idling and 
highway driving condition. During idling the sensible heat factor is continuously 
increasing from 0.73 at 15?C ambient temperature to 1.0 at 45?C. During 
highway driving, the lowest SHF was measured at 35?C with 0.64, while the 
highest SHF was measured at 45?C with 0.88. 
Figure ?I.3 shows a summary of steady-state compressor efficiencies for 
the R134a DX system. Volumetric efficiency, shown in Figure ?I.3 a), decreases 
with increasing ambient temperature. On average, idling volumetric efficiency 
was calculated to be 0.91, while volumetric efficiency during highway driving 
conditions was calculated to be 0.84. Figure ?I.3 b) presents isentropic efficiency 
results. Isentropic efficiency decreased with increasing ambient temperature. 
During idling conditions, isentropic efficiency decreased from 0.85 (15?C) to 0.77 
(45?C). During highway driving conditions, isentropic efficiency decreases from 
0.69 (25?C) to 0.66 (45?C).  
280 
 
Figure ?I.2: Steady-state Sensible/Latent Performance Summary (R134a DX) 
0.50
 0.55
 0.60
 0.65
 0.70
 0.75
 0.80
 0.85
 0.90
 0.95
 1.00
 T15       
idle
 T25       
idle
 T35       
idle
 T45       
idle
 T15 
hwy
 T25 
hwy
 T35 
hwy
 T45 
hwy
 S
 H
 F
  [
 -]
 Ambient Temperature [?C] ; Compressor Speed [rpm]
 Sensible Heat Factor (R134a DX)
 .0
 0.5
 1.0
 1.5
 2.0
 2.5
 T15       
idle
 T25       
idle
 T35       
idle
 T45       
idle
 T15 
hwy
 T25 
hwy
 T35 
hwy
 T45 
hwy
 C
 a
 p
 a
 c
 it
 y
  [
 k
 W
 ]
 Ambient Temperature [?C] ; Compressor Speed [rpm]
 Latent Capacity (R134a DX)
 0.0
 0.5
 1.0
 1.5
 2.0
 2.5
 3.0
 3.5
 4.0
 4.5
 5.0
 T15       
idle
 T25       
idle
 T35       
idle
 T45       
idle
 T15 
hwy
 T25 
hwy
 T35 
hwy
 T45 
hwy
 C
 a
 p
 a
 c
 it
 y
  [
 k
 W
 ]
 Ambient Temperature [?C] ; Compressor Speed [rpm]
 Sensible Capacity (R134a DX)
 a)
 b)
 c)
281 
 
Figure ?I.3: Steady-state Compressor Efficiency Summary (R134a DX) 
  
0.60
 0.65
 0.70
 0.75
 0.80
 0.85
 0.90
 0.95
 1.00
 T15       
idle
 T25       
idle
 T35       
idle
 T45       
idle
 T15 
hwy
 T25 
hwy
 T35 
hwy
 T45 
hwy
 Ef
 fi
 ci
 en
 cy
  [
 -]
 Ambient Temperature [?C] ; Compressor Speed [rpm]
 Isentropic Efficiency (R134a DX) 
.6
 .6
 .7
 0.75
 0.80
 0.85
 0.90
 0.95
 1.00
 T15       
idle
 T25       
idle
 T35       
idle
 T45       
idle
 T15 
hwy
 T25 
hwy
 T35 
hwy
 T45 
hwy
 Ef
 fi
 ci
 en
 cy
  [
 -]
 Ambient Temperature [?C] ; Compressor Speed [rpm]
 Volumetric Efficiency (R134a DX)
 a)
 b)
282 
References 
[1] J. E. Rubio-Quero,?W.?E.?Dunn,?and?N.?R.?Miller,??A?Facility?for?Transient 
Testing?of?Mobile?Air?Conditioning?Systems,??1995. 
[2] C.?D.?Collins?and?N.?R.?Miller,??Experimental?Study?of?Mobile?Air?
 Conditioning?System?Transient?Behavior,??1996. 
[3] P.?G.?Weston,?W.?E.?Dunn,?and?N.?R.?Miller,??Design?and?Construction?of?a?
 Mobile?Air?Conditioning?Test?Facility?for?Transient?Studies,??1996. 
[4] P.?Hrnjak,??Some?Lessons?Learned?from?SAE?AR?CRP,??in?SAE Phoenix, 
2002. 
[5] P.?Hrnjak?and?B.?Hill,??OVERVIEW?OF?THE?SAE?ALTERNATE?
 REFRIGERANT?COOPERATIVE?RESEARCH?PROGRAM.??2003. 
[6] P.?Hrnjak,??Challenges?and?Results?of?MAC?Cooling?and?Energy?Efficiency?
 Performance?Exploration?Project?at?the?University?of?Illinois,??in?MAC 
Summit, 2004. 
[7] B.?Hill,??Overview?of?Systems?under?Evaluation?in?SAE?Alternate?
 Refrigerant Cooperative Research?Program,??in?SAE Phoenix, 2002. 
[8] J.?Wertenbach,??SAE?Alternate?Refrigerant?Cooperative?Research?Project?
 Presentation - Energy?Analysis?of?Refrigerant?Cycles,??SAE Phoenix, p. 37, 
2003. 
[9] W.?Atkinson,??SAE?Alternate?Refrigerant?Cooperative?Research?Project - 
Project?Overview,??in?SAE Phoenix, 2003. 
[10] W.?Atkinson,??SAE?Alternate?Refrigerant?Cooperative?Research?Program?
 And?Industry?Impact,??MAC Summit, pp. 1?19, 2006. 
[11] J.?Baker,?M.?Ghodbane,?J.?Rush,?and?W.?Hill,??Alternative?Refrigerant?
 Demonstration?Vehicles,??2007. 
[12] A. E.-S.?Gado,??Development?of?a?Dynamic?Test?Facility?For?Environmental?
 Control?Systems,??Thesis, University of Maryland, 2006. 
[13] A.?Gado,?Y.?Hwang,?and?Reinhard?R,??Dynamic?Behavior?of?Mobile?Air-
 Conditioning?Systems,??HVAC&R Research, vol. 14, no. 2, p. 16, 2008. 
[14] M.?B.?Yahia?and?C.?Petitjean,??Transient?phenomena?effect?on?energy?
 consumption?Environmental?background,??in?SAE Phoenix, 2002. 
283 
[15] C.-C.?D.?Huang,??A?Dynamic?Simulation?Model?For?Automobile?Passenger?
 Compartment?Climate?Control?And?Evaluation,??1998. 
[16] M.?Wang,?D.?L.?Farley,?and?L.?L.?Leitzel,??Air?Conditioning?System?Head?
 Pressure?Spike?During?Vehicle?Acceleration,??in?SAE, 2000, no. 724. 
[17] T.?J.?Hendricks,??Optimization?of?Vehicle?Air?Conditioning?Systems Using 
Transient?Air?Conditioning?Performance?Analysis,??in?SAE VTMS, 2001. 
[18] R.?Marzy,?J.?Hager,?and?C.?Doppelbauer,??Optimization?of?Warm?Up?Using?
 Simulation?Tools,??SAE VTMS, 2001. 
[19] C.?Tian?and?X.?Li,??Transient?behavior?evaluation?of?an?automotive air 
conditioning?system?with?a?variable?displacement?compressor,??Applied 
Thermal Engineering, vol. 25, no. 13, pp. 1922?1948, Sep. 2005. 
[20] H.?Tummelscheit?and?D.?Limperich,??The?AirConditioning?library?for?
 simulation of advanced vehicle A/C systems,??in?SAE VTMS, 2007, pp. 23?
 32. 
[21] I.?Bayraktar,??TIME?DEPENDENT?SIMULATION?METHODS?FOR?
 VEHICLE?THERMAL?MANAGEMENT,??GVSETS Proceedings, pp. 1?4, 
2009. 
[22] B.?Li?and?A.?G.?Alleyne,??A?Full?Model?of?a?HVAC?Vapor?Compression?
 Cycle Interacting with a Dynamic?Environment,??in?American Control 
Conference, 2009, pp. 3662?3668. 
[23] K. Sandhu, C. Chatham, A. Milosevic, D. Adekeye, and M. Warner, 
?Method?of?Evaluating?the?Effect?of?Air?Conditioning?on?Vehicle?Energy?
 Management,??p.?20. 
[24] ASHRAE,??ANSI/?ASHRAE Standard 34, Designation and Safety 
Classifications of Refrigerants, American Society of Heating, Refrigerating, 
and Air-Conditioning?Engineers.??Atlanta,?GA,?2004. 
[25] ISO,??ISO?Standard?817:?Refrigerants?Designation?System,?International?
 Organization?for?Standardization.??Geneva,?Switzerland,?2005. 
[26] K.?Wang,?M.?Eisele,?Y.?Hwang,?and?R.?Radermacher,??Review?of?
 secondary?loop?refrigeration?systems,??International Journal of 
Refrigeration, vol. 33, no. 2, pp. 212?234, Mar. 2010. 
[27] E.?Union,??Directive 2006/40/EC of the European parliament and of the 
Council?of?17?May?2006,?Official?J.?Eur.?Union?L?161?(12).??2006. 
284 
[28] M.?Spatz?and?B.?Minor,??HFO-1234yf ? a?low?GWP?refrigerant?for?MAC,??in?
 VDA Alternative Refrigerant Winter Meeting, 2008. 
[29] E.?Granryd,??Hydrocarbons?as?refrigerants,??International Journal of 
Refrigeration, no. 24, pp. 15?24, 2001. 
[30] J.?M.?Corberan,?J.?Segurado,?D.?Colabourne,?and?J.?Gonzalvez,??Review?of?
 standards for the use of hydrocarbon refrigerants in a/c, heatpump and 
refrigeration?equipment,??International Journal of Refrigeration, no. 31, pp. 
748?756, 2008. 
[31] CEN,??prEN?378-2, Refrigerating Systems and Heat Pumps ? Safety and 
Environmental?Requirements,?European?Committee?for?Standardization.??
 Brussels, Belgium, 2006. 
[32] BSI,??BS?EN?378-1: Refrigerating Systems and Heat Pumps, Safety and 
Environmental?Requirements,?Basic?Requirements,?Definitions,?
 Classification?and?Selection?Criteria,?British?Standard?Institute,?Bristol,?U.K.??
 Bridtol, U.K., 2008. 
[33] DIN,??DIN?8975-12: Refrigerating Plants ?Safety Principles for Design, 
Equipment?and?Installation;?Test?Certificate?and?Identification?Plate?- Part 
12. Recovery Systems, Deutsches Institut fuer Normung e.V., Berlin, 
Germany.??Berlin, Germany, 2004. 
[34] D. Colbourne, ?An?overview?of?hydrocarbons?as?replacement?refrigerants?in?
 commercial refrigeration and air-conditioning, Refrigeration Northern 
Ireland?Centre?for?Energy?Research?and?Technology.??2000. 
[35] D.?Colbourne?and?K.?O.?Suen,??Assessment?of?performance?of?hydrocarbon 
refrigerants,??Science et technique du froid, pp. 149?156, 2000. 
[36] D.?Colbourne?and?T.?J.?Ritter,??Compatibility?of?non-metallic materials with 
hydrocarbon?refrigerant?and?lubricant?mixtures,??Science et technique du 
froid, pp. 413?418, 2000. 
[37] I. L. Maclaine-Cross?and?E.?Leonardi,??Comparative?performance?of?
 hydrocarbon?refrigerants,??I.I.F. ? I.I.R. ?Commissions E2, E1, B1, B2, 
Melbourne, Australia, 1996. 
[38] K.?A.?Joudi,?A.?S.?K.?Mohammed,?and?M.?K.?Aljanabi,??Experimental?and?
 computer performance study of an automotive air-conditioning system with 
alternative?refrigerants,??Energy Conversion and Management, no. 44, pp. 
2959?2976, 2003. 
285 
[39] M.?Ghodbane,??An?Investigation?of?R152a?and?Hydrocarbon?Refrigerants?in?
 Mobile?Air?Conditioning,??in SAE, 1999, no. 724. 
[40] I. L. Maclaine-cross?and?E.?Leonardi,??Hydrocarbon?refrigerant?risk?in?car?
 air-conditioners,??in?International CFC and Halon Alternatives Conference, 
1995. 
[41] V.?Razmovski,??Safety?of?hydrocarbon?refrigerants?for?car?air- conditioning 
systems,?School?of?Mechanical?Engineergin,?UNSW,?Sydney,??UNSW,?
 Sydney, 1994. 
[42] ?Arthur?D.?Little?Ltd.,?Risk?Assessment?of?Flammable?Refrigerants,?Part?3:?
 Car Air-Conditioning.?Final?Report?to?Calor?Gas?Ltd.,?Cambridge?UK,??1995. 
[43] D. J.T., J. Bentley,?and?A.?Varone,??Non-Inert Refrigerant Study for 
Automotive?Applications.??1991. 
[44] S.?J.?Elbers?and?M.?Verwoerd,??Quantitative?risk?assessment?of?a?heat?
 pump system with propane refrigerant, Final Report of TNO (for Lodam 
Energi A/S), Apeldoorn,?Netherland,??Apeldoorn,?Netherland,?1997. 
[45] T.?J.?Ritter?and?D.?Colbourne,??Quantitative?risk?assessment:?hydrocarbon?
 refrigerants, IIF - IIR ?Sections?B?and?E,?Oslo,?Norway,??1998. 
[46] J.?Jetter,?F.?Reynaldo,?and?R.?Rubenstein,??Fault?tree?analysis?for exposure 
to refrigerants used for automotive air-conditioning?in?the?United?States,??
 Risk Analysis, no. 21, pp. 157?171, 2001. 
[47] D.?Colbourne?and?K.?O.?Suen,??Appraising?the?flammability?hazards?of?
 hydrocarbon refrigerants using quantitative risk assessment model, partII, 
model?evaluation?and?analysis,??International Journal of Refrigeration, 2004. 
[48] B.?Palm,??Hydrocarbons?as?refrigerants?in?small?heat?pump?and?
 refrigeration systems ? a?review,??Internal Journal of Refrigeration, no. 31, 
pp. 552?563, 2008. 
[49] Danfoss,??Practical?Application?of?Refrigerant?R290?Propane?in?Small?
 Hermetic?Systems.??2000. 
[50] J. M. Corberan, J. F. Urchueguia, I. Navarro, J. Gronzalvez, and A. Gala, 
?Performance?of?a?reciprocating?hermetic?refrigerant?compressor?using 
propane?as?working?fluid,??in?IIR - Gustav Lorentzen Conference of Natural 
Working Fluids, 2000, pp. 225?232. 
286 
[51] C.?L.?Pellec,?C.?Marvillet,?and?D.?Clodic,??Experimental?study?of?flat?heat?
 exchangers?in?ammonia?refrigeration?unit,??IIR - Gustav Lorentzen 
Conference of Natural Working Fluids, pp. 785?794, 1996. 
[52] T. Setaro, G. Boccardi, J. M. Corberan, J. Urchueguia, and J. Gonzalvez, 
?Comparative?study?of?evaporation?and?condensation?of?propane?and?R22?
 in?a?brazed?plate?heat?exchanger?and?a?tube?and?fins coil,??in?IIR - Gustav 
Lorentzen Conference of Natural Working FluidsGustav Lorentzen, 2000. 
[53] P.?S.?Hrnjak?and?M.?R.?Hoehne,??Charge?minimization?in?systems?and?
 components using hydrocarbons as a refrigerant, ACRC TR-224.??2004. 
[54] P.?Fernando,?B.?Palm,?T.?Ameel,?P.?Lundqvist,?and?E.?Granryd,??Propane?
 heat?pump?with?low?refrigerant?charge:?design?and?laboratory?tests,??
 International Journal of Refrigeration, no. 27, pp. 761?773, 2004. 
[55] P. Fernando, B. Palm, T. Ameel, P. Lundqvist,?and?E.?Granryd,??A?
 minichannel aluminum tube heat exchanger ? part I: evaluation of single- 
phase?heat?transfer?coefficients?by?the?Wilson?plot?method,??International 
Journal of Refrigeration, no. 31, pp. 669?680, 2008. 
[56] P. Fernando, B. Palm, T. Ameel,?P.?Lundqvist,?and?E.?Granryd,??A?
 minichannel aluminum tube heat exchanger ? part II: evaporator 
performance?with?propane,??International Journal of Refrigeration, no. 31, 
pp. 681?695, 2008. 
[57] P.?Fernando,?B.?Palm,?T.?Ameel,?P.?Lundqvist,?and?E.?Granryd,??A?
 minichannel aluminum tube heat exchanger ? part III: condenser 
performance?with?propane,??International Journal of Refrigeration, no. 31, 
pp. 696?708, 2008. 
[58] P.?Likes,??Secondary?refrigerant?systems?for?supermarket?equipment,??in?
 Conference on Ozone Protection Technologies, 1996, pp. 158?164. 
[59] L.?Rolfsman,??CO2?and?NH3?in?the?supermarket?ICA-focus,??in?IIR - 
Conference of on Applications for Natural Refrigerants, 1996, pp. 219?225. 
[60] K.?Evenmo,??A?new?secondary?fluid,??in?IIR - Gustav Lorentzen Conference 
on Natural Working Fluids, 1998, pp. 679?688. 
[61] W.?T.?Horton,?E.?A.?Groll,?and?E.?Sharp,??Testing?of?a?drop-in secondary 
loop refrigeration system for medium temperature supermarket?application,??
 in IIR - Gustav Lorentzen Conference on Natural Working Fluids, 1998, pp. 
690?700. 
287 
[62] J.?Nyvad?and?S.?Lund,??Indirect?cooling?with?ammonia?in?supermarkets,??in?
 IIR - Gustav Lorentzen Conference on Natural Working Fluids, 1998, pp. 
725?734. 
[63] ?Tyforop?Chemie?Gmbh,?SecondaryRefrigerants(OrganicSalts),?
 http://www.tyfo.de/index_english.html.??2009. 
[64] J.?Arias?and?P.?Lundqvist,??Field?experiences?in?three?supermarkets?in?
 Sweden,??in?IEA Annex 26 Workshop, IEA Heat Pump Centre, 2000. 
[65] R.?T.?Faramarzi?and?D.?H.?Walker,??Investigation?of?Secondary?Loop?
 Supermarket Refrigeration Systems, Consultant Report, Foster-Miller, Inc. 
CaliforniaEnergy Commission 500-04-0132004,??2004. 
[66] V.?Minea,??Supermarket?refrigeration?system?with?completely secondary 
loops,??ASHRAE Transactions, no. 49, p. 40, 2007. 
[67] V.?Minea,??Low?charge?and?low?emission?supermarket?refrigeration?
 systems,??in?IIR - Gustav Lorentzen Conference on Natural Working Fluids, 
2008, pp. 901?908. 
[68] N.?Rivers,??Unconventional secondary?refrigeration?in?a?UK?supermarket,??
 in IEA Heat Pump Centre Newsletter 18, 2000, pp. 18?19. 
[69] V.?D.?Baxter,??Advances?in?Supermarket?Refrigeration?Systems,??in?IEA 
Annex 26 Summary, 2006. 
[70] D.?K.?Choi,?P.?A.?Domanski,?and?D.?A.?Didion,??Evaluation?of?flammable?
 refrigerants for use in a water-to-water?residential?heat?pump,??in?IIR - 
Conference on Applications for Natural Refrigerants, 1996, pp. 467?476. 
[71] Y.?S.?Chang,?M.?S.?Kim,?and?S.?T.?Ro,??Performance?and?heat?transfer?of?
 hydrocarbon refrigerants?and?their?mixtures?in?a?heat?pump?system,??in?IIR - 
Conference on Applications for Natural Refrigerants, 1996, pp. 477?486. 
[72] O.?Pelletier?and?B.?Palm,??Performance?of?plate?heat?exchangers?and?
 compressor in a domestic heat pump using propane,??IIR - Conference on 
Applications for Natural Refrigerants, pp. 497?506, 1996. 
[73] W.?V.?Payne,?P.?A.?Domanski,?and?J.?Muller,??A?study?of?a?water-to-water 
heat?pump?using?flammable?refrigerants,??in?IIR - Gustav Lorentzen 
Conference on Natural Working Fluids, 1998, pp. 658?667. 
[74] J.?Stene,??Investigation?of?a?residential?brine-water CO2 heat pump for 
combined low-temperature?space?heating?and?hot?water?preparation,??in?IIR 
288 
- Conference on Applications on Natural Working Fluids, 2002, pp. 268?
 275. 
[75] T.?Yanagisawa,?M.?Fukuta,?N.?Ogura,?and?H.?Kaneo,??Operating?
 characteristics of natural circulating CO2 secondary loop refrigeration 
system?working?with?NH3?primary?loop,??in?IIR - Gustav Lorentzen 
Conference on Natural Working Fluids, 2004. 
[76] P. A. Domanski?and?D.?Yashar,??Comparable?performance?evaluation?of?
 HC?and?HFC?refrigerants?in?an?optimized?system,??in?IIR - Gustav 
Lorentzen Conference on Natural Working Fluids, 2006. 
[77] K.?Mani?and?V.?Selladurai,??Experimental?analysis?of?a?new?refrigerant 
mixture as drop-in?replacement?for?CFC12?and?HFC134a,??International 
Journal of Thermal Sciences, no. 47, pp. 1490?1495, 2008. 
[78] W.?Hill,??Risk?assessment?and?performance?evaluation?of?HFO-1234yf,??in?
 VDA Alternative Refrigerant Winter Meeting, 2008. 
[79] EPA,??USA?Environmental?Protection?Agency,?Legal?Status?of?HC-12a ?, 
DURACOOL 12a ?, and OZ-12??,??2010.?[Online]. Available: 
http://www.epa.gov/ozone/snap/refrigerants/hc-12a.html. 
[80] ?USA?Environmental?Protection?Agency,?Unacceptable?Substitute?
 Refrigerants,??2010.?[Online].?Available:?
 http://www.epa.gov/ozone/snap/refrigerants/lists/unaccept.html. 
[81] Z.?Ure,??Secondary?Refrigeration?- European?Experiences,??ASHRAE 
Transactions, vol. 109, pp. 1?6, 2003. 
[82] R. Delventura, C. L. Evans, and I. Richter,??Magazine:?Secondary?loop?
 SyStemS?for?the?Supermarket?induStry?White?Paper,??White Paper, no. 
September, 2007. 
[83] A.?Melinder,??Thermophysical?properties?of?aqueous?solutions?used?as?
 secondary?working?fluids,??Royal?Institute?of?Technology,?2007. 
[84] P.?W.?Egolf?and?M.?Kauffeld,??From?physical?properties?of?ice?slurries?to?
 industrial?ice?slurry?applications,??International Journal of Refrigeration, no. 
28, pp. 4?12, 2005. 
[85] Z.?Ure?and?M.?Mashrae,??Slurry?ice?based?cooling?systems,??in?20th 
International Congress of Refrigeration, 1999. 
289 
[86] J.?W.?Meewise?and?C.?A.?I.?Ferreira,??Optimal?properties?of?ice?slurries?in?
 secondary?cooling?systems,??in?IIR - Gustav Lorentzen Conference of 
Natural Working Fluids, 2000, pp. 513?520. 
[87] M. J. Wang and?V.?Goldstein,??A?novel?ice?slurry?generation?system?and?its?
 application,??in?IIR - Conference on Applications for Natural Refrigerants, 
1996, pp. 543?551. 
[88] K.?G.?Christensen?and?M.?Kauffeld,??Ice?slurry?accumulation,??in?IIR - 
Gustav Lorentzen Conference on Natural Working Fluids, 1998, pp. 701?
 711. 
[89] M.?Wang,?J.?Y.?Inoue,?and?V.?Goldstein,??Ice?thermal?storage?in?modern?
 building,??in?20th International Congress of Refrigeration, 1999, pp. 19?24. 
[90] S. Fukusako, Y. Kozawa, M. Yamada, and M. Tanino,??Research?and?
 development?activities?on?ice?slurries?in?Japan,??in?First Workshop on Ice 
Slurries of the International Institute of Refrigeration: Proceedings of the 
International Meeting in Yverdon-les-Bains, Switzerland, 1999, pp. 83?105. 
[91] A. Saito,??Recent?advances?in?research?on?cold?thermal?energy?storage,??
 International Journal of Refrigeration, no. 25, pp. 177?189, 2002. 
[92] D.?Hinde,?S.?Zha,?and?L.?Lan,??CO2?experiences?in?North?American?
 supermarkets,??in?IIR - Gustav Lorentzen Conference on Natural Working 
Fluids, 2008, pp. 1098?1104. 
[93] A.?Melinder,??Secondary?Fluids?for?Low?Operating?Temperatures,??in?IIR - 
Gustav Lorentzen Conference on Natural Working Fluids, 2004. 
[94] A.?C.?Pachai,??Experience?with?CO2?as?refrigerant?in?supermarkets,??in?IIR - 
Gustav Lorentzen Conference on Natural Working Fluids, 2004. 
[95] A.?Pearson,??Carbon?dioxide?- new?uses?for?an?old?refrigerant,??Internal 
Journal of Refrigeration, no. 28, pp. 1140?1148, 2005. 
[96] S.?F.?Pearson,??Cooling?Method?and?Apparatus, British Patent Number 
2258298,??1995. 
[97] K.?G.?Christensen,??Use?of?CO2?as?primary?and?secondary?refrigerant?in?
 supermarket?application,??in?20th International Congress of Refrigeration, 
1999, pp. 1936?1942. 
[98] P. O. Nilsson, J. Rogstam, S. Sawalha, and?K.?Shahzad,??Ice?rink?
 refrigeration?system?with?carbon?dioxide?as?secondary?fluid?in?copper?
290 
tubes,??in?IIR - Gustav Lorentzen Conference on Natural Working Fluids, 
2006. 
[99] K.?Jahn,??Natural?refrigerants?in?dairy?processing,?supermarket?refrigeration?
 and?airconditioning,??GTZ Proklima: Natural Refrigerants - Sustainable 
Ozone - and Climate - Friendly Alternatives to HCFCs Eschborn Germany, 
Deutsche Gesellschaft fuer Technische Zusammenarbeit (GTZ) GmbH, 
German Technical Cooperation ? Programme Proklima, 2008. 
[100] L.?Dentis,?A.?Mannoni,?and?M.?Parrino,??HC?refrigerants:?an?ecological?
 solution?for?automotive?a/c?systems,??in?Vehicle Thermal Management of 
Systems Conference, 1999, pp. 133?147. 
[101] M.?Ghodbane,??On?Vehicle?Performance?of?a?Secondary?Loop A/C?System,??
 SAE Technical Paper, no. 724, p. 7, 2000. 
[102] M.?Ghodbane?and?J.?A.?Baker,??Secondary?loop?system?for?passenger?
 compartment?heating?and?cooling,?Unites?States?Patent?Number?6405793,??
 2002. 
[103] P.?S.?Kadle?and?M.?Ghodbane,??Heat?pump?with?secondary loop air-
 conditioning?system,?United?States?Patent?Number?7063137,??2006. 
[104] M.?Ghodbane,?T.?D.?Craig,?and?J.?A.?Baker,??Demonstration?of?an?Energy-
 Efficient Secondary Loop HFC-152a?Mobile?Air?Conditioning?System,??2007. 
[105] ?ANSI/ASHRAE?41.2-1987 (RA92) - Standard Methods for Laboratory 
Airflow?Measurement,??Order A Journal On The Theory Of Ordered Sets 
And Its Applications. American Society Of Heating Refrigerating And Air-
 Conditioning Engineers, Atlanta, GA, 1992. 
[106] ?http://www.ni.com/dataacquisition/.??. 
[107] ?http://www.optimizedthermalsystems.com/.??. 
[108] E.?W.?Lemmon,?M.?L.?Huber,?and?M.?O.?McLinden,??NIST?Standard?
 Reference Database 23: Reference Fluid Thermodynamic and Transport 
Properties-REFPROP, Version 9.0, National Institute of Standards and 
Technology,?Standard?Reference?Data?Program.??Gaithersburg,?2010. 
[109] M.?Conde?Engineering,??Properties?of?Working?Fluids,?THERMOPHYSICAL?
 PROPERTIES?OF?BRINES.??Zurich,?2011. 
[110] ?USA?Environmental?Protection?Agency,?Dynamometer?Drive?Schedules,??
 2012. [Online]. Available: 
http://www.epa.gov/nvfel/testing/dynamometer.htm. 
291 
[111] ?National?Renewable?Energy?Laboratory,?U.S.?Department?Of?Energy,?
 Typical?Meteorological?Year,?Version?2,??2005.?[Online]. Available: 
http://rredc.nrel.gov/solar/old_data/nsrdb/1961-1990/tmy2/State.html. 
[112] M.?Ghodbane,??On?vehicle?performance?of?a?secondary?loop?system,??in?
 SAE Phoenix, 1999. 
[113] D.?A.?Didion,??The?Influence?of?the?Thermophysical?Fluid?Properties?of?the?
 New Ozone-Safe?Refrigerants?on?Performance,??International Journal of 
Applied Thermodynamics, vol. 2, no. 1, pp. 19?35, 1999. 
[114] M.?O.?McLinden?and?D.?A.?Didion,??Quest?for?Alternatives,??ASHRAE 
Journal, vol. 29, pp. 32?42, 1987. 
[115] G.?Alefeld,??Efficiency?of?compressor?heat?pumps?and?refrigerators?derived?
 from?the?2nd?Law,??International Journal of Refrigeration, vol. 10, no. 6, pp. 
331?341, 1987. 
[116] F. Ayad, L. Leitzel, and?F.?D.?Sciullo,??Numerical?and?experimental?study?of?
 a secondary cooling loop including a liquid cooled charge air cooler and a 
liquid?cooled?condenser,??in?4th European Workshop: Mobile Air 
Conditioning and Vehicle Thermal Systems, 2011. 
[117] SAE,??Surface Vehicle Standard, SAE J2765 OCT2008, Procedure for 
Measuring System COP (Coefficient of Performance) of a Mobile Air 
Conditioning?System?on?a?Test?Bench.??2008. 
[118] S.?K.?Fischer,??Comparison?of?Global?Warming?Impacts?of?Automobile?Air-
 Conditioning Concepts,??in?International CFC and Halons Alternative 
Conference, 1995. 
[119] S.?Papasavva,?W.?R.?Hill,?and?R.?O.?Brown,??Green-MAC-LCCP: A Tool for 
Assessing?Life?Cycle?Greenhouse?Emissions?of?Alternative?Refrigerants,??
 SAE International Journal of Passenger Cars - Mechanical Systems, vol. 1, 
no. 1, pp. 746?756, 2008. 
[120] U.S.?Department?of?Commerce?and?U.S.?Census?Bureau,??Commuting?in?
 the?United?States?:?2009,??Washington,?DC,?2011. 
[121] D.?Clodic,?E.?Zgheib,?and?S.?Mortada,??Impacts?of?Heating?and?Cooling on 
Electrified?vehicles,??in?4th European Workshop: Mobile Air Conditioning 
and Vehicle Thermal Systems, 2011. 
[122] C.?Reynolds?and?M.?Kandlikar,??How?hybrid-electric vehicles are different 
from conventional vehicles: the effect of weight and power on fuel 
292 
consumption,??Environmental Research Letters, vol. 2, no. 1, p. 014003, 
Jan. 2007. 
[123] V.?H.?Johnson,??Fuel?Used?for?Vehicle?Air?Conditioning?:?A?State-by-State 
Thermal Comfort-Based?Approach,??Society of Automotive Engineers, 
2002. 
[124] R. Farrington?and?J.?Rugh,??Impact?of?Vehicle?Air- Conditioning on Fuel 
Economy , Tailpipe Emissions , and Electric Vehicle Range - Preprint,??in?
 Earth Technologies Forum, 2000, no. September. 
[125] K.?Umezu?and?H.?Noyama,??Air-Conditioning system For Electric Vehicles 
(i-MiEV),??in?SAE Automotive Refrigerant & System Efficiency Symposium, 
2010. 
[126] W.?O.?Forrest?and?M.?S.?Bhatti,??Energy?Efficient?Automotive?Air?
 Conditioning?System,??SAE 2002 World Congress, no. 724, 2002. 
[127] D. Roy, P. P. Jean, D. Clodic, and?K.?E.?Khoury,??Influence?of?thermal?
 preconditioning?technologies?on?A/C?system?power?and?passengers??
 thermal?comfort,??IMechE, pp. 205?214, 2003. 
[128] D.?Roy,?P.?Petitjean,?D.?Clodic,?and?K.?E.?Khoury,??Experimental?
 Investigation of a Thermal Preconditioning?of?a?Car?Cabin,??SAE Technical 
Paper Series, vol. 01, no. 2057, 2005. 
[129] H.?Qiao,?V.?Aute,?and?R.?Radermacher,??Comparison?of?Equation-Based 
and Non-Equations-Based Approaches for Transient Modeling of a Vapor 
Compression?Cycle,??in?International Refrigeration and Air Conditioning 
Conference at Purdue, 2012, p. 2190. 
[130] F.?W.?Dittus?and?L.?M.?K.?Boelter,??Heat?Transfer?in?Automobile?Radiators?
 of?the?Tubular?Type,??International Communications in Heat and Mass 
Transfer (originally published in University of California Publications in 
Engineering), vol. 12, no. 1, pp. 3?22. 
[131] F. P. Incropera and D. P. DeWitt, Introduction to heat transfer (3rd ed.). 
New York: John Wiley & Sons, 1996. 
[132] A. A. Eilemann, J. Diem, H. Hofmann, B. Gmbh, and C. Kg,??From?indirect?
 Charge Air Cooling to a 2-layer?Cooling?Module?:?Assumptions?,?
 Preconditions?and?Perspectives,??in?4th European Workshop: Mobile Air 
Conditioning and Vehicle Thermal Systems, 2011. 
[133] R.?Cipollone,?D.?D.?Battista,?and?A.?Gualtieri,??INTEGRATION OF 
COOLING VEHICLE THERMAL NEEDS ON ENGINE COOLING 
293 
SYSTEM,??in?4th European Workshop: Mobile Air Conditioning and Vehicle 
Thermal Systems, 2011. 
[134] F. D. Sciullo, W. Ferraris, C. Malvicino, C. Ricerche, S. C. Fiat, F. Vestrelli, 
and F. Beltramelli,??Two?Levels?Cooling?(?Smart?Cooling?)?Application?to?B-
 segment?vehicle?Sc,??in?4th European Workshop: Mobile Air Conditioning 
and Vehicle Thermal Systems, 2011. 
[135] J.?Benouali,?R.?Beauvis,?and?C.?Petitjean,??R-1234yf A/C system 
improvements & further?link?to?engine?cooling?loop,??in?4th European 
Workshop: Mobile Air Conditioning and Vehicle Thermal Systems, 2011. 
[136] R.?Seccardini,?F.?Mattiello,?and?C.?Malvicino,??TIFFE?thermal?system?Light?
 commercial?vehicle?application,??in?4th European Workshop: Mobile Air 
Conditioning and Vehicle Thermal Systems, 2011.