experimental study of zeotropic refrigerant mixture hfc-407c as … · 2016-12-02 · 1...
TRANSCRIPT
1
Experimental study of Zeotropic refrigerant mixture
HFC-407C as a replacement for HCFC-22 in
Refrigeration and air conditioning systems
By
Changiz M. Tolouee
B. Sc. (Mech. Eng.) and M. Sc. (Mech. Eng.)
A thesis submitted for the degree of
DOCTOR OF PHILOSOPHY
School of Engineering and Science Swinburne University of Technology
2006
2
DECLARATION
This thesis contains no material which has been accepted for the award of any other
degree or diploma, except where due reference is made in the text of the thesis. To the
best of my knowledge, this thesis contains no material previously published or written
by another person except where due reference is made in the text of the thesis.
Signed …………………………….
3
ABSTRACT
HCFC-22 is the world’s most widely used refrigerant. It serves in both residential and
commercial applications, from small window units to large water chillers, and
everything in between. Its particular combination of efficiency, capacity and pressure
has made it a popular choice for equipment designers. Nevertheless, it does have some
ODP, so international law set forth in the Montreal Protocol and its Copenhagen and
Vienna amendments have put HCFC-22 on a phase out schedule. In developed
countries, production of HCFC-22 will end no later than the year 2030.
Zeotropic blend HFC-407C has been established as a drop-in alternative for HCFC-22
in the industry due to their zero Ozone Depletion Potential (ODP) and similarities in
thermodynamic properties and performance. However, when a system is charged with a
zeotropic mixture, it raises concerns about temperature glide at two-phase state,
differential oil solubility and internal composition shift.
Not enough research has been done to cover all aspects of alternative refrigerants
applications in the systems. This research intended to explore behavior of this
alternative refrigerants compare to HCFC-22 and challenges facing the industry in
design, operation service and maintenance of these equipments.
The purpose of this research is to investigate behavior of R407C refrigerant in chiller
systems. This includes performance and efficiency variations when it replaces R22 in an
existing system as well as challenges involved maintaining the system charged with
R407C. It is a common practice in the industry these days to evacuate and completely
recharge when part of the new refrigerant blend was leaked from the system. This has
proved to be extremely costly exercise with grave environmental ramifications.
This research is intended to address challenges faced in the real world and practical
terms.
Theoretical and experimental approaches used as a methodology in this work. The
system mathematically modeled to predict detailed system performance and effect of
the leak at various conditions. To make this feasible and accurate enough, two separate
approaches made, first system performance for pure R22 and R407C, and second
4
system subjected to range of leak fractions. The earlier model was relatively straight
forward when compared to the latter. Modeling a system charged with R407C ternary
mixture and subjected to range of leaks posed enormous challenges.
A sophisticated experimental test apparatus was also designed and built. Comprehensive
and detailed tests at various conditions were conducted with special attention on
instrumental accuracy and correct methodology.
The first part has been successfully modeled and predicted all the factors and
performance with excellent accuracy when compared to the test results. In these
approaches pure refrigerants R22 and R407C were used and simulated the system
behavior at range of conditions.
However, the second part was the most challenging ever. Comprehensive leak process
simulations produced trends of R32/R125/R134a composition change as function of rate
of leak. Starting from this point, equations have been created to represent the
composition change as function of percentage of the leak. The system thermodynamic
cycle was also modeled to calculate capacity, power input and COP at the range of the
conditions. Despite many affecting parameters and complexity of the model, the
mathematical model successfully predicted the test outcome with a very reasonable
accuracy, averaging around 3% with some times reaching to 5 to 6%.
On the experimental stage the system charged with the new HFC-407C was deliberately
subjected to refrigerant leak at various leak stages. The aim was to objectively
determine to what extend the gas leak can be still acceptable without going through the
expensive complete gas charge. The effect of leak was tested and verified at 10% steps,
from 10% up to 50% mass fraction for the total charge.
It has been observed that at the leaks beyond 30%, the adverse effect on the capacity
becomes more significant, from 8 to about 15% decrease. While the power input
decreased at slower pace, from 3% up to about 8% depending on the test conditions.
This translated to COP decrease ranging from 4 to about 7%. This capacity loss and
efficiency decrease are significant figures which suggests that the system, here chiller,
can not be allowed to degrade the performance to that extend and still continue
operating.
5
ACKNOWLEDGMENTS I would like to express my gratitude to my supervisor Prof. Yos Morsi, for his
encouragement, guidance and advice throughout the course of this research project. His
supervision and contribution to this research work has been very valuable. Prof. Yos
Morsi’s initiative in establishing link between the Swinburne University of Technology
and the refrigeration and air-conditioning industry was a key factor in continuation of
this research work.
I would like to thank Dr Wei Yang for his advice, guidance and full support in bringing
together all necessary facilities from literature to building test rig. His continual support
and dedication throughout the course of this work proved invaluable.
The early stage of this work goes back to my research work in the University of New
South Wales where I was guided by Professor Masud Behnia and Professor Eddie
Leonardi to whom I would like to forward my appreciation for their contribution.
Thanks to Mr. Giovanni Giofre for his assistance during construction and
commissioning of the test apparatus.
Thanks to all workshop staff for the construction and maintenance of the experimental
apparatus.
6
TABLE OF CONTENTS
LIST OF FIGURES ..........................................................................................................8
LIST OF TABLES ..........................................................................................................11
NOMENCLATURE........................................................................................................12
1 INTRODUCTION ..................................................................................................14 1.1 PREFACE .......................................................................................................14 1.2 WORLDWIDE OZONE DEPELETION LEGISLATION ............................14
1.2.1 Kyoto Protocol ........................................................................................18 1.2.2 HCFC-22 (R22) Substitutes ....................................................................18 1.2.3 Properties of HCFC-22 (R22) Substitutes ..............................................19 1.2.4 Retrofitting Existing systems ..................................................................20 1.2.5 Refrigerant Solutions for Today’s Environmental Challenges ...............21 1.2.6 HCFC-22 Phase out and Recycling ........................................................23
1.3 PURPOSE OF THIS RESEARCH .................................................................23
2 LITERATURE REVIEW .......................................................................................25
3 THEORY ................................................................................................................36 3.1 INTRODUCTION ..........................................................................................36 3.2 AZEOTROPIC BLEND..................................................................................37 3.3 ZEOTROPIC BLENDS ..................................................................................39 3.4 THERMODYNAMIC PROPERTIES DETERMINATION ..........................42
3.4.1 Redlich-Kwong-Soave (RKS) (1980) Equation of State ........................42 3.4.2 Vapor Density .........................................................................................44 3.4.3 Liquid Density.........................................................................................45 3.4.4 Enthalpy ..................................................................................................45 3.4.5 Entropy....................................................................................................46
3.5 OTHER EQUATIONS OF STATE................................................................47 3.5.1 Pure Refrigerants.....................................................................................47 3.5.2 Mixed Refrigerants .................................................................................48
3.6 CYCLE COMPONENTS ...............................................................................50 3.6.1 Compressor .............................................................................................50 3.6.2 Evaporator ...............................................................................................52 3.6.3 Condenser................................................................................................52 3.6.4 Expansion Valve .....................................................................................53 3.6.5 Suction and discharge Lines ...................................................................53
3.7 Determination of the Polytropic Exponent .....................................................53 3.8 CONSTANTS .................................................................................................54 3.9 RESULTS .......................................................................................................56
4 CHILLER TECHNOLOGY....................................................................................57
7
4.1 INTRODUCTION ..........................................................................................57 4.2 SYSTEM FUNDAMENTALS .......................................................................57 4.3 COMPRESSORS ............................................................................................60
5 EXPERIMENTAL APPARATUS..........................................................................75 5.1 THE TEST APPARATUS ..............................................................................75
5.1.1 Compressor .............................................................................................75 5.1.2 Condenser................................................................................................75 5.1.3 Evaporator ...............................................................................................76
5.2 INSTRUMENTATION AND DATA LOGGING..........................................81 5.2.1 Thermocouples........................................................................................81 5.2.2 Pressure Transducers...............................................................................82 5.2.3 Data logging and calibration of the sensors ............................................82
5.3 TEST PROCEDURE ......................................................................................83
6 TEST RESULTS.....................................................................................................86 6.1 ANALYSIS OF THE COLLECTED DATA..................................................86
7 SIMULATION AND MODELING RESULTS ...................................................104 7.1 R22 and Pure R407C.....................................................................................104 7.2 Modeling of the system charged with R407C and then subjected to leak ....109
7.2.1 Calculating the new mixture properties after the leak ..........................110 7.2.2 Calculating the system performance including capacity, power input and COP 111
8 CONCLUSIONS AND RECOMMENDATIONS ...............................................119 8.1 GENERAL CONCLUSIONS .......................................................................119 8.2 RECOMMENDATIONS FOR FUTURE WORK .......................................121
REFERENCES..............................................................................................................123
8
LIST OF FIGURES Figure 1-1 Ozone distribution over earth’s atmosphere, as on August 1, 1998..............15 Figure 1-2 European Union HCFC production phase out schedule................................16 Figure 1-3 Mean Condensing Pressure comparison of the alternative refrigerants to
HCFC-22 at 50 C CT. ....................................................................................21 Figure 2-1 Various mixture combinations of the Chlorine free R125/R32/R134a gases26 Figure 2-2 Schematic diagram of vapor and liquid leak models ....................................33 Figure 3-1 Ideal binary blends at constant temperature ..................................................36 Figure 3-2 Ideal binary blend at constant pressure .........................................................37 Figure 3-3 Azeotropic Blend with Minimum Boiling point variety, Temperature vs.
concentration ..................................................................................................38 Figure 3-4 Azeotropic binary blend with minimum boiling point, pressure vs.
concentration ..................................................................................................38 Figure 3-5 Zeoptropic blend vapor compression cycle using non-isothermal phase
change.............................................................................................................39 Figure 3-6 Graph illustration of a vapor compression cycle using zeotropic blend with
specific composition.......................................................................................40 Figure 3-7 Ternary blend at constant pressure................................................................41 Figure 3-8 Compression process diagram.......................................................................52 Figure 3-9 Various compression processes represented on a P – V diagram .................54 Figure 4-1 A typical water-cooled chiller system schematic with shell & tube condenser
and evaporator, and with screw or semi-hermetic reciprocating compressor 58 Figure 4-2 A typical Two stage centrifugal chiller (water-cooled) with economizer.....59 .Figure 4-3 Compression process of a typical reciprocating compressor .......................61 Figure 4-4 Capacity variation as a function of SST of a typical reciprocating compressor
........................................................................................................................61 Figure 4-6 A typical dual rotor – male and female - screw compressor assembly .........63 Figure 4-7 A typical dual rotor – male and female - screw compression concept..........63 Figure 4-8 Suction, trap, compression and discharge process of the scroll concept.......64 Figure 4-9 Rotation geometry of motor shaft, journal bearing and driven scroll ..........65 Figure 4-10 Capacity variation of positive displacement compressors versus SST at
three constant SDTs .......................................................................................66 Figure 4-11 A typical single stage centrifugal compressor cross section ......................67 Figure 4-12 Velocity increase by impeller and conversion of velocity into static
pressure in a centrifugal compression process ...............................................68 Figure 4-13 Performance of a typical centrifugal compressor over a range of inlet vane
positions .........................................................................................................69 Figure 4-14 A cut-away section view of the new totally oil-free compressor ................70 Figure 4-15 Cross section view of two stage centrifugal gas compression assembly ....71 Figure 4-16 Shaft and impellers assembled with magnetic bearings..............................72 Figure 4-17 Cross section schematic of the front radial, rear radial and axial bearings
with sensor rings.............................................................................................72 Figure 4-18 Operating map of the new totally oil free and variable speed centrifugal
compressor .....................................................................................................73 Figure 5-1 Heat transfer coefficient as function of boiling regimes and quality inside a
tube, Cengel 1998...........................................................................................77 Figure 5-2 Schematic of the experimental apparatus......................................................78
9
Figure 5-3 Evaporator piping and sensor arrangement isometric schematic ..................78 Figure 5-4 Temperature sensor (Thermocouple) location schematic .............................79 Figure 5-5 Evaporator Thermocouple tip installation schematic....................................79 Figure 5-6 General view of the test rig ...........................................................................80 Figure 5-7 Data acquisition and instrumentation set up of the experimental apparatus .81 Figure 6-1 Evaporator pressure data acquisition for pure R22 .......................................86 Figure 6-2 Evaporator temperature data acquisition vs. time for pure R22....................87 Figure 6-3 Evaporator temperature data acquisition for pure R407C.............................87 Figure 6-4 Evaporator temperature data acquisition for 10% leaked R407C .................88 Figure 6-5 Evaporator temperature data acquisition for 30% leaked R407C ................88 Figure 6-6 Pressure data acquisition for pure R407C .....................................................89 Figure 6-7 Condensing pressure data acquisition for pure R407C .................................89 Figure 6-8 Condensing, average evaporating and suction pressure data acquisition for
10% leaked R407C.........................................................................................90 Figure 6-9 Average evaporating and suction pressure data acquisition for 10% leaked
R407C.............................................................................................................90 Figure 6-10 Average evaporating and suction pressure data acquisition for 30% leaked
R407C.............................................................................................................91 Figure 6-11 Condensing pressure data acquisition for 30% leaked R407C ...................91 Figure 6-12 Average evaporating and suction pressure data acquisition for 40% leaked
R407C.............................................................................................................92 Figure 6-13 Condensing pressure data acquisition for pure 30% leaked R407C ...........92 Figure 6-14 Refrigerant temperature along the evaporator at CT 40 C for pure R22....93 Figure 6-15 Refrigerant temperature along the evaporator for pure R22 .......................93 Figure 6-16 Suction and average evaporating pressures vs. CHWLT at CT 40 and 50 C
for R22............................................................................................................94 Figure 6-17 Suction and average evaporating pressure vs. CHWLT at 40 C and 50 C CT
for 10% leaked R407C ...................................................................................95 Figure 6-18 Suction and average evaporating pressure vs. CHWLT at 40 and 50 C CT
for 30% leaked and topped up........................................................................96 Figure 6-19 Suction and average evaporating pressure vs. CHWLT at CT 40 and 50 C
for 50% leaked and topped up........................................................................97 Figure 6-20 Average evaporating pressure variation vs. pure R22 and percentage leaked
R407C at various CHWLT for 40 C CT ........................................................97 Figure 6-21 Average evaporating pressure variation vs. pure R22 and percentage leaked
R407C at various CHWLT for 50 C CT ........................................................98 Figure 6-22 cooling capacity variation vs. pure R22 and percentage leaked R407C at
various CHWLT for 40 C CT ........................................................................99 Figure 6-23 cooling capacity variation vs. pure R22 and percentage leaked R407C at
various CHWLT for 50 C CT ........................................................................99 Figure 6-24 Power input variation vs. pure R22 and percentage leaked R407C at various
CHWLT for 40 C CT ...................................................................................100 Figure 6-25 Cooling capacity variation vs. pure R22 and percentage leaked R407C at
various CHWL for 50 C CT.........................................................................101 Figure 6-26 Coefficient of Performance (COP) variation vs. pure R22 and percentage
leaked R407C at various CHWLT for 40 C CT...........................................102 Figure 6-27 Coefficient of Performance (COP) variation vs. pure R22 and percentage
leaked R407C at various CHWLT for 50 C CT...........................................103 Figure 7-1 Genetron software calculation results screens for R22 ...............................105 Figure 7-2 Genetron software calculation results screens for R407C ..........................106
10
Figure 7-3 Capacity and power input comparisons, Genetron vs. test results, R22 at 40 C CT .............................................................................................................107
Figure 7-4 Capacity and power input comparisons, Genetron vs. test results, R22 at 50 C CT .............................................................................................................107
Figure 7-5 Capacity and power input comparisons, Genetron vs. test results, R407C at 40 C CT ........................................................................................................108
Figure 7-6 Capacity and power input comparisons, Genetron vs. test results, R22 at 40 C CT .............................................................................................................108
Figure 7-7 Mass fraction change of the ternary blend R134a/R32/R125 and polynomial curve fitted models .......................................................................................110
Figure 7-8 Comparison of the capacities from the test results and the modeling at various CHWLTs and various leak percentages, 40 C CT ..........................112
Figure 7-9 Comparison of the capacities from the test results and the modeling at various CHWLTs and various leak percentages, 50 C CT ..........................113
Figure 7-10 Power input comparison from the test results and the modeling at various CHWLT temperatures and various leak percentages, 40 C CT ...................114
Figure 7-11 Power input comparison from the test results and the modeling at various CHWLTs and various leak percentages, 50 C CT .......................................115
Figure 7-12 COP comparison from the test results and the modeling at various CHWLTs and various leak percentages, 40 C CT .......................................116
Figure 7-13 COP comparison from the test results and the modeling at various CHWLTs and various leak percentages, 50 C condensing ..........................117
11
LIST OF TABLES Table 1–1 Montreal Protocol Production Caps...............................................................16 Table 1–2 European Union ban schedule on use of HCFC-22 in the new equipment....17 Table 1–3 The United States phase out schedule for the HCFCs ...................................17 Table 1–4 Characteristics of the HCFC22 and potential substitutes...............................20 Table 3–1 m and n constants...........................................................................................43 Table 3–2 Calculated ijk values.......................................................................................44 Table 3–3 Constants for vapor pressure equations .........................................................55 Table 3–4 Constants for ideal gas heat capacity equation ..............................................55 Table 4–1 Chiller Technology alternatives – Vapor Compression cycle .......................60 Table 5–1 Details of the conditions of series of the tests conducted upon. ....................85 Table 7–1 The calculated new compositions after the system subjected to various leak
percentages without adding..........................................................................111 Table 7–2 The calculated new compositions after the system subjected to various leak
percentages and topping up the system ........................................................111
12
NOMENCLATURE
Avg EP Average Evaporating Pressure
c Compressor clearance fraction
CFC Chloro Floro Carbon
Const Constant value
COP Coefficient of Performance
Cp Specific heat at constant pressure
Cv Specific heat at constant volume
CP Condensing Pressure
CT Condensing Temperature
DX Direct Expansion
EP Evaporating Pressure
ET Evaporating Temperature
GWP Global Warming Potential
GTD Gliding Temperature Difference
h Enthalpy
HCFC Hydro Chloro Floro Carbon
HFC Hydro Floro Carbon
HVACR Heating, Ventilation, Air Conditioning and Refrigeration
ID Internal Diameter
IGV Inlet Guide Vanes
LCHWT Leaving Chilled Water Temperature
LLSL-HX Liquid Line Suction Line Heat Exchanger
N Number of cylinders
n Polytropic Exponent
ODP Ozone Depletion Potential
P Pressure
Pa Pressure component a
Pb Pressure component b
Pc Critical pressure
PD Piston displacement
Pt Total pressure
13
Q Heat transfer rate
RPM Compressor speed (rev/min)
S Entropy
SP Suction Pressure
T Temperature
Tc Critical temperature
TEWI Total Environment Warming Impact
UV Ultra Violet
v Specific volume
W Compressor power
X Mole fraction of liquid in mixture
aRiX 134 Mass fraction of R134a in liquid after the leak
125RiX Mass fraction of R125 in liquid after the leak
32RiX Mass fraction of R32 in liquid after the leak
aRiXN 134 New mass fraction of R134a in the liquid after leak and topping up
125RiXN New mass fraction of R125 in the liquid after leak and topping up
32RiXN New mass fraction of R32 in the liquid after leak and topping up
iLR Leak mass fraction
Y Mole fraction of vapor in mixture
Z Compressibility factor
14
1 INTRODUCTION
1.1 PREFACE
The HVACR industry faces two major environmental challenges today: stratospheric
ozone depletion and global climate change.
Stratospheric Ozone Depletion is believed to be caused by the release of certain
manmade ozone-depleting chemicals into the atmosphere. A compromised ozone layer
results in increased ultraviolet (UV) radiation reaching the earth’s surface, which can
have wide ranging health effects. Legislation has been enacted worldwide through the
Montreal Protocol to phase out the production of these chemicals. This phase out is in
progress, and the scientific evidence indicates repair to the ozone layer is underway.
Global climate change is believed to be caused by buildup of greenhouse gases in the
atmosphere. The primary greenhouse gas is carbon dioxide (CO2), created by fossil-
fuel-burning power plants. These gases trap the earth’s heat, causing global warming.
Legislation is not yet in place, but any policy changes or legislation will be enacted
through the Kyoto Protocol. CFC and HCFC refrigerants used by the HVACR industry
are suspected ozone-depleting substances. CFC, HCFC and HFC refrigerants are
considered greenhouse gases. In addition, HVACR equipment is a major power
consumer. Therefore, the industry is part of these environmental challenges.
1.2 WORLDWIDE OZONE DEPELETION LEGISLATION
Worldwide legislation has been enacted through the United Nations environmental
Program (UNEP) to reduce stratospheric ozone depletion. The Montreal Protocol was
approved in 1987 to control production of the suspected ozone-depleting substances,
among them chlorofluorocarbons (CFCs) and hydro chlorofluorocarbons (HCFCs),
commonly used as refrigerants in the HVACR industry. The Montreal Protocol has a
provision to conduct and review future scientific, technical and economic assessments,
and adjust the legislation accordingly. Further evidence in the early 1990s did suggest a
phase out of ozone-depleting substances was in order. Amendments were made to the
15
Protocol in London (1990), in Copenhagen (1992) and in Vienna (1995). No changes
have been made since 1995.
Figure 1-1 Ozone distribution over earth’s atmosphere, as on August 1, 1998
CFCs, which have the highest Ozone Depletion Potential (ODP), were phased out of
production on January 1, 1996. Today, the only available CFC refrigerants for
replacement in existing equipment are those that were stockpiled prior to 1996.
Montreal Protocol initially scheduled HCFCs including HCFC-22 (R22) for total phase
out by 2030. HCFCs Production caps begin January 1, 1996, based on 2.8% of CFCs
used in 1989, weighted by ozone depletion potential (ODP); plus
ODP-weighted 1989 HCFC consumption, thus on January 1:
16
Table 1–1 Montreal Protocol Production Caps Ozone Depleting
Substances Developed Countries Developing Countries
Hydro chlorofluorocarbons (HCFCs)
Freeze: beginning of 199635% reduction by 2004 65% reduction by 2010 90% reduction by 2015 99.5% reduction by 2020 Total phase out by 2030
Freeze: 2016 at 2015 base level Total phase out by 2040
Production and import of HCFCs in developed countries are currently capped at the
1989 baseline values. There will be a 35 percent reduction in the cap in 2004, a 65
percent reduction in 2010, a 90 percent reduction in 2015, a 99.5 percent reduction in
2020 and a complete phase-out in 2030
Developing countries have an additional 10 years from the dates for developed
countries. There is also a monetary fund established to help them make the transition to
alternatives.
The recent Montreal Protocol meeting in Beijing resulted in a proposed amendment to
limit the production of HCFCs. It will require developed countries to freeze the
production of HCFCs in 2004 at 1989 levels and developing countries to do so in 2016
with a 2015 baseline. Production of 15% above baseline will be permitted to meet the
"basic domestic needs" of developing countries. The Beijing amendment will enter into
force after it has been ratified by 20 governments.
Figure 1-2 European Union HCFC production phase out schedule
17
The European Union has accelerated the reduction schedule for HCFCs with an
eventual phase-out in 2015.
Table 1–2 European Union ban schedule on use of HCFC-22 in the new equipment 1/1/2001 HCFC-22
Ban on using in all new refrigeration equipment with a few exceptions
1/1/2002 HCFC-22 Start of the ban on the new refrigeration and air-conditioning equipment with cooling capacity less than 100 kW
1/1/2004 HCFC-22 Start of the ban on the new Reversible Heat Pumps
1/1/2020 HCFC-22 Start of the ban on the use of virgin HCFC-22 refrigerant for service purposes
1/1/2015 HCFC-22 Total phase out
The United States of America implemented the Montreal Protocol agreement as part of
the Clean Air Act Amendments of 1990.
Table 1–3 The United States phase out schedule for the HCFCs 1/1/2003 HCFC-141b
Ban on production and consumption
1/1/2010 HCFC-22/142b Freeze on production and consumption Ban on virgin unless used as a feedstock or refrigerant in appliance manufactured prior to 1 January 2010.
1/1/2015 All other HCFCs Freeze on production and consumption. Ban on all other virgin HCFCs used as a feedstock or refrigerant in appliance manufactured prior to 1 January 2020.
1/1/2020 HCFC-22/HCFC-142b Ban on production and consumption
1/1/2030 All other HCFCs Ban on production, consumption
Note: Consumption is defined as Production plus Imports minus Exports.
18
The United States of America adopted much relaxed policy by imposing freeze on
portion of HCFCs from 1st January 2010, 14 years later than the Montreal Protocol time
table. While, Australia continues to be a world leader in the phase out of ozone
depleting substances, in many cases well ahead of the Protocol requirements. Australia's
approach has been based on a highly co-operative partnership between industry, the
community, and all levels of government
The countries of the European Community have adopted even stricter measures and new
equipment banned from using HCFC-22.
After the phase-out dates, supplies of HCFC-22 should still be plentiful thanks to
conservation and recycling techniques developed for CFCs. For instance, CFC-11 hasn’t
been produced since January 1996, but because of recycling, it is still widely available.
Recycled HCFC-22 should be even more plentiful, because it is more widely used than
CFC-11. It is important to remember that the phase out affects only production, not use.
1.2.1 Kyoto Protocol
The Kyoto protocol offers new incentives for technological creativity and adoption of
kind of solutions that make economic sense irrespective of climate change. Because
activities and products with zero or low emissions will gain competitive advantage, the
energy, transport, industrial, housing and agricultural sectors will gradually move
toward more climate-friendly technologies and practices. The protocol targets 5.2%
average greenhouse effect reduction and 6 greenhouse gases in 38 countries with
commitment period of 2008-2012 and demonstrated progress to goal by 2005 as well as
banking and carry over of reductions.
1.2.2 HCFC-22 (R22) Substitutes
HCFC-22, introduced about 60 years ago, is the world’s most widely used refrigerant. It
serves in both residential and commercial applications, from small window units to
large water chillers, and everything in between. Its particular combination of efficiency,
capacity and pressure has made it a popular choice for equipment designers. Recently,
extensive use of HCFC-22 has made it possible to reduce the use of CFC refrigerants,
because its Ozone Depletion Potential (ODP) is as much as 95% lower than CFCs.
19
Nevertheless, it does have some ODP, so international law set forth in the Montreal
Protocol and its Copenhagen and Vienna amendments have put HCFC-22 on a phase-
out schedule. In developed countries, production of HCFC-22 will end no later than the
year 2030. In intervening years, production is reduced in a series of specified steps.
Detailed phase-out schedules vary from country to country
The evaluation of the Chlorine free substitutes for HCFC-22 indicates that the
possibilities of direct, comparable single substance refrigerants are limited. Apart from
Ammonia and hydrocarbons, with their special application criteria, as a single substance
only the refrigerants HFC-32, HFC-125 and HFC-134a remain as direct potential
substances for HCFC-22.This is due to their specific characteristics such as zero ODP,
flammability and thermodynamic properties. A refrigerant blend containing these
substances could match characteristics of HCFC-22.
Obviously, new substitutes should possess what HCFC-22 lacks from an environmental
standpoint, while retaining its good thermodynamic characteristics. Environmentally,
the ideal candidate must have an ODP of zero and a low Global Warming Potential
(GWP). Thermodynamic similarity is important when retrofitting existing chillers;
otherwise, capacity and efficiency may be lost. In new chiller designs, thermodynamic
similarity is less of an issue because the equipment can be specifically designed for the
replacement refrigerant’s particular characteristics.
Currently, there are four leading replacement candidates: HFC-407C, HFC-404A, HFC-
134a, and HFC-410A.
1.2.3 Properties of HCFC-22 (R22) Substitutes
All four replacement candidates are Huffs; their molecules do not include chlorine, so
their ODP is zero. As far as GWP, however, the numbers do vary, and according to
Dupont, the GWP ratings are illustrated in Table 1–4.
20
Table 1–4 Characteristics of the HCFC22 and potential substitutes Refrigerants Global
Warming Potential (GWP)
ODP* GTD**K
Critical Temperature
C
Composition
HFC-134a 1300 0 0 101 Single Substance HFC-407C 1600
0 7.4 87 R32/R125/R134a
(Blend) HFC-410A 1890 0 <0.2 72 R32/R125 (Blend) HFC-404A 3750 0 0.7 73 R143a/R125/R134a
(Blend HCFC-22 1700 0.05 0 96 Single Substance Notes: (*) Ozone Depletion Potential (ODP) with reference to CFC-11(R-11) = 1.0
(**) Gliding Temperature Difference (GTD), difference between bubble and dew points at 100 kPa absolute pressure. This value varies depending on the pressure.
While there is a variance in these numbers, it is not significant. According to the TEWI
(Total Equivalent Warming Impact) formula being proposed by GWP researchers, there
are two ways in which refrigeration system can affect global warming: refrigerant leaks
and electric-energy consumption. According to refrigerant manufacturers, all HCFC-22
substitutes will be available for the long-term. All are non-flammable and have low
toxicities like HCFC-22. However, their capacities, efficiencies, and pressures vary,
which affects their application.
1.2.4 Retrofitting Existing systems
To be successful in a retrofit situation, a replacement refrigerant needs to have capacity,
efficiency and pressure characteristics comparable to HCFC-22. These thermodynamic
properties will ensure satisfactory equipment capacity and efficiency.
21
0
500
1000
1500
2000
2500
3000
3500
1
Refrigerant
kPa
HCFC-22HFC-134aHFC-407CHFC-404AHFC-410A
Figure 1-3 Mean Condensing Pressure comparison of the alternative refrigerants to HCFC-22 at 50 C CT.
HFC-407C is very similar to HCFC-22 in all thermodynamic respects, which makes it a
good retrofit candidate. Consequently, HFC-407C is being used for retrofits in some
areas now. The capacity of HFC-134a is much lower than HCFC-22. The fact that HFC-
134a requires about 35% more volumetric capacity on the compressor usually results in
substantial de-rate, when used as a drop-in replacement for HCFC-22. So it is not a
good retrofit candidate. The pressures of HFC-404A and HFC-410A are much higher
than HCFC-22, so they are not good retrofit candidates either.
1.2.5 Refrigerant Solutions for Today’s Environmental Challenges
Table 1–4 shows several alternative refrigerants for the systems that use positive
displacement compressors. There are enough differences between each of them that one
refrigerant would not be the likely choice for all chillers. Each manufacturer will
determine which refrigerant makes the most sense for its product line replacement for
HCFC-22.
Progress is being made, and new systems using HFC-407C, HFC-410A and HFC-134a
are now entering the market to replace HCFC-22. The incremental phase out of HCFCs
was a necessary step to make a smooth transition from CFCs to more environmentally
friendly refrigerants. The fact that HFC-134a requires about 35% more volumetric
capacity on a compressor usually results in substantial de-rate, when used as a drop-in
replacement for HCFC-22.
22
HFC-407C is a high-pressure refrigerant, similar to HCFC-22 with ODP of zero. As
such, it is a long-term environmental solution to ozone depletion and is not scheduled
for production phase-out. HFC-407C is a blend refrigerant with a low toxicity level. In
selecting a refrigerant to replace HCFC-22, HFC-407C was found to be an ideal
alternative for air-cooled DX chillers. Its operating characteristics are so similar to
HCFC-22 that it has been used as a drop-in alternative in machines originally designed
for HCFC-22. This allows the continued use of the proven design and components of
air-cooled chillers. Additionally, a unique feature of HFC-407C – the property known as
“temperature glide” – presents intriguing implications for DX chiller design. This
temperature glide phenomenon, when matched with DX cooler design modifications,
can improve energy efficiency. HFC-407C is a blend of three refrigerants: HFC-32,
HFC-125 and HFC-134a. This composition exhibits the characteristics of a zeotropic
blend, meaning that the resulting mixture does not act as a single substance. At a given
pressure, it evaporates over a range of temperatures, rather than at a single temperature.
Thus, the term, temperature glides.
HFC-407C has a gliding temperature difference (GTD) of approximately 6°K, which
can be leveraged to give opportunities for greater operating efficiency. If the DX
evaporator is designed as a counter-flow heat exchanger, refrigerant and water enter at
opposite ends, and the leaving refrigerant temperature can be greater than the leaving
chilled water temperature. The higher leaving refrigerant temperature means the
compressor does less work, resulting in lower power consumption.
There has been more focus on R407C as a possible R22 substitute. Preliminary
theoretical analyses based on its thermodynamic properties, clearly shows R407C as the
best possible drop in substitute for R22. R407C is a zeotropic mixture of
R32/R125/R134a (23/25/52% in weight) with a gliding temperature difference in
evaporation and condensation of average about 6°K. R407C can be considered as a long
term substitute of R22 because it is an HFC.
HFC-410A is a high-pressure refrigerant, having approximately 50% higher pressure
than HCFC-22. As a result, it can provide significant capacity gain to a compressor
designed to handle the pressure. HFC-410A is a leading candidate for unitary residential
and commercial equipment. HFC-410A is a blend refrigerant consisting of HFC-125
and HFC-32 (50/50%). It has a low temperature glide of 0.5 °C, small enough to have
23
little or no effect on heat exchanger performance. HFC-410A has an ODP of zero. As
such, it is a long-term environmental solution to ozone depletion and is not scheduled
for production phase-out. However, due to its high pressure, the existing systems need
to be redesigned, manufactured and certified for the new pressure level. This in turn
poses enormous challenges in terms of engineering and manufacturing costs while try to
stay competitive in a price driven market. Therefore, HFC-410A cannot be considered
as a drop-in alternative for the existing R22 systems.
1.2.6 HCFC-22 Phase out and Recycling
International law set forth in the Montreal Protocol and its Copenhagen, Vienna and
Kyoto amendments have put HCFC-22 on a phase out schedule. In developed countries,
production of HCFC-22 will end no later than the year 2030. In the years leading up to
the proposed year, production is reduced in a series of specified steps. Detailed phase
out schedule vary from country to country.
In Australia, HCFC-22 is frozen since 1996 and has become a controlled substance.
While in the United States, production of this gas will be frozen on at baseline level on
January 1, 2010, and the production of virgin refrigerant will be banned unless it is used
as a feed stock for other refrigerants, or equipment manufactured prior to this date.
The countries of the European Community have more restricted measures by banning
all new equipment with HCFC-22.
In the mean time, during the scheduled cut in production and phase out, supplies of
HCFC-22 should still be available due to conservation and recycling techniques
developed during the last few years.
In Australia, Industry Codes of Practice is designed to minimize escape of HCFC-22 gas
into the atmosphere. It is expected that, due to the cost involved, the recycled HCFC-22
will become far more expensive than the alternative gas like HFC-407C.
1.3 PURPOSE OF THIS RESEARCH
The purpose of this research is to investigate behavior of R407C refrigerant in chiller
systems. This includes performance and efficiency variations when it replaces R22 in an
existing system as well as challenges involved maintaining the system charged with
24
R407C. Various researches have been conducted on this refrigerant, which were
touched base earlier ranging from performance comparisons in evaporator, condenser
and system to leak simulations and effect of the leak in more general terms.
However, this research is intended to address challenges faced in the real world and
practical terms. First, to evaluate precisely the performance of a given R22 chiller when
charged with R407C at different conditions that are relevant to real design and operating
conditions. Second, to investigate chiller systems that subjected to substantial gas leaks.
These cases are not uncommon in the industry that, due to composition change of the
remaining refrigerant, requires replacement of entire gas.
Majority of chillers are large machines with high capacity and are installed in plant
rooms of central cooling systems. They contain large amount of refrigerant, which can
range from about 50 kg up to 1000 kg. Thus expensive exercise as chillers contain large
amount of refrigerant, while, it is desirable to top up the leaked gas to the required
amount. In reality, majority of the leaks are slow and in vapor form, therefore, can be
considered as isothermal vapor leaking process. It is important to investigate to what
extend composition change would affect the performance of the system after the leak as
well as after the remaining refrigerant topped up with an original refrigerant.
25
2 LITERATURE REVIEW Popular interest in the use of refrigerant blends started in the late 1950’s.The emphasis
was placed on energy savings through the reduction of irreversibility in the heat
exchanger and on capacity variation during operation through the control of the fluid
composition. Cooper (1982) performed tests to determine the heating and cooling
performance of standard air conditioned unit using the refrigerant mixture of
R13B1/R152a to replace R22. An improvement of 27% in COP was obtained for
heating at very low outdoor temperatures, while an improvement of only 4% could be
achieved for cooling.
Kruse (1998) studied chlorine free azeotropic and zeotropic mixtures to understand
concerns about the internal composition shift, differential oil solubility and the leakage.
When leakage occurs in a system, the remaining mixture has higher concentration of the
higher boiling point component than the original blend. In order to evaluate this effect,
ternary zeotropic blend of R32/R125/R134a and azeotropic blend of
R143a/R125/R134a were studied. While the separation effect by leakage with the
azeotropic mixture is negligible and below 1% until 40% leaked out, a change of
composition between 5 to 10% could be noticed with the zeotropic mixture.
26
Figure 2-1 Various mixture combinations of the Chlorine free R125/R32/R134a gases
Due to the composition shift during evaporation process, Zeotropic mixtures are not
suitable for flooded evaporators,. Composition shift of up to 3% have been calculated
with dry expansion evaporators using a simulation program. Therefore, more
investigations are required on application of zeotropic mixtures in the dry expansion
systems.
Differential solubility of the individual component in the oil is another concern with the
refrigerant mixtures. This effect applies to both azeotropic and zeotropic refrigerant
mixtures depending on their individual behavior with the oil which influenced by
temperature and pressure in the system, especially in the oil receiver. Concentration
shifts by oil solubility of around 5 to 7% for the low boiling components have been
measured in a water chiller charged with ternary mixture R32/R125/R134a.
Tolouee et al (1991) examined blend of R22/R142b as a drop in replacement for R12. A
computer simulation program using refrigerant blends had been developed and the
predictions compared with the test results. The computer model predicted the
experimental values with a maximum deviation of five percent. It became evident that
the saturation pressure of a mixture with 60% mass fraction of R22 closely follows that
of R12.
27
Gabrielii et al (1998) compared R22 and R407C in terms of their thermodynamic and
energetic performances. The experimental tests have been carried out in a vapor-
compression refrigerating plant. The final conclusion reached on the basis of the
experimental results is that the efficiency of R407C is significantly lower than that
pertaining to R22. Therefore, a plant working with R407C requires higher electric
power consumption in order to provide the same refrigerating load. Besides the direct
costs, environmental impact is the indirect cost, since more fuel must be burned and
higher amounts of carbon dioxide discharged in the atmosphere. The efficiency of
R407C must be increased substantially prior to any practical application. To achieve
this, practical improvements that suggested by a detailed analysis on each device of the
plant need to be applied. In most cases the thermodynamic design of the plant is perhaps
more focused towards R22 characteristics.
They suggested that the analysis should start from the heat exchangers of the vapor
compression system. In the present works attention has been focused on the evaporator.
The mean heat transfer coefficient of R407C has been evaluated and compared with
R22. By determining these coefficients extra costs due to under design or over design of
evaporators, boilers and other two-phase process equipment can be avoided..
Goto et al (2001) studied condensation and evaporation heat transfer of R410A inside
internally grooved horizontal tubes. Spiral groove tube of 8.01 mm o.d. and 7.30 mm
mean ID, and a herring-born groove tube of 8.00 mm o.d. and 7.24 mm mean i.d. were
used in the experiments. Condensation and evaporation heat transfer coefficient, and
pressure drop were measured with the system was charged with both R410A and R22.
Test results indicated that local heat transfer coefficients of the herring-bone grooved
tube are about twice as large as those of spiral one for condensation and are slightly
larger than those of spiral one for the evaporation. The measured local pressure drop in
both condensation and evaporation is well correlated with the empirical equation
proposed by the authors. The conclusion is that the herring-bone grooved tube is more
effective in enhancing evaporation and condensation heat transfer than the conventional
spiral groove tube. It is possibly that the enhancement of heat transfer during the
evaporation and the condensation is made by the mixing of liquid film of refrigerant and
increasing the effective heat transfer surface area on heat transfer, where the thin liquid
28
film forms. However, the liquid film flow mechanism inside these tubes is not
explained.
The local evaporation and condensation heat transfer coefficient calculated using basic
heat transfer correlations and experimental data. Refrigerant side local equilibrium
quality x is calculated by means of local heat balance and local enthalpies as:
xin=(hin-hLin)/(hVin-hLin), xout=(hout-hLout)/(hVout-hLout) 2.1
The equation predicted the pressure drop inside a grooved tube within ±20% for both
the evaporation and the condensation heat transfer of R22 and R410A.
An experimental apparatus made up of a heat pump refrigerant-loop and two water-
loops was used. The refrigerant-loop consisted of a compressor, an oil separator, a four-
way valve, a condenser, a mass flow meter, an expansion valve, an evaporator and an
accumulator. A constant temperature water tank, a pump, a flow control valve, and a
flow meter used on each water loop to supply the heating/cooling water to the condenser
and the evaporator, respectively. The test heat exchanger consisted of four 2-m long
tube in tube type test sections with the total working length of 8 m. The refrigerant
flows inside an inner tube and the water flows between the inner and outer tubes at a
counter flow regime.
The accuracy of the measurements were predicted to be within ±0.15 K for ‘T’ and K-
type thermocouples, within ±0.05 K for Pt resistance thermometers, within ±0.1% for
the refrigerant mass flow meter, within ±0.3% for the water flow meter, within ±3.9 kPa
for the absolute pressure transducers and within ±0.3 kPa for the differential pressure
transducers, respectively. The accuracy of the measured heat transfer coefficient and
frictional pressure drop PF were estimated within about ±40% and ±5%, respectively.
Aprea et al (2000) carried out experimental evaluation on evaporative heat transfer
coefficient of R22 and R407C in a vapor compression plant. The mean heat transfer
coefficients of a coaxial counter-flow heat evaporator (20 mm ID) in a vapor
compression system were measured. A typical small size refrigeration system working
conditions used as test parameters with the capacity ranging from 1.9 kW to 9.1 kW and
the mass flux varied from 30 to 140 kg/m2 s. The results showed that the R22 heat
29
transfer coefficient is always greater than that of R407C. The over-predicted R407C
coefficients are in line with the most of the theoretical predictions by credited literature.
(a) An overview of heat transfer correlations
A large number of correlations have been proposed to predict the heat transfer
coefficient in forced convective boiling in horizontal and vertical smooth tube for pure
substances and mixture. The prediction of heat transfer coefficient remains essentially
empirical due its complex hydrodynamics and heat transfer processes involved. The first
general correlation for saturated flow boiling of pure substances is the well-known Chen
(1997) correlation. At the start of boiling, both nucleate boiling and liquid convection
may become effective heat transfer mechanisms. While the quality is low, the vapor
void fraction is relatively low and the nucleate boiling is much stronger than the forced
convective effect. Downstream the flow vaporization occurs, the void fraction rapidly
increases. Consequently, the flow must accelerate, which tends to enhance the
convective transport of heat from the heated wall to the evaporating fluid.
(b) Test apparatus and test conditions
A refrigeration system consisted of a semi-hermetic vapor compression compressor, a
coaxial counter-flow condenser connected to a liquid receiver, a coaxial counter-flow
evaporator and a manual expansion device. A Coriolis-effect flow meter with accuracy
of ±0.2% is fitted at the outlet of the liquid receiver to measure refrigerant flow rate. All
the probes are calibrated in the proper operating range in a thermostatic bath equipped
by a precision platinum resistance thermometer; the estimated uncertainty is ±0.1 K.
Absolute pressure throughout the refrigerant loop is measured by piezoelectric pressure
transducers with a range of 0 to 700 or 0 to 3000 kPa and accuracy of ±0.5% of the full
scale value. Volumetric water flow rate is measured using a turbine flow meter with an
uncertainty of ±0.25%.
A typical working condition for small-scale refrigeration systems is used to determine
heat transfer coefficients of the evaporator. The evaporating pressure ranging from 360
to 420 kPa, quality after expansion of 15-20%., superheat 1-2 K, heat flux 1.9-9.11
kW/m2, refrigerant mass flux of 30-140 kg/m2 s and glycol solution mass flow rate:
0.16-0.30 kg/s.
30
Test results showed the R22 heat transfer coefficient is always greater than R407C for
any given refrigerant mass flux, while, the difference percentage decreases with
increasing refrigerant mass flux.
Experimental heat transfer coefficient for R22 and R407C were compared to the heat
transfer coefficient evaluated with the correlations of Shah (1982), Kandlikar (1989)
and Chen (1966) with about ±20% deviation over 80% of the experimental points.
Similar results obtained with the comparison of the heat transfer coefficients predicted
with the correlations of Gungor et al (1986), Yoshida et al (1994). and Kattan et al
(1998), with the experimental values.
In conclusion, it is shown that the heat transfer coefficient of R22 is always higher than
that of R407C. The difference ranges from 54 to 24% which decreases with increasing
the refrigerant mass flux.
The experimental data have been compared with predictions by available correlations.
Predictions by these correlations were compared with the experimental data. Theoretical
calculations results were under predicted for R22 and over predicted for R407C in most
of the cases. The Kandlikar (1989) correlation seems to be the best-fitting correlation
for R22 experimental data, while, the Gungor et al (1986) correlations seem to be the
best-fitting correlations for R407C.
Johansson et al (2000) have developed a method to estimate the circulated composition
in refrigeration and heat pump systems using zeotropic refrigerant mixtures. The
method has been tested and evaluated on a well-equipped experimental rig and tests
show that it is possible to estimate the composition of the circulated refrigerant mixture
to within 2%, by measuring only two temperatures and pressures.
When a system is charged with zeotropic refrigerant mixtures, the circulated
composition may change from the nominal to a different composition. The composition
may change in the system due to a few reasons such as leakage one phase or
accumulation of refrigerant in liquid or vapor phase in certain cycle components. The
shifts in composition would affect thermodynamic characteristic of the system by means
of capacity, efficiency, pressures, temperatures, etc. It is necessary to know the
circulated mixture properties at all stages of the cycle in order to evaluate the system
properly.
31
The molar fraction is used in the leakage model for computational reasons. While the
nominal mass fraction composition of R407C is R134a=0.52, R32=0.23 and
R125=0.25, its nominal mole fraction composition becomes R134a=0.4393,
R32=0.3811 and R125=0.1796.
A container with the volume V is presumably charged with nz moles of nominal
composition of R407C at the temperature T, until a vapor mole quality Q has been
obtained. At the equilibrium state ny, i mol of the i-th component is in vapor phase and
nx, i mol is in liquid phase. The fraction of the i-th component in the bulk, vapor and
liquid is zi, yi and xi, respectively.
Bulk: zR125=1-zR134a-zR32 2.2Vapor: yR125=1-yR134a-yR32 2.3Liquid: xR125=1-xR134a-xR32 2.4
Two major leak types of isothermal and adiabatic leakage have been studied. A small
leak from a system simulates the isothermal and a fast leak simulates the adiabatic.
The temperature, T, of the refrigerant is kept constant in the isothermal model; the
system pressure slowly decreases with vapor leakage, and slowly increases in the liquid
leak case. The bulk is slowly enriched with the two more volatile components, R32 and
R125, in the liquid leak case, while it is slowly enriched with the less volatile
component, R134a.
Both vapor and liquid leakage simulations have been performed at three different
temperatures -10, 0 and +10°C.
In the adiabatic leakage case no heat transfer is allowed through the container wall, the
temperature and pressure of the system drops rapidly. In other words, the internal
energy at any stage is the sum of the internal energy of the liquid and vapor phase
before the leak, minus the enthalpy of the leaked out refrigerant.
Refrigerant leakages have been simulated using purpose built vessel for isothermal and
adiabatic leak. The composition of the bulk, liquid and vapor changes, when refrigerant
vapor or liquid leaks out from a vessel containing two phase equilibrium of a zeotropic
refrigerant mixture. The test results shows that as the vapor leak continues, during an
isothermal vapor leak, share of R134a in the bulk increases, while, it decreases during
an isothermal liquid leak at 10°C, the differences in initial composition between bulk
32
and vapor phase are smaller at higher temperatures, and larger at lower temperatures. It
was also observed that the composition change of the vapor phase is much higher than
liquid phase in an adiabatic vapor leak.
The author has also simulated two consecutive leaks. A bottle charged with nominal
composition of R407C allowed leaking vapor out to another initially evacuated bottle
and accumulated there. In the second bottle, the accumulated composition is richer with
R32 and R125. When the refrigerant in the second vessel subjected to a new isothermal
leak, the accumulated bulk composition follows near vicinity of the same normal leak
curve.
It has concluded that compositions actually fit themselves more or less on a curve in any
leak process. In the case of R407C, this curve can be explained by ternary function or
mole fraction of the first component (R134a) as function of mole fraction of the last one
(R32). The latter was curve fitted to the following polynomial function, provided that
the original composition was the nominal composition of R407C.
Polynomial correlations also can be generated for the other ternary zeotropic refrigerant mixtures, as they show similar leak behavior to R407C.
Circulated composition in a refrigeration system was estimated using in-situ method by
means of iterative solution of multiple equations. So, three equations are needed to find
mole fraction of each component in the circulated R407C as it composed of three
components, R134a, R32 and R125. The first equation is based on the principle of
adiabatic expansion that means enthalpy of the refrigerant is constant before and after
the expansion Equation 2.6. The second equation is the conservation of mass theory
Equation 2-7, and the third equation is polynomial equation generated earlier i.e.
Equation 2.5, with the assumption of composition shift, in case, due to leakage or other
reasons.
h1s(p1s,t1s,R134a,R32,R125)=h2s(p2s,t2s,R134a,R32,R125) 2.6 R125=1-R134a-R32 2.7
The estimated circulated composition values using this method with absolute deviation
of approximately 2% compare to values measured by gas chromatography.
33
Kim et al (1995) simulated leak processes of zeotropic refrigerant mixtures. Two ideal
leak scenarios isothermal and adiabatic leak processes were considered in this study.
The isothermal leak process represents an ideal case of a very slow leak, and the
adiabatic leak process represents an ideal case of a very fast leak from a system or a
refrigerant storage bottle. Two refrigerant mixtures R-32/134a (30/70) and R-
32/125/134a (30/10/60) have been selected for calculations.
The modeled system is consisted of a cylinder with a small hole through to let the
refrigerant leak through in liquid or vapor forms, as shown in the schematic diagrams
(a) and (b) of Figure 2-2.
Figure 2-2 Schematic diagram of vapor and liquid leak models
zi is defined as the overall mole fraction of the i-th component in the cylinder, which is
sum of the mole fractions of the component i in both liquid and vapor phases. While, xi
is the mole fraction of the i-th component in the liquid phase and yi is the mole fraction
of the i-th component in vapor phase.
T, P, V, Z
nV y i
VAPOR
T, P, V, Z
nV yi
VAPOR
LIQUID xi
ni
iv ynn .∆=∆
LIQUID xi
ni
ii Xnn .∆=∆
(a) Vapor leak (b) Liquid leak
34
Initially, the cylinder is assumed full of the saturated liquid of mole fraction xi,
thus xi = zi, the overall mole fraction. During the isothermal leak process, Pressure, and
both liquid and vapor mole fractions are interactive. Liquid and vapor mole fractions are
determined from the equation of state, once temperature and pressure are given.
Once all the liquid is evaporated and the system is full of saturated vapor leakage from
this point would result in pressure decrease and no further composition shift.
During isothermal leak of R-32/134a (30/70) and R-32/125/134a (30/10/60) mixtures,
vapor and liquid mass fractions of the more volatile component (YR-32 and XR-32)
decreases. While, in adiabatic leak case, liquid mass fraction of the more volatile
component (R-32) decreases and the vapor mass fraction of this component increases.
In fact, a refrigerant charging process is close to the adiabatic leak process, and during
this charging process, the mass fraction change for a zeotropic mixture is negligible.
Man-Hoe Kim (2002) evaluated Performance of R-22 and four alternative refrigerants
R-134a, R-32/134a(30/70%), R-407C, and R-410A. Testing at operating conditions
typical for a residential air conditioner with pure cross-flow condenser and counter-flow
evaporator heat pump carried out the experimental evaluation. Counter- flow evaporator
and cross flow condenser with water/ethylene and water used as heat transfer fluid.
Test results at 1800 RPM indicated similar capacities for R-22, R-407C and R-32/134a,
while the capacity of R-410A is 40% greater than the R-22, and the capacity of R-134a
is 32% lower than R22. As for COP, R22 and R407C have almost the same; R32/R134a
is 4.7% higher and R134a only 2% improvement compare to R22, while R410A was 7%
worse.
However, at 1000 RPM, COP of R410A is 22% higher and R134a much lower than R22
and R407C. The higher RPM resulted in negative effect on the R134a efficiency (COP)
due to the higher friction losses and excessive pressure drop across the heat exchanger
as a consequence of high specific volume of the refrigerant. The condenser temperature
profiles showed steeper temperature slope in zeotropic refrigerant than R22. This is due
to the temperature glide associated with zeotropic mixture’s composition shift and
influence of pressure drop during a phase change.
35
The temperature profile of R407C showed flatter slope than that of R22 in the
evaporator, which indicates that phase change pressure drop offset temperature glide
during the evaporation process.
The two zeotropic mixtures R-407C and R32/R134a had a discharge pressure similar to
that of R-22, and the R-134a pressure was lower, while R410A had the highest
discharge pressure. All three zeoptropic refrigerants R410A, R407C and R32/R134a had
mass flow rate of within 10% of the mass flow of R22, but mass flow rate of T134a was
much lower.
This research concludes that zeotropic mixtures, R-407C and R-32/134a (30/70), have
the closest performance characteristics to R-22, with R-32/134a having a slightly better
COP. The major thermodynamic properties such as evaporation and condensing
pressures did not deviate significantly from the R-22 values. The binary near-azeotrope,
R-410A, displayed a 44% higher capacity than R-22 when tested at the same
compressor rpm. However, at a reduced compressor speed at which R-410A capacity
matched that of R-22, the COP of R-410A was 22% better than the COP of R-22. It has
to be realized that this COP improvement resulted from significantly lower pressure
losses especially in the evaporator, suction line, and at compressor valves and from
reduced friction losses in the compressor running at a lower speed.
Use of the LLSL-HX, in other words economizer, improved COPs of all fluids tested
with exception of R-410A which was not included in these tests. Presence of two phase
refrigerant on the suction side did not help to improve COP for R-22 and R-134a.
However, for the zeotropic mixtures, R-407C and R-32/134a, a COP improvement was
measured even at small superheat values leaving the LLSL-HX. Although the COP
increases using LLSL-HX was predicted in theory, but has not been quantified. The
author therefore does not recommend any further study on that direction. Effect of oil
miscibility has also been mentioned as another factor in the COP increase as using the
same oil with different refrigerants would have advantaged some and disadvantaged
other refrigerants heat transfer. That is why it is difficult to objectively evaluate the
performance potential of different fluids; even if the tests are performed using the same
laboratory apparatus. The use of the same oil may penalize the refrigerants those are not
miscible with it.
36
3 THEORY
3.1 INTRODUCTION
In a binary solution, the fluid properties ideally show a proportional relationship with
the concentration of the mixture. According to Raoult’s Law, partial vapor pressure of
each compound in a liquid solution is proportional to the concentration of an ideal
mixture. Total pressure will be equal to sum of the partial pressures, when Dalton’s Law
is satisfied at the low total pressure, see Figure 3-1 and 3-2. However, the composition
of vapor in equilibrium with the liquid will be richer in the more volatile component
than the composition in the liquid phase. When the deviation from Raoult’s Law
becomes very significant, the total vapor pressure curve may pass through a maximum
or minimum, Figures 3-3 and 3-4.
0 20 40 60 80 100
Pt
Pa
Pb
T = Constant
Figure 3-1 Ideal binary blends at constant temperature
37
0 20 40 60 80 100
Figure 3-2 Ideal binary blend at constant pressure
3.2 AZEOTROPIC BLEND
In this case, the composition of vapor in equilibrium must be the same as liquid at that
specific concentration. This blend is referred as azeotropic composition, which behave
just like pure refrigerants. The 500 series azeotropic mixed refrigerants are commonly
used in the industry for quite long time. Their vapor pressure is higher than of either
component. For example, azeotropic blend of R502 is composed of R22 and R115 and
has slightly higher pressure than R22; it is used in low temperature applications instead
of R22 where high discharge pressure may have adverse effect on the system.
38
0 20 40 60 80 100
Figure 3-3 Azeotropic Blend with Minimum Boiling point variety, Temperature vs. concentration
0 20 40 60 80 100
T = Constant
Figure 3-4 Azeotropic binary blend with minimum boiling point, pressure vs. concentration
39
3.3 ZEOTROPIC BLENDS
Mixtures that their total pressure curve at any concentration does not pass through a
maximum value are defined as zeotropic blends. The composition difference between
liquid and vapor cause non-isothermal condensation and evaporation. This variation
which is obvious on dew point and bubble-point curves is referred as gliding
temperature difference (GTD). Ignoring pressure drop effects, pure refrigerants phase
change process takes place at a constant saturated temperature corresponding to the
condenser or evaporator temperatures. There is no constant condensing and evaporating
temperatures with the zeotropic blends, but rather a range of gliding temperature.
Figure 3-5 Zeoptropic blend vapor compression cycle using non-isothermal phase change
A graphic illustration is used to explain thermodynamic cycle of a basic refrigeration
system charged with a fixed composition of a zeotropic blend, Figure 3-5. An
azeotropic blend behaves like a single substance with theoretically constant evaporation
and condensing temperatures at given pressures. Zeotropic blends however experience a
temperature glide during the evaporation and condensing cycle, because the phase
changes don’t occur isothermally. This is due to the fact that each of the compounds
tends to keep its dominant thermodynamic characteristic while acting as a blend. The
Figure shows a vapor compression cycle using a zeotropic blend where, evaporating
temperature at the inlet t01 is lower than the outlet t02. The difference between these
two temperatures is called Gliding Temperature Difference (GTD). A similar
phenomenon occurs during condensing, where the condensing temperature at the
40
beginning, saturated vapor line tx1, is higher than at saturated liquid line tx2.
Therefore, GTD evap = t02 – t01 and GTD cond = tx1 – tx2.
Vapor-liquid equilibrium properties represented by two upper and lower envelopes at
two pressure levels corresponding to the evaporator P1 and the condenser P2.The cycle
beginning at the compressor; the zeotropic blend is compressed and hot gas temperature
is raised above the corresponding saturated condensing temperature. During
condensation process, the total vapor fraction is reduced and the vapor-phase
composition shifts to the right along the dew-point line reflecting the ongoing depletion
of component A, which is the higher boiling point fluid. Along the bubble point line,
concentration of the component A is initially high, in the liquid phase. However, as the
process proceeds and the liquid fraction increases, the composition shifts to the right
until all of vapor is condensed, where the liquid composition becomes the same as the
initial charge. The temperature declines during entire phase-changing process and
further cooling results in sub-cooled liquid.
Reverse of the condensation process takes place in the evaporation process, with the
more volatile component B evaporating at a higher rate initially
Figure 3-6 Graph illustration of a vapor compression cycle using zeotropic blend with specific composition
41
The non-isothermal evaporation and condensation can be shown on temperature -
entropy and pressure-enthalpy diagrams, see Figure 3-6. These represent steady state
operation at a fixed composition of the zeotropic blend. If the composition change
occurs, then the fluid properties change and would result in a modified diagram.
Ternary blends pressure and temperature behavior can be illustrated in a three
dimensional curves, Figure 3-7. Axis AC, BC and AB represent mass fraction of each
binary when mass fraction of the third is assumed zero. It can be seen that on the AB
axis the binary is Azeotropic, while the other two binaries demonstrate Zeotropic (Non-
Azeotropic) behavior. However, the resulting ternary blend would be Zeotropic blend.
Figure 3-7 Ternary blend at constant pressure
42
3.4 THERMODYNAMIC PROPERTIES DETERMINATION
In this section thermodynamic properties of pure refrigerants are theoretically analyzed
and extended to determination of binary as well as ternary and high order blend
properties. To do this, analyzing Equation of State is the major and the most critical part
of this theoretical study.
Depending on suitability, the program uses an Equations of States for a specific family
of refrigerants. The Equations of State that used in this prediction software are discussed
in this section in details. Pure and mixed refrigerants properties are calculated using
relevant Equation of State to get the best accuracy possible. This method makes it
flexible in order to get the best and most accurate results.
3.4.1 Redlich-Kwong-Soave (RKS) (1980) Equation of State
For each pure component, it is necessary to have the critical temperature and pressure,
molecular weight, ideal gas heat capacity coefficients, and some pressure –temperature
data. It is thus possible, using the RKS equation of state and appropriate thermodynamic
relations to calculate a variety of thermodynamic properties (enthalpy, entropy and
fugacity) of a blend, from a very limited set of experimental pure and mixture data. The
original Redlich-Kwong equation of state was proposed in 1949,
3.1
Where P is pressure, T is temperature, v is molar volume and, a and b are composition
and temperature dependent parameters. Soave (1980) modified this equation by
including an expanded ‘a’ parameter for pure components to include a temperature
dependent function with two constants m and n. This equation is known as the Redlich-
Kwong-Soave (1980) equation of state (RKS).
Thus, for each pure component i, 3.2
Where
( ) ( ))bvva
bvRTP
−−
−=
cii aa α=
43
3.3
3.5
3.6
The constants m and n have been determined from experimental data for each pure
component, using a least square curve fitting technique.
Calculated m and n values for a number of refrigerants are listed in
Table 3–1 m and n constants Refrigerant m n
R11 R12 R22 R23 R114 R142b R152a R500 R502
R13B1 R13
0.6417 0.5927 0.6737 1.15 0.6420 0.7502 0.9331 0.668 -0.1165 0.6209 0.5756
0.179 0.1937 0.1944 -0.11560.2577 0.1601 0.677 0.1792 0.6696 0.1684 0.1986
Furthermore, to determine mixture properties, Soave’s (1980) method includes an
experimentally determined interaction constant ijk in the mixing rule for a.
For mixture of components i and j,
3.7
3.8
ci
cici
ci
cici
ciii
cii
PRT
bb
PTR
a
TT
nmTT
08664.0
42748.0
11
22
==
=
⎟⎠
⎞⎜⎝
⎛+⎟⎟
⎠
⎞⎜⎜⎝
⎛−+=α
( )( )ijj
i ji
ijij
axxa
aiajka
∑∑=
−= 2/11
44
3.9
Where, ix and jx are the qualities of each component.
The interaction constants ijk are determined, for each binary mixture, from
experimental bubble point pressures using a least square curve fitting technique.
Calculated ijk values for a number of mixtures are listed in table 3-2
Table 3–2 Calculated ijk values Refrigerant mixture ijk Average error (%)R22/R12 R152a/R13B1 R12/R114 R13/R12 R22/R114 R22/R142b
0.03510.08330.0027
0.03 0.060 -0.018
0.5 0.26 1.1 1.1
3. 19 3.24
3.4.2 Vapor Density
Edmister (1968) has shown that the R-K-S equation of state can be rewritten as a cubic
polynomial of the compressibility factor Z. For certain values of temperature and
pressure this is found to have three real roots. The largest root represents the vapor
density; the smallest root is the liquid density, whilst the third has no physical
significance.
This polynomial takes the form;
Where 3.10
3.11
( ) 0223 =−−−+− ABZBBAZZ
RTbPB
TRaPA
RTPvZ === ,, 22
⎟⎟⎠
⎞⎜⎜⎝
⎛ +=∑∑ 2
jij
i ji
bbxxb
45
Although this equation is satisfactory for calculating the vapor density, it is generally
recognized that it does not result the desired accuracy for calculating liquid density.
3.4.3 Liquid Density
The following correlation is recommended by Downing (1974) for the calculation of the
liquid density,
3.12
Where fρ is the liquid density (lb/ft3), T and Tc are the temperature and critical
temperature (R) respectively.
The constants used in the liquid density equation, for a number of refrigerants, are listed
in Table 3–1 m and n constants
3.4.4 Enthalpy
The vapor and liquid enthalpies, H, at the state (P, T) are evaluated using,
3.13
In which the isothermal change in enthalpy is computed using,
3.14
Where
⎟⎟⎠
⎞⎜⎜⎝
⎛−+⎟⎟
⎠
⎞⎜⎜⎝
⎛−+
⎟⎟⎠
⎞⎜⎜⎝
⎛−+⎟⎟
⎠
⎞⎜⎜⎝
⎛−+⎟⎟
⎠
⎞⎜⎜⎝
⎛−+⎟⎟
⎠
⎞⎜⎜⎝
⎛−+=
cf
cf
cf
cf
cf
cfff
TTG
TTF
TTE
TTD
TTC
TTBA
11
1111
2/1
4/33/23/1
ρ
1ln'
−+⎟⎠⎞
⎜⎝⎛
+−
=∆ Z
BZZ
bRTTaa
RTH
DDT
T
TDp RT
RTHRT
RTHdTCH ⎟
⎠⎞
⎜⎝⎛ ∆−⎟
⎠⎞
⎜⎝⎛ ∆+= ∫
46
3.15
And
3.16
PC is the mixture ideal gas heat capacity at constant pressure, DT is the reference
datum temperature (chosen to be –40 F/C) and the subscript D denotes the datum
temperature. The reference enthalpy is at 0.0 BTU/h/lb liquid enthalpy at –40 F/C.
3.4.5 Entropy
The vapor or liquid entropies at the (P, T) are evaluated using,
3.17
in which,
3.18
For each pure component, at least three experimental vapor pressure data points must be
specified. These data points are input to a regression program which determines the
characteristic constants m and n for each pure substance. For each binary mixture, at
least two experimental pressures versus composition data point must be specified. These
data are input to regression program which determines the characteristic interaction
constant kij for each pair. The constants m and n for each component and the constant
kij for each pair are input to a set of subroutines for computing properties over the entire
range of composition.
( ) ( )''
2/1
'
21
ijjiji
cjciijji aakxxa αααα
αα+⎟
⎟⎠
⎞⎜⎜⎝
⎛−=∑∑
ci
iciii T
mTTn
−−= 2'α
⎭⎬⎫
⎟⎠⎞
⎜⎝⎛ ∆−
⎩⎨⎧
⎟⎠⎞
⎜⎝⎛ ∆+= ∫
DT
T
TP
RS
RSRdT
TCS
D
⎟⎠⎞
⎜⎝⎛ +
′+⎟⎠⎞
⎜⎝⎛ −
ZB
bRa
PBZ 1ln=
∆RS
47
Computation of Enthalpy and Entropy requires pure component ideal-gas heat capacity
data in the form of polynomial coefficients. Ideal gas heat capacity equation constants
for several refrigerants are given in table 3-4. Coefficients for the liquid density
equation must be provided by the user in the data file.
If the user inputs the condenser dew point temperature, it is necessary first to have a
preliminary estimate of the pressure prior to starting an iteration to determine the actual
pressure. The vapor pressure equation and constants used, for a number of refrigerants,
are listed in table 3-3.
3.5 OTHER EQUATIONS OF STATE
The thermodynamic properties can also be calculated using comprehensive equations of
state. Using this approach allows calculations at all conditions with thermodynamic
consistency. Other methods, such as the combination of a vapor-phase model with vapor
pressure and liquid density equations may not be applicable in the compressed liquid
and supercritical regions.
3.5.1 Pure Refrigerants
Depending on the availability of data, there are three other models that are widely used
for the thermodynamic properties of pure components. However, one of them
considered to be of high accuracy for refrigerants other than R123. This high-accuracy
pure-fluid equation of state is expressed in terms of reduced molar Helmholtz (1882)
free energy:
a idaRTA== ra+ = ( )kkki ldt
kk
t
ii γδδταταδ −++ ∑∑ expln 3.19
Where, the first two terms on the left side of the Equation 3.19, represent the ideal-gas
property effect. The second summation is the real-fluid, effect ra . Two dimensionless
variables expressing the temperature and density, TT *=τ and *ρρδ = , where the
parameters T* and *ρ have reducing effect and are often equal to the critical
parameters.
48
The parameters, iα and kα are coefficients obtained from the experimental data, and the
exponents kt , it , and kd are determined by processing a large number of data on a
software program. The parameter γ is equal to 0 when kl = 0; [and it is equal to 1
when 0≠kl . This model is used for the international standard calculation of R134a. Due
to its complete description of the thermodynamic properties, this model is sometimes
considered as the fundamental equation
3.5.2 Mixed Refrigerants
A new model is developed by Tillner-Roth ( 1997) to calculate the thermodynamic
properties of mixtures. This model is based on the Helmholtz (1882) energy that applies
mixing rules to the mixture components:
( )[ ] excesspqpqq
n
p
n
pqpjj
rj
idjjmix aFxxxxaaxa ∑ ∑∑
−
= +=
+++=1
1 1
ln 3.20
( )dVRTPRT
av
r ∫∞
−= ρ1 3.21
dTT
CR
dTCRThT
TR
SRTh
aT
T
idp
T
T
idp
refref
refrefid
refref
∫∫ −+⎟⎟⎠
⎞⎜⎜⎝
⎛+−−=
11ln1ρρ 3.22
Where the id subscript denominates the ideal gas and the superscript r denominates the
real fluid terms for each of the pure fluids in the n component mixture The first
summation in Equation. 3.20 represent the ideal solution. The jj xx ln terms are coming
from the entropy of ideal gases mixture where jx is the mole fraction of component j.
The double summation in the same equation refers to the “excess” deviation from the
ideal solution. The parameters and term(s) pqF and its multiplier excesspqa are generalizing
parameters, relate to the behavior of one binary pair with another, are empirical
functions related to the experimental binary blend data. The ra and excesspqa functions are
determined at a reduced temperature and density τ andδ . The Parameters refh and refS
are reference enthalpy and entropy at the defined conditions of refT and refρ .
49
mixTT *
=τ and *ρρδ mix= with
( )critq
critpq
n
p
n
qppqT TTxxkT += ∑∑
= = 21
1 1,
* 3.23
⎟⎟⎠
⎞⎜⎜⎝
⎛+= ∑∑
= =critq
critp
qp
n
p
n
qpqv xxk
ρρρ11
211
1 1,* 3.24
Where, the pqTk , parameter is determined by taking into account the bubble point
pressures that shows azeotropic behavior. The volume changes on a binary blend
represented by the pqvk , parameter. Ternary and higher order blend grouped into binary
pairs with pqTk , =1 and 1, =pqvk for qp = .
In the cases where limited vapor-liquid equilibrium data can be obtained, the excesspqa term
can be eliminated. This term can only be determined if extensive single-phase pressure-
volume-temperature and heat capacity data is obtainable. A function to determine a excesspqa has been developed by Lemmon using data for 28 binary pairs of various type of
gasses including HFC’s.
The thermodynamic criteria for the critical components for three component or higher
order blends, is a complex calculation exercise. Therefore, combination of the binary
critical lines used to estimate the critical parameters.
jki
kk
critjj
critii
crit xxcTxTxT ∑=
++=6
1
3.25
The values are constants that can be obtained from test results for each combination of
binary mixtures.
For three component and higher number of component blends, the following equation is
used.
50
( )
∑ ∑
∑ ∑−
= +=
−
= +== 1
1 1
1
1 1n
i
n
ijji
n
i
n
hij
critijji
crit
xx
ZTxxT 3.26
Where, the )(ZT critij is the binary critical temperature for the pair )(ij mixture, based on
the mixture composition x , and at a combining-unauthentic-composition Z .
Pure-component and binary-pair mixture boundary conditions or empirical methods
used to estimate the critical volume. Given the critical temperature and volume
calculated by the method outlined above, the critical pressure is determined using the
mixture equation of state.
3.6 CYCLE COMPONENTS
The major components used in the vapor compression refrigeration cycle include the
compressor, evaporator, condenser, expansion valve, suction line and discharge line.
The steady state conservation and performance equations are similar to those found in
many refrigeration and air conditioning texts.
3.6.1 Compressor
A typical reciprocating compression process is used in this modeling. Thermodynamic
cycle of a reciprocating compression process is basically similar to all positive
displacement compressors including screw and scroll with different volumetric
efficiencies. However, the method of suction and compression is different to each of
them, more details in Chapter 4.
In the calculation of the compressor power consumption the effects of clearance
volume, polytrophic compression, leakage losses, pressure drops across the valves and
superheating on the intake stroke, have been included.
The compressor power consumption is given by,
51
3.27
In which n is the poly tropic Index, the subscript b and c denote the cylinder suction and
discharge conditions respectively, and bb vP , , and m are evaluated using,
And the volumetric efficiency is obtained from,
3.33
For most cases, a clearance volume of 6-7% may be considered as reasonable.
A low overall volumetric efficiency may be caused by leakage losses passed the
discharge and suction valves, piston and stuffing boxes. Valve leakage may be due to un
satisfactory grinding of the valve or ring to the valve plate.
Hence, the actual volumetric efficiency may be written as follows,
3.34
Where vlc is assigned a value of 0.01, however, for old compressors this value may be
higher.
⎥⎥⎥
⎦
⎤
⎢⎢⎢
⎣
⎡−⎟⎟
⎠
⎞⎜⎜⎝
⎛⎟⎠⎞
⎜⎝⎛
−−=
−
11
1n
n
b
cbb p
pvp
nnmW &&
⎟⎟⎠
⎞⎜⎜⎝
⎛××=
=
∆+==
∆+=∆−==
4
2
4
4
5
4
BreStrNrpmPD
vPDm
TTTTT
ppppppp
v
cylab
a
exc
inab
π
η&
b
n
b
cv v
vpp
cc 4/1
1⎥⎥⎦
⎤
⎢⎢⎣
⎡⎟⎟⎠
⎞⎜⎜⎝
⎛−+=η
⎟⎟⎠
⎞⎜⎜⎝
⎛−=
b
ccvva p
pvlcηη
52
The heat transfer rate from the compressor to the surroundings can be determined using,
3.35
Figure 3-8 Compression process diagram
3.6.2 Evaporator
The evaporator capacity may be calculated using,
In which h3 is the enthalpy at evaporator outlet that includes pressure drop in the
evaporator as well as superheating, and h2 is the enthalpy at evaporator inlet.
3.6.3 Condenser
In a similar fashion the heat rejected in the condenser is evaluated using,
3.36( )23 hhmQe −= &&
( ) WhhmQComp&&& +−= 45
53
In which h6 is the enthalpy at condenser inlet (including discharge line losses) and h1 is
the condenser outlet enthalpy (including sub cooling and condenser pressure drop).
3.6.4 Expansion Valve
The expansion valve is considered to be adiabatic thus the exit enthalpy is equal to the
inlet enthalpy.
3.6.5 Suction and discharge Lines
Heat gain in the suction line accounted for and calculated from,
Whilst, the heat loss in the discharge line that connects the compressor outlet to the
condenser inlet is obtained using,
3.7 Determination of the Polytropic Exponent
In practice the compression process is neither isothermal nor adiabatic, but polytropic.
Figure 3-9 illustrates the typical Process in equation 3.40 for various values of the
exponent.
The value of the exponent n may vary from 0 to 1.4, however for reciprocating
compressor this lays within the limits of 1.0 and 1.3.
3.37
3.38
3.39
3.40
( )16 hhmQc −= &&
( )34 hhmQsl −= &&
( )65 hhmQdl −= &&
.constPvn =
54
P
V Figure 3-9 Various compression processes represented on a P – V diagram
The numerical value for the exponent n may be determined using,
In which P1, v1, P2 and v2 are the gas pressure and specific volume after intake valve
(before compressor) and before exhaust valve in the compressor cylinder, indicated as
the points b and c in Figure 3-8.
3.8 CONSTANTS
The vapor pressure equation which was used is given by,
In which P has the units of psia, T in Rankin and the constants are presented in Table 3–3.
12
21
loglogloglog
vvPPn
−−
= 3.41
( )TFT
TFEDTTCTBAP −⎟
⎠⎞
⎜⎝⎛ −
++++= logloglog 3.42
Exp. = 1 Isothermal
Exp. = n Polytropic
Exp. = γ Adiabatic
v
p
CC
n
=
⟨⟨
γ
γ1
55
Table 3–3 Constants for vapor pressure equations Refrigerant A B C D E F R11 R12 R22 R23* R114 R500 R502
42.147028 39.883817 29.357544 328.90853 27.071306 17.780935 10.644955
4344.3438 -3436.6322 -3845.1931 -7952.7691 -5113.7021 -3422.6971 -3671.1538
-12.8459 -2.4715 -7.86103 -144.514 -6.30867 -3.63691 -0.36983
0.0040083725 0.004730442 0.002190939 0.24211502 0.006913003 0.0005.02722 -0.001746352
0.0313605 0 -0.4457467 0.000212806 0.7814211 0.4629401 0.8161139
862.07 0 686.1 0.000000094349 768.35 0.0 695.57 654.0
* The form of the equation used is,
32loglog FTETDTTCTBAP +++++= 3.43
Ideal gas heat capacity ⎟⎠⎞
⎜⎝⎛
°− RlbBTU is evaluated using the polynomial,
In which the polynomial coefficients for several refrigerants are given in Table 3–4.
232
TfdTcTbTaCP ++++= 3.44
Table 3–4 Constants for ideal gas heat capacity equation Refrigerant a b c d f R11 0.02691
3.16266× 410− -.39982× 710−
6.77887× 1110−
-380.59
R12 9.2196× 310− 3.789× 410−
-2.7494× 710− 7.6582× 1110−
0
R22 0.033191 2.66138× 410−
-7.6813× 810−
0 303.66
R23 0.09306 -9.2254× 610−
4.76601× 710− -2.9949× 1010−
0
R114 0.019075 3.804× 410−
1.8203× 710−
0 0
R500 0.030552 3.2694× 410−
-1.10771× 710−
0 0
R502 0.0231755 4.19845× 410−
-1.59926× 710−2.5093× 1110− 0
56
3.9 RESULTS
Various compositions of R22 and R142b were used to validate calculation results versus
test outcomes. Comparison of the computed, using the above equations, and
experimental COPs, for the entire range of mass fraction, proved that the predicted
results followed the trends with the maximum deviation of about 5%. The pressure
ratio, mixture vapor-liquid pressures and GTD versus composition showed that the
minimum pressure ratio and maximum GTD occur at a mass fraction of 30%, which
coincides with the maximum COP.
It has been demonstrated that the computer model determines steady-state performance
characteristics with reasonable accuracy.
The model predicted compressor power consumption and COP for the mixed refrigerant
system experimentally investigated with a maximum deviation of 5%.
Further comparisons were performed using a software program calculating
thermodynamic properties of the pure and mixed refrigerants in the upcoming chapter 7.
Also, a simple refrigeration cycle is used to calculate the system performance. This
program uses the relevant Equations of State out lined in section 3.5 for pure and mixed
refrigerants. The test results showed that they are in agreement with the calculation
results with a good accuracy of somewhere between -1 to +4%.
57
4 CHILLER TECHNOLOGY
4.1 INTRODUCTION
Liquid Chillers are large capacity cooling machines that produce cold water, brine or
other secondary coolants for industrial and commercial air-conditioning or refrigeration
use. The system usually is factory assembled, shipped and installed as one machine.
Liquid chillers are either absorption or vapor compression type. Absorption liquid
chillers are not subject of discussion in this work.
A vapor compression liquid chiller consists of basic components vapor-compression
compressor with drive, liquid cooler (evaporator), condenser, flow control device,
electrical and control. Depending on the application and design, it may also include
expansion turbine, economizer and a receiver. Some auxiliary components such as oil
cooler, oil separator, oil pump and purge unit may be used.
4.2 SYSTEM FUNDAMENTALS
A liquid chiller system with a single compressor and one refrigeration circuit with a
water-cooled condenser are shown in Figure 4-1. Usually chilled water enters at 12 C
and leaves at 7 C. Nominal design point condenser water enters a cooling tower at about
35 C and leaves at 30 C. Air-cooled and evaporative cooled condensers are also widely
used.
58
Figure 4-1 A typical water-cooled chiller system schematic with shell & tube condenser and evaporator, and with screw or semi-hermetic reciprocating compressor
Multiple compressors with multiple circuit chillers, specifically dual systems, are
widely used in the industry. Multiple chillers are often used to add more flexibility and
redundancy.
Depending on the type of condensers used, three main categories of chillers are widely
used in the industry: water-cooled, evaporative cooled and air-cooled. Chillers also can
be either flooded or Direct Expansion (DX) type, depending on which method of
evaporation is utilized. Due to their lower condensing temperature, water-cooled and
evaporative cooled chillers are more efficient than the air-cooled, and are widely used in
high capacity applications. However, their higher cooling tower maintenance cost and
associated legion Ella risk has made unpopular for low capacity applications. Therefore,
the air-cooled chillers are more widely used in low capacity end of the market.
59
Figure 4-2 A typical Two stage centrifugal chiller (water-cooled) with economizer
Due to their high evaporating temperature and lower approach, the chillers with flooded
evaporators are more efficient than the DX types. But they are more expensive to
manufacture and specifically a lot of complexity and challenges involved with oil
management. For the aforementioned reasons, they are economically feasible with the
large capacities only. But this is about to change with commercialization of the new
revolutionary totally oil free centrifugal compressors.
The novel invention of the modular chillers invented in Australia and installed
worldwide, offers many advantages such as flexibility for manufacturer and end user,
compactness and redundancy as well as part load efficiency.
The chilled liquid is supplied to the medium cooling heat exchangers such as fan coils.
They are used in central comfort cooling as well as industrial process cooling systems.
Their capacity ranges from about 70 to up to 7000 kW cooling capacity, but mostly fit
into 200 to 1000 kW chiller range.
60
Table 4–1 Chiller Technology alternatives – Vapor Compression cycle Compressor Typical Capacity Range Drop in Refrigerant AlternativesReciprocating 70 to 500 kW HFC-407C
Screw 150 to 1500 kW HFC-407C
HFC-134a Scroll 70 to 300 kW HFC-407C
HFC-134a Centrifugal Over 500 kW HFC-134a
HCFC-123
4.3 COMPRESSORS
Two main energy sources are available for running chillers: electricity and natural gas --
this includes steam and waste heat.
Electric chillers include reciprocating, screw, scroll and centrifugal compressors. The
first three categorized as positive displacement compressor, whereas the last one uses
centrifugal force to push the gas.
a) Reciprocating
Reciprocating technology, which has been around since the beginning of
commercial refrigeration, is the most noted for its rugged design and ability to be
rebuilt in the field, but less so for its energy efficiency. Space requirements are
minimal, and noise and vibration levels are medium to high. Their capacity
roughly ranges from 70 to 500 kW and their overall efficiency is relatively
unspectacular. This type of compressor increases gas pressure by means of
displacing pistons.
61
Compression stroke
Discharge stroke
Suctionstroke
Figure 4-3 Compression process of a typical reciprocating compressor
Saturated Suction Temperature (SST)Saturated Suction Temperature (SST)
Capa
city
Capa
city
1010°°CC44°°CC--77°°CC --11°°CC
140 kW140 kW
70 kW70 kW
0 kW0 kW
6 cylinders6 cylinders
4 cylinders4 cylinders
2 cylinders2 cylinders
Figure 4-4 Capacity variation as a function of SST of a typical reciprocating compressor
62
A compressor capacity versus saturated suction temperature (SST) at a constant
saturated discharge temperature (SDT), Figure 4-4, reveals that the capacity of the
compressor increases as the suction temperature increases. In a given compression
cycle, a greater mass flow rate of refrigerant can be compressed at higher suction
pressure and therefore higher capacity of the compressor.
b) Screw
As one of the rotary family, screw technology recently crossed over from the
industrial refrigeration sector to the comfort cooling sector. Like centrifugals, they
don't require much space in a mechanical room, but their noise profiles are
significantly higher than centrifugals. Their capacity ranges and efficiencies are
similar to centrifugals; however they cannot utilize variable speed drives, so they
can not be as efficient. Refrigerants that can be used include R-22, R-134a, and
most recently R-407C. Single and dual rotor screws are the most common types
of screw compressors. They trap gas in hollow spaces of female rotor and
compress it by means of male rotor or two sun wheels type pinions squeezing it
towards the discharge port. This technology has bees around for decades;
however, until recently, machining of the rotor profiles had always been a major
challenge.
In the air-conditioning industry, helical-rotary compressors are most commonly
used in water chillers ranging from 150 to 1500 kW.
63
male rotormale rotor
female rotorfemale rotor
housinghousing
Figure 4-5 A typical dual rotor – male and female - screw compressor assembly
meshing pointmeshing point dischargedischargeportport
Figure 4-6 A typical dual rotor – male and female - screw compression concept
Figure 4-6 is a graphical illustration of a dual rotor screw compression process.
Continued rotation of the meshed rotor lobes drives the trapped refrigerant vapor
(to the right), toward the discharge end of the compressor, ahead of the meshing
point. This action progressively reduces the volume of the pockets, compressing
the refrigerant. Finally, when the pocket of refrigerant reach the discharge port,
64
the compressed vapor is released and the rotors force the remaining refrigerant
from the pockets.
c) Scroll
Scroll compressor technology is a major invention which increases efficiency and
reliability - and reduces noise levels. Its highly complex scroll profiles of these
two components optimize the hermetic technology. Scroll technology is rapidly
overtaking the niche of reciprocating chillers in comfort cooling. They provide
small size, low noise and vibration, and good efficiency. Available in air-cooled
and water cooled configurations, scroll chiller capacity can reach approximately
300 kW, which makes them good candidates for spot cooling or make-up cooling
applications. Scroll technology is pretty new compare to reciprocating and screw.
The recent advances in CNC and precise machining paved the way for cost
effective scroll manufacturing. The compression process in the Scroll compressor
is of course based on the Scroll technology which is described below.
intakeintakephasephase
compressioncompressionphasephase
dischargedischargephasephase
Figure 4-7 Suction, trap, compression and discharge process of the scroll concept
65
As shown in Figure 4-7, Scroll compression concept consists of two main
components, involutes scrolls that intermesh. The top scroll which contains the
gas discharge port is fixed and the bottom scroll orbits. This principle does not
require either suction or discharge valves and the absence of clearance volume
makes it a very high volumetric efficiency machine. The two scrolls are
maintained with a fixed angular phase relation (180º) by an anti-rotation device.
As the bottom scroll orbits within the fixed one crescent shape gas pockets are
formed, their volumes are reduced until they concentrate at the center section of
the scroll. All Suction, compression and discharge processes are performed
simultaneously.
journaljournalbearingbearing
motormotorshaftshaft
drivendrivenscrollscroll
directiondirectionof rotationof rotation
Figure 4-8 Rotation geometry of motor shaft, journal bearing and driven scroll
As the orbiting motion continues, Figure 4-8, the relative movement between the
orbiting scroll and the stationary scroll causes the pockets to move toward the
discharge port at the center of the assembly, gradually decreasing the refrigerant
volume and increasing the pressure. Two or three orbits are required to
66
accomplish the volume reduction or compression process allowing very smooth
operation.
Saturated Suction Temperature (SST)Saturated Suction Temperature (SST)
Capa
city
Capa
city
SDT3 ConstSDT3 Const
SDT2 ConstSDT2 Const
SDT1 ConstSDT1 Const
Figure 4-9 Capacity variation of positive displacement compressors versus SST at three constant SDTs
All the abovementioned three types of compressors, namely reciprocating, screw
and scroll fall into positive displacement category. Figure 4-9 illustrates capacity
variation of positive displacement compressors as function of SST and constant
SDTs, with SDT1 being higher than SDT2 and SDT3 respectively. Capacity
increases as SST increase, however, it decreases with discharge pressure rise.
d) Single speed centrifugals
Electric centrifugals have moderate space requirements and noise levels. They
range in capacity from 500 kW to 7000 kW (150 to 2000 Tons) for standard
production units. This technology has the most flexibility in refrigerant usage
including, R-134a, and R-123.
67
volutevolute
diffuserdiffuserpassagespassages
radial radial impeller impeller passagespassages
bladesblades impellerimpeller
Figure 4-10 A typical single stage centrifugal compressor cross section
The inlet of impeller is fitted with blades that draw refrigerant vapor into radial
passages that are internal to the impeller body. The rotation of the impeller causes
the refrigerant vapor to accelerate within these passages, increasing its velocity
and kinetic energy, Figure 4-10.
68
refrigerant refrigerant enters enters diffuserdiffuser
Path through compression processPath through compression process
Veloc
ity or
Kine
tic
Veloc
ity or
Kine
tic
Pres
sure
Pres
sure
Stati
c pre
ssur
eSt
atic p
ress
ure refrigerantrefrigerant
enters impellerenters impellerrefrigerant refrigerant enters voluteenters volute
Figure 4-11 Velocity increase by impeller and conversion of velocity into static pressure in a centrifugal compression process
The accelerated refrigerant vapor leaves the impeller and enters the diffuser
passages. As size of the diffuser passage increases, the velocity and therefore the
kinetic energy of the refrigerant decreases. Thus, the refrigerant’s kinetic energy
(velocity) is converted to static energy, or static pressure. The volute also becomes
larger as the refrigerant travels through it. As the size of the volute increases, the
kinetic energy is converted to static pressure, Figure 4-11.
69
2525
7575
1010 1414
6363
pres
sure
diffe
renc
epr
essu
re di
ffere
nce
capacitycapacity
surgesurge 9090
36365151
A
unloading lineunloading line
vane positionvane position(degrees)(degrees)
CB
Figure 4-12 Performance of a typical centrifugal compressor over a range of inlet vane positions
Performance of a typical centrifugal compressor over a range of inlet guide vane
positions is shown in Figure 4-12. The pressure difference between the
compressor suction and discharge is on the vertical axis and compressor capacity
is on the horizontal axis. Most of the cases pressure ratio is used instead of
pressure differential; however, it does not make difference on the graph trends. As
the load on the compressor decreases from the full-load operating point A, the
inlet guide vanes partially close, reducing the flow rate of refrigerant vapor and
balancing the compressor capacity with the new load B. At the new part load,
required heat rejection reduced, hence reducing approach and consequently
condensing pressure. This results in reduced the pressure difference between the
evaporator and the condenser. Continuing along the unloading line, the
compressor remains within its stable operating range until it reaches C.
Surge line (dashed) is where compressor can not operate beyond it due to surge
phenomenon caused by less than minimum required gas velocity to overcome
discharge pressure. If compressor forced to operate beyond surge line, it would
become very unstable and potentially cause damage. Inlet vanes on a centrifugal
70
compressor allow it to unload over a broad capacity range while preventing the
compressor from operating in the surge region.
e) The new totally oil-free centrifugal
It is fair to say that, for the last few decades, the compressor technology has been
very slow in adapting innovations and has been stuck with single speed motors
and oil dependant lubricating system. This is about to change by invention of the
revolutionary world’s first totally oil free, variable speed and high efficiency
centrifugal compressor, which was developed in Australia and is commercialized
in North America, Europe and Australia, Figure 4-13.
Pressure and temperature
sensors
Inverter speed control
Synchronous brushless DC motor
Motor and bearing control
Inlet Guide Vanes
2 stage centrifugal compressor
Figure 4-13 A cut-away section view of the new totally oil-free compressor
Dual stage centrifugal compression is used to maximize full load efficiency, as
centrifugals provide the industry’s highest efficiency. In addition, when coupled with a
variable speed drive, they provide the highest part load efficiency. It has only one
moving part, a shaft fitted with two stage impellers, Figures 4-13 and 4-14.
71
Figure 4-14 Cross section view of two stage centrifugal gas compression assembly
The new totally oil free compressor uses the latest magnetic bearing technology to
levitate the shaft, only moving part, by a digitally controlled magnetic bearing system
consisting of two radial and one axial bearing, Figure 4-15 and 4-16. Position sensors
provide real-time repositioning of the rotor over one hindered thousand times a second.
The levitated shaft rotates floated on air without friction at high speed up to 48,000
RPM.
72
Figure 4-15 Shaft and impellers assembled with magnetic bearings
Figure 4-16 Cross section schematic of the front radial, rear radial and axial bearings with sensor rings
It also utilizes advantages of variable speed drive as integral part of the compressor to
modulate the capacity by adjusting speed to the load, hence achieving unprecedented
efficiencies that never seen before.
Being oil-free offers enormous advantages in sustainable performance over the entire
lifetime of the application. It simplifies design of the systems and cuts costs and
complexity associated with oil management as well as the failure factor due to lack oil
in the system. By eliminating oil from the system, improves heat transfer coefficient
significantly, as oil coats tubes surfaces and degrades its conductivity.
73
Figure 4-17 Operating map of the new totally oil free and variable speed centrifugal compressor
Physics of centrifugal compression theory is fundamentally different from positive
displacement. Positive displacement method is simply to compress gas by means of
reducing volume hence increasing gas pressure, while centrifugal method converts
kinetic energy that created by centrifugal force into high pressure. It therefore is
sometimes called dynamic compression that can vary depending on flow rate, and low
and high pressure changes. Each compressor map plays an important role on operating
range and limits of that compressor. We briefly discussed about single speed centrifugal
map earlier in this chapter.
However, as it can be seen on Figure 4-17, the new compressor has multiple numbers of
map curves, each for a specific constant speed. In fact, due to infinite speed steps,
numbers of curves exist in reality, but a few specific speeds selected for better
understanding of the map. As it can be seen on the map, Figure 4-17, each curve
represent and specific constant speed that follow pattern of upward to a maximum or
74
turning point and continues downward. Turning point is where inlet guide vanes (IGV)
starts to close by modulating vanes. From that point to the right all the way to the end
tail of the curve is where IGV is fully open and is efficient operating range until the tail
end to the right where compressor reaches its maximum flow rate limit. This is also
called chock limit which can be determined using Mach number of gas at inlet
conditions. However, from that point on to the left is where operation enters surge area
and IGV is partly closed depending on gas flow rate and pressure ratio. The end tail to
the left of the curve is the minimum load limit and real surge point.
The compressor full map is stored in the compressor controller memory which uses it to
know the position at each operating condition and therefore try to optimize as well as
reposition it from the limiting boundaries. Due to this and other reasons, this
compressor is also called the first intelligent compressor ever in the industry.
75
5 EXPERIMENTAL APPARATUS
5.1 THE TEST APPARATUS
The test apparatus used in this work contains the basic components found in any
refrigeration system. The system is basically a single stage vapor compression
refrigeration cycle with the nominal capacity of about 3.0 kW. The other major parts are
air cooled condenser with air flow control, specially designed and built counter flow
tube in tube evaporator and thermostatic expansion valve. Special attention were paid to
the evaporator design to resemble small scale chiller as well as a near ideal counter flow
heat exchanger. The evaporator was also fitted with number of thermocouples and
pressure sensors, so the boiling and superheating process could be better viewed and
analyzed. All the sensors then connected to corresponding special data loggers. Various
components of the plant are briefly discussed below.
5.1.1 Compressor
This is a hermetic type reciprocating compressor suitable for R22 and R407C with
pressure and temperature transducers are fitted at the discharge and suction ports. The
compressor is a small swept volume of positive displacement type which is widely used
in the industry. Performance and behavior of this compressor would resemble
characteristics of the large capacity size compressors currently used in the chiller
industry. Overwhelming majority of the compressors that are used in the chillers
manufacturing is of positive displacement type. When a volume of gas sucked into a
chamber and physically compressed to higher pressure by means of reducing volume is
called positive displacement compression. Unlike centrifugal compression, positive
displacement compression requires high torque and consequently high starting current
draw. In some cases, especially residential air conditioning system, soft starters are
fitted to reduce the starting current. The starting current sometimes also called locked
rotor amps. Nevertheless, this was not an affecting factor in the intended tests, as the
performance data were collected at stable steady operating conditions.
5.1.2 Condenser
The condenser used in this rig is of air cooled coil type to reject heat by transferring
heat from inside tubes where condensation phase change is taking place to blowing
76
through air on the coil fins. Copper tubes expanded to aluminum fins, air cooled coil
with below through axial fan. Majority of heat exchange takes place from the
aluminum fins to air. To enhance the performance, internally grooved tubes were fitted
which improves heat transfer by creating turbulence inside tubes. Air inlet adjustments
and heaters were used to maintain the desired conditions. The condensation phase
change process has always been predicted with reasonable accuracy using heat transfer
correlations. The heat transfer research works have proved that internal grooved tube
significantly improves condensation efficiency. For this reason, almost all of the
compact and high performance condensers equipped with internally grooved tubes.
Various fin shapes are proven to enhance heat transfer on the air side by introducing
turbulence on the air flow stream.
5.1.3 Evaporator In the design of this test plan, special attention has been given to the evaporator. The
evaporator is a tube in tube, dual pass and counter flow U shaped heat exchanger. The
intension was to allow the refrigerant effect to occur within the inner tube alone, be able
to observe and measure evaporation and superheating behavior all along the evaporation
process. The inner tube was of special type internally rifled and externally finned with
nominal 3/4 inch diameter. The internally rifled tubes are specifically designed and
manufactured to improve boiling phase change. Unlike condensation phase change,
evaporation process is extremely challenging to predict with reasonable accuracy.
Extensive research work has been carried out on this topic to model various types of
evaporators. Each and any type has its unique behavior represented in complex
correlations with defined boundary conditions. Of all those, boiling inside a tube with
heat transfer fluid at outer diameter have attracted good research work with reasonable
modeling correlations. One example of those works describes the relationship between
the heat transfer coefficient of pure substance and different flow regimes as a function
of the vapor quality in Figure 5-1. Other studies of R407C boiling inside a tube showed
similar supporting evidence that the maximum heat transfer coefficient obtained at 65-
80% vapor quality and then drops dramatically characterizing the “dry out” region. The
only noticeable difference is that the annular flow is achieved earlier in pure substance
than the mixtures. This can be explained due to the effect of composition on nucleation
boiling and a change in physical properties of mixture with the composition shifts from
boiling
77
Figure 5-1 Heat transfer coefficient as function of boiling regimes and quality inside a tube, Cengel 1998
The outer jacket tube diameter was 1 5/8” inch with total length of 2.4 m at two parallel
lines. The refrigerant flows inside the inner diameter and the chilled liquid flows
counter currently in the annular space which is driven by a pump, Figure 5-1. The outer
jacket was properly insulated to stop heat transfer between the evaporator and ambient.
A liquid thermal reservoir tank fitted with adjustable electric heaters was used to
maintain desired entering chilled liquid temperature. Baffles also installed to properly
mixing the water and providing with more stable bulk temperature. Small pump is used
to circulate the liquid. The chilled liquid was mixture of water and glycol, which
referred to in the industry as brine. Calibrated chilled liquid side flow meters were also
used.
78
CONDENSER
COMPRESSOR
TX VALVE
PUMP
HEATER
Glycol Solution Reservoir
EVAPORATOR
Refrigerant Line
Figure 5-2 Schematic of the experimental apparatus Figure 5-3 Evaporator piping and sensor arrangement isometric schematic
The evaporator was fitted with series of temperature sensors as shown on Figures 5-2
and 5-3. Total of 22 temperature sensors used, 16 for refrigerant side which the tip of
the sensors were installed at the centre of the inner tube and the rest used for measuring
chilled liquid, suction, discharge and liquid temperature measurements, Figures 5-4 and
5-5. These sensors provided very useful detailed behavior information, specifically
along the evaporator. They make it possible to plot and view refrigerant temperature
change trend during boiling and superheating along the evaporator. Pure refrigerants
theoretically boil at constant temperature at constant pressure and rise during
superheating process. It is very important to view this in practice where pressure drop is
Chilled Glycol OUT
Refrigerant Liquid From Condenser
To Compressor Suction
TX Valve
TemperatureSensors
Chilled GlycolIN
79
more than zero and rate of heat transfer is not necessarily uniform. For the mixed
refrigerants, the boiling process is not isothermal. Instead, the evaporating temperature
glides during boiling along the evaporator. This phenomenon is called Gliding
Temperature Difference (GTD). The temperature sensor set up would provide
opportunity to log and view this process therefore more detailed insight of the
evaporation process of pure and mixed refrigerants as well as the leaked system.
Figure 5-4 Temperature sensor (Thermocouple) location schematic
Figure 5-5 Evaporator Thermocouple tip installation schematic
6 5 7
4
8
3
9
2 1
10 1117
16
1312 14
pump Glycol Solution
Reservoir
condenser unit
19
20 22
18
15
compressor
80
For convenience and accuracy sake, the pressure sensors were connected to flare end of
the capillary tubes extended from the position where the pressures were intended to be
measured. Due to lack of flow and velocity inside the capillary tubes, the measured
pressures represent exact pressures on the evaporator.
The thermocouples had to be built specifically for this purpose by the University
technical staff. Special attention was paid to tip, the sensing point, of the thermocouples
to make sure that bulk temperature is measured. This part was a challenging part of
building the evaporator, see Figure 5-4, as they had to be extended past through the
outer jacket all the way to the middle of evaporator cross section. Measuring the
temperatures on the tube wall would not be considered accurate enough for research
work due to temperature gradients from middle of the tube and outside temperature
effect.
Figure 5-6 General view of the test rig
81
Figure 5-7 Data acquisition and instrumentation set up of the experimental apparatus
5.2 INSTRUMENTATION AND DATA LOGGING
5.2.1 Thermocouples
The evaporator was fitted with sixteen K type thermocouples, all along the evaporator,
to measure refrigerant temperatures. These were penetrated all the way from outer
jacket through to the middle of the inner tube. Another 4 thermocouples were fitted to
measure chilled water side temperatures, and additional sensors were fitted for
82
measuring discharge, liquid and the suction gas temperatures. These thermocouples had
to be fitted to specially built fittings to seal from outside as well as the water ingress.
Unlike pressure sensors, temperature sensors have to be fitted precisely at exact position
for accurate and reliable measurement. The temperature sensors in this test rig were
positioned on the very spot of the intended measurement. Tip of the thermocouples were
inserted from top side of the tube into precisely middle of the evaporator tube cross
section.
5.2.2 Pressure Transducers
One low and one high range pressure transducers were fitted on the suction and
discharge side of the compressor, respectively. Additional two were used to alternate
between a few points. All were initially calibrated using a certified reference pressure
gauge. Measuring three pressures on the evaporator side made it possible to calculate
pressure drop and relate it to temperature change behavior of the evaporator. For
accurate and apple to apple comparison of performances at various conditions, average
evaporating temperature corresponding to average pressure was used.
5.2.3 Data logging and calibration of the sensors
All the thermocouples and pressure transducers were connected to a data logger. Due to
different signal generated by each category sensors, special data logger was used for
each purpose, one for thermocouples and another for the pressure transducers. All
signals and data were relayed back from there to a computer. All the data then were
logged into the computer transferable to excel file which then used as raw data and then
processed.
Before starting any test and data logging, all the pressure transducers were calibrated
using certified accurate high resolution pressure gauge against the values read from the
computer. Discharge pressure transducer was calibrated to high pressure range, while
the other three evaporator side transducers were calibrated to lower pressure range.
Calibration graphs were prepared for each one and used in correcting the measured data.
The thermocouples were also calibrated using certified accurate and high resolution
thermocouple as reference.
83
Current were measured using current transducers by measuring inductance around the
cable to accurately determine the current passing through. The voltages were also
measured simultaneously that correspond to the measured current to take into account
minor voltage fluctuations. To excel accuracy, currents and voltages were measured
over a period of time and average taken to calculate power consumption.
5.3 TEST PROCEDURE
After the test rig set up was complete, the rig was thoroughly checked and
commissioned. All the major components, piping, electrical, control and data logging
were precisely inspected and tested for system function. The system then subjected to
series of tests at various conditions, fine tuned charge, evaporator water circulation and
flow meter tuned and expansion valve adjusted. Sample data were taken by the
computer and format was set for better reading and test identification. Once all these
initial set up, inspection, adjustment, tuning, etc completed, the research work tests
started as per the following plan.
At the first stage, series of tests have been carried out using R22 refrigerant under
various conditions. At the second stage the system was charged with R407C refrigerant.
Series of comprehensive tests have been carried out at the identical various conditions.
Details of the various test conditions are outlined in Table 5-1. Before logging and
recording data for each test at the specific condition, utmost attention and effort were
made to stabilize the system at the intended conditions and let it run at this condition for
a period of about 20 minutes. Data would only be recorded if the system proved stable
for this period of time.
First, the system was charged with pure R22, tested at the intended various indicated
conditions, and logged and recorded the data. Tests were carried out the same way on
pure R407C refrigerant by following the same procedures. The tests with pure R22 were
carried out so they can be used for comparison purposes. The reason being, R407C is
considered as the most suitable drop in alternative for R22. This stage of tests were
carried out without a major hick up, except had to make sure the conditions that the
system kept were identical. For each condensing temperature, leaving chilled water and
sub cooling temperatures as well as the chilled water flow rate had to be kept at certain
value. This was very important to resemble real world chiller behavior.
84
The second phase of the experiments was the most important, time consuming and the
most challenging part of this work. The system was tested using several combination of
leaked gas to simulate, as close as possible, what would happen to chiller equipment if
subjected to gas leak. The phenomenon were analyzed from the wide variety of
prospective such as effect on capacity, suction and discharge pressures and
temperatures, power input, evaporation and condensation, etc. This time the system was
tested with a refrigerant blend which its composition was changed due to leak and was
not any more a pure defined refrigerant. A fixed percentage of the gas was let to leak
out, in vapor form, and measured by collecting it in a bottle. The most volatile substance
of the mixture tends to leak during an isothermal vapor leak. The remaining refrigerant
becomes more concentrated in less volatile and low pressure substances. The situation
gets worse as more percentages of the refrigerant are leaked. This would have
consequences in two fronts. Capacity and efficiency drop due to shortage of gas in the
system, and refrigerant characteristics change due to high concentration of the less
volatile substances in the composition. In academic world, these effects could be
interesting from system behavior stand point, but it means much more in the industry.
Besides what interests the academic work, performance and efficiency loss, it poses
system failure risk due to chilled water freeze up, lack of oil return, high discharge gas
temperature, etc.
The next step of the experiments conducted in a way that simulated what the industry
has been doing for many years by simply topping it up the system with pure R407C.
This is a desired way by the industry; however, this can be applied when using pure
refrigerants, not zeotropic mixture like R407C. The system subjected to isothermal
vapor leak, at various proportions, and then compensated by adding up R407C
refrigerant in liquid form. Special attention was paid to the leak process to ensure that
the process was carried out as per the definition and identical at other conditions. It has
been established earlier that the vast majority of leaks occur in the real world systems
are of isothermal type and very rarely adiabatic. So it is very important to apply
precisely the isothermal leak which means slow gas leak. The leaked gas then directed
to specially prepared gas cylinder on a scale to weigh the amount using certified
weights. Specially fitted valves were used to simulate slow leak process. The idea was
to observe and record system behavior, at each proportion, from boiling characteristics
to pressures, temperatures, capacity variations and most important of all efficiencies.
85
From there, to find out the acceptable leak proportion and top up level. This is a very
important side of this research, as the solution as a result of this work has a potential to
save enormous amount of time and money as well as inconveniences for the industry,
specially the central equipment owner and maintenance staff.
Table 5–1 Details of the conditions of series of the tests conducted upon.
In order to have an apple-to-apple comparison, it is utmost important to set correct test
parameters at various conditions, especially evaporator. A small variation on evaporator
conditions would result in considerable effect on the performance of the system, while,
condenser side has less effect. Before each test stage, the system was properly
evacuated, charged with the required refrigerant, measurement devices and data logger
connected to be outlined later.
Condensing Temperature
40 C
Condensing Temperature
50 C Chilled Water Leaving Temperature (CHWL) C
7 10 12 7 10 12
Pure R22 * * * * * *
Pure R407C * * * * * *
R407C Leaked 10% Vapor and topped up * * * * * *
R407C Leaked 20% vapor and topped up * * * * * *
R407C Leaked 30% Vapor and topped up * * * * * *
R407C Leaked 40% Vapor and topped up * * * * * *
R407C Leaked 50% Vapor and topped up * * * * * *
86
6 TEST RESULTS
6.1 ANALYSIS OF THE COLLECTED DATA
The data that collected through the data acquisition system included temperatures,
pressures, flow rates, voltages and current. Upon stabilization of the system, series of
data acquired every 10 seconds for the duration of ranging from two to five minutes.
Samples of data series has been plotted to show stability of the system during the data
collection for each test. Average of each series of data has been used for calculations.
Data Aquisition,40 C CT, 10 C CHWLT
400
420
440
460
480
500
1 2 3 4 5 6 7 8 9 10Logged every 10 seconds intervals
Pres
sure
(kPa
)
Suction Pressure (SP)
Average Evaporating Pressure(Avg EP)
Figure 6-1 Evaporator pressure data acquisition for pure R22
Figure 6-1 shows very stable and constant pressures after the 3rd interval with
maximum deviation of only about 2 kPa. This represents less than 0.5% fluctuations
during the data logging period, which is better than expected for a thermal system. Due
to the multiple affecting factors involved, it is a challenge to keep the instability at
minimal level. However, even this negligible variation is absorbed by taking average
and using them in the data processing.
87
Data aquisition, 40 C CT, 10 CHWLT
0
2
4
6
8
10
12
1 2 3 4 5 6 7 8 9 10
Logged every 10 seconds intervals
Tem
pera
ture
C Temp 12Temp 13Temp 14Temp 15Temp 16
Figure 6-2 Evaporator temperature data acquisition vs. time for pure R22
Data aquisition, 50 C CT, 10 C CHWLT
0123456789
10
1 3 5 7 9 11 13 15 17 19
Logged every 10 seconds intervals
Tem
pera
ture
C
Temp1Temp1Temp2Temp3Temp4Temp5Temp6Temp7Temp8Temp9Temp10Temp11
Figure 6-3 Evaporator temperature data acquisition for pure R407C
88
Data Acquisition, 50 C CT, 7 C CHWLT
0
1
2
3
4
5
6
7
1 2 3 4 5 6 7 8 9 10 11
Logged Every 10 seconds intervals
Tem
pera
ture
CTemp 1Temp 2Temp 3Temp 4Temp 5Temp 6Temp 7
Figure 6-4 Evaporator temperature data acquisition for 10% leaked R407C
Data aquisition, 40 C CT, 12 C CHWLT
0
2
4
6
8
10
12
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Logged data every 10 seconds
Tem
pera
ture
C
Temp 1Temp 2Temp 3Temp 4Temp 5Temp 6Temp 7
Figure 6-5 Evaporator temperature data acquisition for 30% leaked R407C
Figure 6-5 shows the maximum deviation of only about 0.2 C on all of the graphs
With exception of two other graphs represent very stable and constant temperatures
after the 2nd interval. The other two graphs show about 0.4 C during the last four
readings. This variation is further minimized by using average value of those data.
Therefore, the accuracy was not compromised even with the interactive thermal system.
89
Data aquisition, 50 C CT, 7 C CHWLT
400
420
440
460
480
500
1 2 3 4 5 6 7 8 9 10 11 12
Logged every 10 seconds intervals
Pres
sure
kPa
Avg Ep
SP
Figure 6-6 Pressure data acquisition for pure R407C
Data aquisitio, 50 C CT, 7 C CHWLT
1800
1850
1900
1950
2000
1 2 3 4 5 6 7 8 9 10 11 12
Logged every 10 seconds intervals
Pres
sure
kPa
CP
Figure 6-7 Condensing pressure data acquisition for pure R407C
90
Data aquisition, 50 C CT, 7 C CHWLT
400
600
800
1000
1200
1400
1600
1800
2000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Logged every 10 seconds
Pres
sure
kPa SP
Avg EPCP
Figure 6-8 Condensing, average evaporating and suction pressure data acquisition for 10% leaked R407C
Data aquisition, 50 C CT, 7 C CHWLT
400
410
420
430
440
450
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Logged every 10 seconds intervals
Pres
sure
kPa
SP
Avg EP
Figure 6-9 Average evaporating and suction pressure data acquisition for 10% leaked R407C
91
Data aquisition,40 C CT, 12 C CHWLT
400
420
440
460
480
500
1 2 3 4 5 6 7 8 9 10 11 12
Logged every 10 seconds
Pres
sure
kPa
Avg EPSP
Figure 6-10 Average evaporating and suction pressure data acquisition for 30% leaked R407C
Data aquisition, 40 C CT, 12 C CHWLT
1400
1420
1440
1460
1480
1500
1 2 3 4 5 6 7 8 9 10 11 12 13
Logged every 10 seconds
Pres
sure
kPa
CP
Figure 6-11 Condensing pressure data acquisition for 30% leaked R407C
92
Data aquisition, 50 C CT, 7 C CHWLT
400
410
420
430
440
450
1 2 3 4 5 6 7 8 9 10 11 12
Logged every 10 seconds
Pres
sure
kPa
SPAvg EP
Figure 6-12 Average evaporating and suction pressure data acquisition for 40% leaked R407C
Data aquisition, 50 C CT, 7 C CHWLT
1600
1700
1800
1900
2000
1 2 3 4 5 6 7 8 9 10 11 12
Logged every 10 seconds
Pres
sure
kPa
CP
Figure 6-13 Condensing pressure data acquisition for pure 30% leaked R407C
93
Refrigerant tempreature along the evaporator, at 40 C CT, R22
2
4
6
8
0 2 4 6 8 10Sensor position along the evaporator
Tem
pera
ture
(C)
CHWLT 7 CCHWLT 10 CCHWLT 12 C
Figure 6-14 Refrigerant temperature along the evaporator at CT 40 C for pure R22
Refrigerant tempreature along the evaporator at 50 C CT, R22
2
4
6
8
0 2 4 6 8 10
Sensor position along the evaporator
Tem
pera
ture
(C)
CHWLT 7 CCHWLT 10 CCHWLT 12 C
Figure 6-15 Refrigerant temperature along the evaporator for pure R22
94
SP & avg EPs , R22
430440450460470480490500510
7.0 8.0 9.0 10.0 11.0 12.0CHWLT C
Pres
sure
(kPa
)
SP @ 40 C CT
Avg EP @ 40 C CT
SP @ 50 C CT
Avg Ep @ 50 C CT
Figure 6-16 Suction and average evaporating pressures vs. CHWLT at CT 40 and 50 C for R22
95
Figure 6-17 Suction and average evaporating pressure vs. CHWLT at 40 C and 50 C CT for 10% leaked R407C
Figure 6-17 shows suction and average evaporating pressures trend and pressure drops
for R407C after 10% leaked and then topped up. At 40 C CT, the pressure drop across
the evaporator appears to be slightly high towards lower CHWLT of 7 C, about 25 kPa.
The lower evaporating as a consequence of the lower CHWLT resulted in lower
refrigerant velocity and partly oil clogging inside the evaporator tube. Although an
average 10 kPa pressure drop is not unusual in direct expansion evaporators, the 25 kPa
pressure drop does not pose a serious concern. However, this should be treated as
warning sign which could be exaggerated with higher leak ratios.
SP & Avg EP, R407C 10% leaked
350
370
390
410
430
450
470
7 8 9 10 11 12
CHWLT(C)
kPaSP @ 40 C CT
Avg EP @40 C CT SP @ 50 C CT Avg EP @ 50 C CT
96
Figure 6-18 Suction and average evaporating pressure vs. CHWLT at 40 and 50 C CT for 30% leaked and topped up
As the leak ratio increases, signs of trend shape change for the pressure drop between
average evaporation and suction pressure becoming noticeable, Figure 6-19. The
pressure drop across the evaporator appears consistent with average of about 15 kPa for
both 40 C and 50 C CT. The pressure drop does not vary significantly with increasing
CHWLT. Increasing CHWLT means increasing ET and consequently increasing
capacity. At the 30% leaked R407C, the refrigerant start to show noticeable behavior
change that would have adverse effect on the system operation and performance.
SP & Avg EP, R407C 30% leaked
380
400
420
440
460
480
7 8 9 10 11 12
CHWLT (C))
kPa SP @ 40 C CT
Avg EP @ 40 C CT
SP @ 50 C CT
Avg EP @ 50 C CT
97
SP and Avg Eps, R407C 50% leaked
350
370
390
410
430
450
470
490
7 8 9 10 11 12
CHWLT (C)
kPa
SP @ 40 C CT
Avg EP @ 40 C CT
SP @ 50 C CT
Avg EP @ 50 C CT
Figure 6-19 Suction and average evaporating pressure vs. CHWLT at CT 40 and 50 C for 50% leaked and topped up
Avg EP at 40 C CT
R22 R407C
10% leak
20% leak30% leak 40% leak
50% leak
R22 R407C
10% leak
20% leak30% leak40% leak50% leak
R22R407C
10% leak
20% leak30% leak
40% leak50% leak
360
380
400
420
440
460
480
500
Pres
sure
(kPa
)
CHWLT 7 CCHWLT 10 CCHWLT 12 C
Figure 6-20 Average evaporating pressure variation vs. pure R22 and percentage leaked R407C at various CHWLT for 40 C CT
The average evaporating pressures were plotted for range of refrigerant charges from
pure R22 to R407C and all the leaked compositions. The pressure trend moves upward
from R22 to R407C, and then points downward for the leaked compositions depending
on the ratio of the leakage. The trend is almost consistent from pure R407C down to
98
20% leaked composition, but start to show irrational up and down behavior at higher
leaked. This is an indication of the fact that the refrigerant loosing its oil miscibility and
sitting inside evaporator tube, thus causing more pressure drop. This phenomenon can
pose a very serious reliability problem in the existing refrigeration and air-conditioning
systems
Avg EP at 50 C CT
R22R407C
10% leak
20% leak30% leak 40% leak
50% leak
R22
R407C
10% leak
20% leak30% leak 40% leak
50% leak
R22
R407C
10% leak
20% leak 30% leak40% leak 50% leak
400
420
440
460
480
500
520
Pres
sure
(kPa
)
CHWLT 7 CCHWLT 10 CCHWLT 12 C
Figure 6-21 Average evaporating pressure variation vs. pure R22 and percentage leaked R407C at various CHWLT for 50 C CT
A similar trend is followed with the average evaporating pressure at 40 C CT. All
indications point on behavior change from 20% leak onward. One of the reasons could
be the refrigerant oil miscibility characteristic change which in turn can cause partial oil
blockage in the evaporator tube. When a system start to show this type of operating
indications, then reliability of operation is under big question mark, and something need
to be done. In a real chiller system, these could lead to freeze protection pressure or
temperature cut out, and the system practically inoperable. The test rig is used to carry
out tests is controlled manually, and none of the pressure and temperature safeties
installed in order to make it possible to carry out tests in extreme conditions.
Each pressure line corresponds to specific LCHWT, i.e. 7, 10 and 12 C. Also, the higher
CT or condensing temperature indicates the higher average evaporating temperature.
This is due to the fact that the higher condensing pressure has its effect even after the
expansion device such as thermal expansion valves. The expansion devices cause
specific pressure drop at constant enthalpy.
99
Capacity variations at 40 C CT
R22
R407C 10% leak20% leak 30% leak
40% leak50% leak
R22
R407C 10% leak20% leak 30% leak
40% leak50% leak
R22
R407C 10% leak20% leak
30% leak40% leak 50% leak
2000210022002300240025002600270028002900
Cap
acity
(W)
CHWLT 7 CCHWLT 10 CCHWLT 12 C
Figure 6-22 cooling capacity variation vs. pure R22 and percentage leaked R407C at various CHWLT for 40 C CT
Capacity variations at 50 C CT
R22
R407C 10% leak20% leak 30% leak
40% leak 50% leak
R22
R407C10% leak 20% leak
30% leak40% leak 50% leak
R22
R407C10% leak
20% leak 30% leak
40% leak50% leak
17001800190020002100220023002400250026002700
Capa
city
(W)
CHWLT 7 CCHWLT 10 CCHWLT 12 C
Figure 6-23 cooling capacity variation vs. pure R22 and percentage leaked R407C at various CHWLT for 50 C CT
Figure 6-22 and Figure 6-23 show declining trend of the capacity from pure R22
towards R407C and continue to decrease as the leak percentage increases. The trend is
similar for both 40 C and 50 C SCT. With R407C when compared with R22, the system
capacity drops between 4 to 6%, depending on the pressure ratio. The capacity is higher
for 40 C than the 50 C CT, and higher CHWLT result in higher capacity. These are due
to two factors, nature of positive displacement compressors and phase change and
enthalpy change behavior of the refrigerant; the higher CHWLT corresponds to higher
100
evaporating temperature and pressure, hence higher density, hence higher mass flow.
This is vice versa with the SCT or condensing pressure due to mainly the gas phase
change behavior on the P-h diagram.
The average capacity loss at 30% leak when compared to pure R407C is about 15%,
however, this loss goes up to about 25% with the 50% leak. In the commercial chiller
systems, a capacity loss of up to 10% may be tolerated, but losses beyond that will not
be acceptable. Therefore, in addition to other factors such as freeze up protection, any
leak over 20% would need serious system evaluation.
Power Input variations at 40 C CT
R22
R407C 10% leak
20% leak30% leak
40% leak
50% leak
R22
R407C10% leak
20% leak 30% leak
40% leak 50% leak
R22R407C
10% leak 20% leak 30% leak
40% leak50% leak
800
820
840
860
880
900
Pow
er In
put (
W)
CHWLT 7 CCHWLT 10 CCHWLT 12 C
Figure 6-24 Power input variation vs. pure R22 and percentage leaked R407C at various CHWLT for 40 C CT
Power input decreases from R22 down towards pure R407C and continuous to decline
as the leak percentage increases, as it can be seen on Figure 6-24. Again we can notice
kind of declining trend down to 20% leak, but unusual behavior from the 30% leak
onward. This confirms the earlier interpretation that a drastic change in the refrigerant’s
thermodynamic and thermo-physical property.
101
Power Input variations at 50 C CT
R22 R407C10% leak20% leak30% leak
40% leak50% leak
R22 R407C10% leak20% leak
30% leak40% leak50% leak
R22 R407C10% leak
20% leak30% leak40% leak
50% leak
800
840
880
920
960
1000
1040Po
wer
Inpu
t (W
)CHWLT 7 CCHWLT 10 CCHWLT 12 C
Figure 6-25 Cooling capacity variation vs. pure R22 and percentage leaked R407C at various CHWL for 50 C CT
Power Input variations at 50 C CT
R22 R407C10% leak20% leak30% leak
40% leak50% leak
R22 R407C10% leak20% leak
30% leak40% leak50% leak
R22 R407C10% leak
20% leak30% leak40% leak
50% leak
800
840
880
920
960
1000
1040
Pow
er In
put (
W)
CHWLT 7 CCHWLT 10 CCHWLT 12 C
Figure 6-25 shows similar trend as of 40 C SDT, Figure 6-25, however, indicates almost
the similar power input values for both pure R22 and pure R407C refrigerants and
slightly different trend on the 12 C CHWLT curve. It is clear that the power input
curves with higher CHWLT are higher than those with lower ones. Also the higher
CHWLT correspond to higher evaporating temperature or pressure and consequently
higher ET and SP. Although the pressure ratio or differential is lower with the higher
102
SP, but increase in the refrigerant mass flow rate reflected in increased capacity requires
higher power input.
The power input is higher in general for the 50 C CT than 40 C CT. This is due to the
higher pressure differential.
COP variations at 40 C CT
R22
R407C 10% leak 20% leak 30% leak 40% leak
50% leak
R22R407C 10% leak
20% leak 30% leak 40% leak50% leak
R22
R407C 10% leak20% leak
30% leak 40% leak50% leak
2.5
2.7
2.9
3.1
3.3
3.5
CO
P
CHWLT 7 CCHWLT 10 CCHWLT 12 C
Figure 6-26 Coefficient of Performance (COP) variation vs. pure R22 and percentage leaked R407C at various CHWLT for 40 C CT
Overall Coefficient of Performance (COP) shows declining trend from the pure R22 to
pure R407C and continuous downward as the leak increases for both 40 C and 50 C CTs
as well as all the CHWLTs. The COP varies from the maximum of about 3.2 at pure
R22 and 12 C LCHWT down to about 2.65 at 50% leak and 7 C CHWLT.
The tests indicate minor COP drop from pure R407C to 20% leak, however, it becomes
significant at 50% leak, Figure 6-26.
103
COP variations at 50 C CT
R22
R407C 10% leak20% leak
30% leak 40% leak
50% leak
R22
R407C 10% leak20% leak
30% leak40% leak 50% leak
R22
R407C 10% leak 20% leak 30% leak
40% leak 50% leak
2
2.2
2.4
2.6
2.8CO
PCHWLT 7 CCHWLT 10 CCHWLT 12 C
Figure 6-27 Coefficient of Performance (COP) variation vs. pure R22 and percentage leaked R407C at various CHWLT for 50 C CT
A similar trend was observed at the 50 C CT for various CHWLTs, with lower overall
COP when compared to the 40 C CT.
Declining COP at higher leaks is not a major factor, but it is another contributing factor
to re-evaluate system beyond 30% leak. This could become more and more influencing
factor as more and more nations are becoming energy efficiency and environmental
conscious.
104
7 SIMULATION AND MODELING RESULTS To simulate and calculate thermodynamic properties of the refrigerants including
blends, a software program has been developed by a major player in the refrigeration
industry, specifically refrigerant producer and supplier, Du Pont. The software is called
Genetron. The ASHRAE standard is used as a reference for the Enthalpy and Entropy,
being 0 at -40 C.
7.1 R22 and Pure R407C
Depending on suitability, the program uses an Equations of States for a specific family
of refrigerants. The Equations of State that used in this prediction software are discussed
in section 3.5 in details. Pure and mixed refrigerants properties are calculated using
relevant Equation of State to get the best accuracy possible. This method makes it
flexible in order to get the best and most accurate results.
The program also is capable of performing calculations for a basic refrigeration system.
It requires inputting the detailed cycle information at the various points. Namely,
average evaporating, average condensing, super heat at the compressor suction, gas
temperature at the evaporator outlet and liquid sub-cool before the expansion system.
On the compressor, it requires swept volume rate and isentropic efficiency to perform
the compression calculations. The volumetric efficiency, however, is not considered as
an input variable. This factor can affect accuracy of the calculations; however, it’s
influence on is considered insignificant on comparison basis. It could vary from one
condition to the other, which will be discussed later in this chapter.
Only the pure refrigerants could be simulated using this software. Due to the
refrigerants composition change that occurs during the process of leak and later top up,
the thermodynamic properties and performance could not be simulated in this section.
105
Figure 7-1 Genetron software calculation results screens for R22
106
Figure 7-2 Genetron software calculation results screens for R407C
Performance calculations were carried out, using the Genetron software, on both R22
and R407C refrigerants at various conditions. It uses basic refrigeration cycle with a few
data input options. The Compressor displacement rate, isentropic efficiency, means
condensing and evaporating temperatures, and other cycle details were used as input
data. The software program was able to come up with all calculated performance of the
cycle and thermodynamic properties of the refrigerant at various states. It should be
noted that due to the limited number of decimal points, the compressor displacement has
not been shown in these screen print outs. The displacement of the compressor that was
used in the testing was 0.0008 m3/s.
107
Genetron vs. Test R22, 40 C CT
500
1000
1500
2000
2500
3000
7 10 12CHWLT (C)
(W)
Capacity-Genetron
Capacity-Test
Power-GenetronPower-Test
Figure 7-3 Capacity and power input comparisons, Genetron vs. test results, R22 at 40 C CT
Genetron vs. Test R22, 50 C CT
500
1000
1500
2000
2500
3000
7 10 12CHWLT (C)
(W)
Capacity-Genetron
Capacity-Test
Power-Genetron
Power-Test
Figure 7-4 Capacity and power input comparisons, Genetron vs. test results, R22 at 50 C CT
Figures 7-3 and 7-4 show capacity and power input comparisons for R22 at 40 C and 50
C saturated discharge temperature, between the test and simulation results at various
chilled water leaving temperatures.
The simulation performance results indicate slightly higher capacity and slightly lower
power than the test results. These differences can be explained due to the fact that the
simulation calculations do not take into account losses associated with motor efficiency
108
and volumetric efficiency. Also, in any thermal testing system, here pilot test rig, some
tolerances attributed to the instrumentation accuracy. Considering the earlier two major
parameters, the variations are at acceptable level.
Genetron vs. Test pure R407C, 40 C CT
500
1000
1500
2000
2500
3000
7 10 12CHWLT (C)
(W) Capacity-Genetron
Capacity-Test
Power-Genetron
Power-Test
Figure 7-5 Capacity and power input comparisons, Genetron vs. test results, R407C at 40 C CT
Genetron vs. Test pure R407C, 50 C CT
500
1000
1500
2000
2500
3000
7 10 12CHWT C
(W)
Capacity-GenetronCapacity-Test
Power-Genetron
Power-Test
Figure 7-6 Capacity and power input comparisons, Genetron vs. test results, R22 at 40 C CT
109
Figures 7-5 and 7-6 shows capacity and power input comparisons for R407 C at 40 C
and 50 C saturated discharge temperature, between the test and simulation results at
various chilled water leaving temperatures.
Capacity and power variations appear to be wider at 7 C chilled water leaving
temperature and get closer as the temperature increases. Higher the chilled water
temperature means lower pressure ratio and also higher the temperature resolution
which helps lesser compression chamber leak ratio hence higher volumetric efficiency
and improving measurement accuracy.
Also, the average variation in the 50 C condensing temperature appears to be lower than
the 40 C one. Given that the earlier has higher pressure ratio than the latter one, the
compression chamber leak ratio would be expected to be higher with the earlier. Hence
lower volumetric efficiency than the latter.
7.2 Modeling of the system charged with R407C and then subjected to leak
Any refrigeration system can be subjected to adiabatic leak or isothermal leak.
However, the adiabatic leak process can happen only at very exceptional conditions
where the leak occurs at a very fast paste with totally insulated piping and vessels. On
the other hand, vast majority of the leak cases occur at isothermal process where the
leak process is slow and can be assumed constant temperature. The test rig in this work
has been subjected to isothermal leak process.
Comprehensive leak process simulations and experiments produced trends of
R32/R125/R134a composition change as function of rate of leak. The vapor and liquid
mass fraction of the most volatile refrigerant (R32) decreases during the vapor leak. As
a result, the mass fraction of the least volatile component increases. Starting from this
point, equations have been created to represent the composition change as function of
percentage of the leak.
Using this method, the liquid composition of the leaked refrigerant has been determined.
Since the system topped up for the leaked amount by liquid R407C, the new refrigerant
composition needs to be determined. These new compositions have been calculated
using proportional combination formulas.
110
7.2.1 Calculating the new mixture properties after the leak
Thermodynamic properties of the new refrigerant mixtures have been determined
Refprop program. The new refrigerant mixtures came to existence after the system
subjected to the isothermal leak at various percentages and topped up with R407C.
Therefore, they can not be considered as standard refrigerant mixture that assigned by
ASHRAE, hence no standard property tables.
This program, developed by the National Institute of Standards and Technology (NIST)
which makes it possible to define a new refrigerant mixture and determine their
properties. It is based on the most accurate pure fluid and mixture models currently
available. It implements three models for the thermodynamic properties of pure fluids:
equations of state explicit in Helmholtz (1882) energy, the modified Benedict-Webb-
Rubin (1995) equation of state, and an extended corresponding states (ECS) model. The
Helmholtz (1882) energy of the mixture components mixing rules applied. This
program makes it possible to define new refrigerant mixtures using non standard
composition values, and determine the thermodynamic properties.
Mass Fraction Change
y = 0.625x3 - 0.375x2 + 0.3x + 0.6R2 = 1
y = 0.2083x3 - 0.375x2 - 0.0833x + 0.3R2 = 1
y = -0.2083x3 + 0.125x2 - 0.0667x + 0.1R2 = 1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mass fraction Leak out
Mas
s Fr
actio
n
X R134aX R32X R125Poly. (X R134a)Poly. (X R32)Poly. (X R125)
Figure 7-7 Mass fraction change of the ternary blend R134a/R32/R125 and polynomial curve fitted models
111
Composition change behavior of a system subjected to isothermal leak is plotted on the
graph, Figure 7-7. The graph clearly indicates that during isothermal leak from the
system, R32 and R125 mass fractions drops while R134a mass fraction increases inside
the ternary mixture. The mass fraction decrease of R32 faster than R125. This is due to
the fact that R32 is more volatile than R125 and therefore tend to boil first and increase
its concentration in the vapor phase. The fitted polynomial equations matched perfectly
the trend so that the trend graphs are completely covered by the graphs plotted by the
equations. The calculated value of 2R = 1 indicates a perfect match with close to zero
deviation between the fitted equations and the data gathered from the tests.
Table 7–1 The calculated new compositions after the system subjected to various leak percentages without adding Leak mass fraction LRi Xi134a Xi125 Xi32 Total composition
0 0.52 0.25 0.23 1 0.1 0.546875 0.2443717 0.2181283 1.009375 0.2 0.57 0.2399936 0.2000064 1.01 0.3 0.593125 0.2356159 0.1768841 1.005625 0.4 0.62 0.2299888 0.1500112 1 0.5 0.654375 0.2218625 0.1206375 0.996875
aRiiiaRi XLRLRXN 134134 52.0)1( ×+×−= 7.1
125125 25.0)1( RiiiRi XLRLRXN ×+×−= 7.2
3232 23.0)1( RiiiRi XLRLRXN ×+×−= 7.3
Table 7–2 The calculated new compositions after the system subjected to various leak percentages and topping up the system
New Composition after leak and top Up % leaked and topped up XN 134a XN 125 XN 32 Total composition
10% 0.5226875 0.24943717 0.22881283 1.000937520% 0.53 0.24799872 0.22400128 1.00230% 0.5419375 0.24568477 0.21406523 1.001687540% 0.56 0.24199552 0.19800448 150% 0.5871875 0.23593125 0.17531875 0.9984375
7.2.2 Calculating the system performance including capacity, power input and COP
Using the above curve fittings, equations and tables, the Refrprop software were used to
calculate thermodynamic properties of the refrigerant mixtures with changed
compositions. Two stages of process applied, first subject the system to leak at various
fractions (percentages), calculated the new compositions’ mass fractions. At the second
112
stage, each percentage leak topped up with the same percentage of the pure R407C
refrigerant which would create new composition mass fractions. Mass fraction of the
new composition then calculated using equations 7.1 to 7.3 as well as the table 7.1.
Thermodynamic properties of the new composition then were calculated using the
Proref software and tables were created.
To complete the modeling, the calculated thermodynamic properties were downloaded
to Excel spread sheet for further system performance determinations including
capacities, power inputs and COPs at various conditions. The formulas from Chapter 3
used to determine capacities, power inputs as well as COPs at the same conditions the
tests were to be conducted. After calculation of the above, comparisons were made with
the test results at the corresponding conditions. The following graphs show the
comparison data which indicates performance prediction with a reasonably good
accuracy. It should be noted that theoretical performance prediction in thermal systems,
specifically refrigeration systems, with fewer than 10% accuracy is considered a good
modeling result.
Figure 7-8 Comparison of the capacities from the test results and the modeling at various CHWLTs and various leak percentages, 40 C CT
Capacity comparison, Test vs modeling at 40 C CT
1500
1700
1900
2100
2300
2500
2700
2900
10% leak 20% leak 30% leak 40% leak 50% leak
(W)
7 C Test7 C Modelling10 C Test10 C Modelling12 C Test12 C Modelling
113
Capacities calculated by modeling predictions and test results at 40 C condensing
temperature were compared on Figure 7.8. Comparisons were also made at various
chilled water temperatures. Both test and model results show similar trend as function of
percentage of leak that the system subjected to. The biggest variation is seen at 20%
leak (about 3% deviations) and smallest at 50% leak (less than 1% deviation). With
such a small variations, this proves an excellent match between the predicted
performance results and tested performance results.
Figure 7-9 Comparison of the capacities from the test results and the modeling at various CHWLTs and various leak percentages, 50 C CT
The capacity comparison between the modeling prediction and the test results at 50 C
condensing temperature is shown on Figure 7-9. The modeling capacity change trends
as function of leak percentage indicates similarity to that of the test results, however,
with slightly higher deviations of up to 5%.
Capacity comparison, Test vs modeling at 50 C CT
1500
1600
1700
1800
1900
2000
2100
2200
2300
2400
2500
10% leak 20% leak 30% leak 40% leak 50% leak
(W)
7 C Test7 C Modelling10 C Test10 C Modelling12 C Test12 C Modelling
114
Figure 7-10 Power input comparison from the test results and the modeling at various CHWLT temperatures and various leak percentages, 40 C CT
Calculating the required power to compress certain amount of gas using theoretical
modeling is a challenge. There are quite a few minor factors that affect the compression
process and are not easy to estimate extend of their effect on the outcome. In this case, a
polytropic compression model is used to compress gas inside a chamber with suction
and discharge valves. This model and fundamentals of thermodynamic properties
determination are discussed in the section 3.6 and chapter 3 in details.
The model prediction of power input at 40 C condensing is shown on the Figure 7-10
with comparison to the test results. The model predicts slight power input decrease slop
as the leak percentage increases. At 12 C CHWLT, the graph indicates accurate enough
prediction with less than 3% deviation which increases to over 5% at 50% leak. The
deviations are slightly higher at 10 and 7 C LCHWT to a bit over 6% and increased gap
at 50% leak.
Power Input comparison, Test vs modeling at 40 C CT
600
650
700
750
800
850
900
10% leak 20% leak 30% leak 40% leak 50% leak
(W)
7 C Test7 C Modelling10 C Test10 C Modelling12 C Test12 C Modelling
115
Figure 7-11 Power input comparison from the test results and the modeling at various CHWLTs and various leak percentages, 50 C CT
Power input comparison at 50% as shown on Figure 7-11, indicates relatively better
prediction with average deviation of less than 4% and maximum of about 5%.
Deviations between the model and tests are generally consistent with various LCHWT
conditions.
Minor trend change on the test result at the 50% leak represents slight inconsistency
which can be explained by variations in which the system components characteristics
may be affected. This could well be behavior of the expansion valve at a very high leak
condition causing unstable operation of the component and consequently pressure
fluctuations. This phenomenon can also be seen at 40 C condensing as well.
Despite complexity of the power input calculations and the fact couple of effecting
parameters are hard to predict accurately enough, the out come is very promising. The
modeling predicted power with encouraging accuracy with average deviation of 3 to 4%
and maximum of around 5% at complete range of conditions.
Power Input comparison, Test vs modeling at 50 C CT
700
750
800
850
900
950
1000
10% leak 20% leak 30% leak 40% leak 50% leak
(W)
7 C Test7 C Modelling10 C Test10 C Modelling12 C Test12 C Modelling
116
Figure 7-12 COP comparison from the test results and the modeling at various CHWLTs and various leak percentages, 40 C CT
The Coefficient of performance (COP) values were calculated by dividing capacity by
power input and plotted on the Figures 7-12 and 7-13 at various conditions vs. the leak
percentages. The model prediction indicates very small change in COP values as
function of the leak percentages with a slight downward trend to 30% leak and slight
upward trend toward 50% leaks. The slight increase at high leak is due to increase in
concentration of R134a refrigerant in the system. R134a is a more efficient refrigerant
when compared to the other components of the ternary blend which constitutes R407C.
At 40 C condensing, the average deviation between the modeling predicted values and
the test results is around 3% with maximum reaches over 6% at the highest leak
percentage. The jump can only be related to combination of system inconsistency,
instrumentation accuracy and test method effects at high leak conditions.
COP comparison, Test vs modeling at 40 C CT
2.5
2.7
2.9
3.1
3.3
3.5
10% leak 20% leak 30% leak 40% leak 50% leak
COP
7 C Test7 C Modelling10 C Test10 C Modelling12 C Test12 C Modelling
117
Figure 7-13 COP comparison from the test results and the modeling at various CHWLTs and various leak percentages, 50 C condensing
The COP comparison graph at 50 C condensing, Figure 7-13, shows similar trend as of
the 40 C condensing. The COP values at the earlier vary from 2.7 up to 3.1, while these
values range from 2.1 up to 2.4. This is due to thermodynamic fundamentals of the
vapor compression cycle which requires higher power to compress at higher pressure
ratios.
On the other hand, the test results show slight drop in COP at the high leak percentages.
This difference in trend may look a bit unusual at fist glance; however, this can be
explained in the context of the system performance. Since COP is product of dividing
capacity by power input, the variations and deviations were better explained on the
related graphs.
Modeling a refrigeration system that is subjected to leaked Zeotropic refrigerant at
various mass fractions and conditions is a very challenging task. Fragmented works
have been done on this subject, but this is the first work that completely models a
system that isothermally leaked and then topped up with the blend r407C refrigerant.
COP comparison, Test vs modeling at 50 C CT
2
2.2
2.4
2.6
2.8
3
10% leak 20% leak 30% leak 40% leak 50% leak
COP
7 C Test7 C Modelling10 C Test10 C Modelling12 C Test12 C Modelling
118
Despite the enormous challenge faced, in overall, the modeling predicted performance
values accurately enough and the out come is very promising. The modeling predicted
power with encouraging accuracy with average deviation of 3 to 4% and maximum of
around 5% at complete range of conditions.
119
8 CONCLUSIONS AND RECOMMENDATIONS
8.1 GENERAL CONCLUSIONS
This research work can be categorized in two major areas.
First, comparing and analyzing system performance when converted from the HCFC-22
to HFC-407C. These types of drop in conversions are not unusual in the industry.
Therefore, it is very important to understand what to expect after the conversion from
all aspects such as capacity, efficiency, pressures and temperatures. The test results
proved that HFC-407C is the closest possible to HCFC-22 when it comes to refrigerant
conversion in those types of equipments with exception of the chillers of flooded
evaporator types.
When charged with the new refrigerant, the system capacity dropped in the range of 4 to
6%, depending on the pressure ratio, and power input remained almost the same. The
condensing and evaporating pressures increased slightly in the vicinity of about 5%,
which has a negligible factor on the pressure vessel and piping design. However, gliding
temperature difference (GTD) phenomenon in both evaporator and condenser was the
major difference, meaning no longer the temperature constant phase change. This
phenomenon has more prominent effect on the evaporation than condensation. It
reduces heat transfer between the refrigerant and to be cooled fluid, here water. And
also for this reason, chillers with flooded evaporators can not be converted to HFC-
407C. Pool boiling would end up concentration of low volatile component of the
mixture in the liquid.
The tests also proved that, for a given condensing temperature or pressure, the capacity
and power input increases as the evaporating temperature or pressure rises. Also for a
given evaporating temperature or pressure, the capacity decreases and power input
increases as the condensing temperature rises. This behavior is line with the single
speed positive displacement compressors such as the reciprocating type used for
purpose of this research work.
The second and most important area to cover was to analyze the effect of system
refrigerant gas leak on performance, efficiency and reliability of the system. The system
120
charged with the new HFC-407C was deliberately subjected to refrigerant leak at
various leak stages. It is a common practice in the industry these days to evacuate and
completely recharge when part of the new refrigerant blend was leaked from the system.
This has proved to be extremely costly exercise with grave environmental ramifications.
The aim was to objectively determine to what extend the gas leak can be still acceptable
without going through the expensive complete gas charge. The effect of leak was tested
and verified at 10% steps, from 10% up to 50% mass fraction for the total charge. At the
leaks beyond 30%, the adverse effect on the capacity becomes more significant, from 8
to about 15% decrease. While the power input decreased at slower pace, from 3% up to
about 8% depending on the test conditions. This translated to COP decrease ranging
from 4 to about 7%. This capacity loss and efficiency decrease are significant figures
which suggests that the system, here chiller, can not be allowed to degrade the
performance to that extend and still continue operating.
The system was also mathematically modeled to predict detailed system performance
and effect of the leak at various conditions. To make this feasible and accurate enough,
two separate approaches made, first system performance for pure R22 and R407C, and
second system subjected to range of leak fractions. The earlier model was relatively
straight forward when compared to the latter. Modeling a system charged with R407C
ternary mixture and subjected to range of leaks poses enormous challenge.
The first part has been successfully modeled and predicted all the factors and
performance with excellent accuracy when compared to the test results. In these
approach pure refrigerants R22 and R407C were used and simulated the system
behavior at range of conditions.
However, the second part was the most challenging ever. Comprehensive leak process
simulations produced trends of R32/R125/R134a composition change as function of rate
of leak. Starting from this point, equations have been created to represent the
composition change as function of percentage of the leak. Using this method, the liquid
composition of the leaked refrigerant has been determined. Since the system topped up
for the leaked amount by liquid R407C, the new refrigerant composition needed to be
determined. These new compositions have been calculated using proportional
combination formulas. The system thermodynamic cycle was also modeled to calculate
121
capacity, power input and COP at the range of the conditions. Despite many affecting
parameters and complexity of the model, the mathematical model successfully predicted
the test outcome with a very reasonable accuracy, averaging around 3% with some
times reaching to 5 to 6%. Needless to say that achieving accuracy of this level should
be considered as fulfilling effort of this work.
In addition to this, another major factor is the effect of large leak on the system
reliability. Chiller systems are normally fitted with pressure and temperature controls set
at specific range to protect the system from various malfunctions such as low pressure,
high pressure and freeze up etc.
The tests indicated that up to about 20% mass fraction leak had adverse effect on the
system capacity of about 5%. While at the same range, the efficiency (COP) decrease
was only less than 3%.
8.2 RECOMMENDATIONS FOR FUTURE WORK
This research work focused on the effect of leak on the system performance and system
reliability as well as its adverse effect on system maintenance and cost associated with
tackling complexity of leak in the systems charged with zeoptropic refrigerant blend. As
these systems, in the event of leak, can not be simply topped up with refrigerant, it
introduces major challenges to the industry dealing with this issue. The chiller systems
contain significant amount of refrigerant due to its shear size and capacity. Any
complete refrigerant replacement would be a costly exercise. This work showed that,
depending on the system and efficiency and safety sensitivity, leaks up to about 10-20%
can be tolerated by just topping up the system charge with R407C. This is due to the
fact that the efficiency and characteristic change of the system would be minimal.
The key point here is how to measure or estimate the amount of leak on a commercial
system installed on site. Pressure and/or temperature on the evaporator is one way of
estimating this. To this end, chilled water leaving temperature is normally set at fixed
value, e.g. 7 C, which corresponds to a certain range of evaporating temperatures
depending on the type the evaporator and condenser is used. Usually amount of the
pressure or temperature drop relates somehow to extend of the leak out of the system.
This suggests that there could be a complex relation between percentage of the leak and
122
variables like temperature and pressure. The refrigerant pressure can easily be measured
on site by the technical staff, however, measuring accurate temperature depends on the
system equilibrium, or in other words the refrigerant has more or less the same
temperature at different parts of the system. It is recommended that the future work
focuses on investigating the effect of leak amount on pressure at given temperatures in
the systems charged with zeotropic refrigerants, more specifically HFC-407C. The
research need to come up with proposals that would result in tangible series of graphs
and correlations accurately predict the ratio of the leaked and remaining mixture as well
as composition of the mixture after leak. Using these graphs and equations, the new
composition of the remaining mixture could be calculated. From here, using
composition mass fraction of the remaining mixture in the system, fraction or
percentage of the leaked mixed refrigerant can be determined. To achieve this, reverse
process of calculations need to be adopted.
In long term, these method and correlations then can be used to develop an instrument
to determine state of the mixture in the system, estimating fraction of leakage. By
knowing the fraction, the leaked amount can be calculated. This would give enough
information together with other findings of this work, to decide whether to top up the
system with R407C or recover and completely recharge the system. In overwhelming
majority of the cases, they would end up only topping up the system, which in turn
would save significant amount of time and money to the chiller plant owners.
123
REFERENCES Apera C., de Roddi, F and Greco, A., “Experimental evaluation of R22 and R407C
evaporative heat transfer coefficients in a vapor compression plant”, International
Journal of Refrigeration, vol. 23, issue 5, August 2000.
Belghazi, M., Bontemps, A., Signe, J.C. and Marvillet, C., “Condensation heat transfer
of a pure fluid and binary mixture outside a bundle of smooth horizontal tubes.
Comparison of experimental results and classical model”, International Journal of
Refrigeration, Vol. 24, No. 8, December 2001.
Cengel. Y, Heat Transfer – A Practical Approach, McGraw-Hill USA, 1998
Chen Jiufa and Kruse Horst, “Concentration Shift Simulation for the Mixed
Refrigerants R404A, R32/R134a and R407C in an Air-Conditioning System”,
HVAC&R Research, Vol. 3 No 2, April 1997
Cheng-Shu Kuo, Chi-chuan Wang, Horizontal Flow Boiling of R22 and R407C in a
9.52mm Micro-fin tube, Applied Thermal Engineering Vol.16 (1996), pp719-736
Chen, L., Taylor, A., Tolouee, C. M., Leonardi,E., Behnia, M., “A non-azeotropic
mixture performance test” , Journal of Australian Refrigeration, Air-Conditioning and
Heating (AIRAH), March 1991.
Connon, H.A., Drew, W.D., “Estimation and calculation of thermodynamic properties
for a non-azeotropic refrigerant mixture”, International Journal of Refrigeration,
Volume 6, No. 4, 1983.
Cooper, M. G., “Saturation nucleate pool boiling: a simple correlation”, International
Chemical Engineering Symposium, Volume 86, 1984.
Devotta, S., Waghmare, A. V., Sawant, N. N. and Domkundwar, B. M., “Alternatives to
HCFC-22 for air conditioners”, Applied thermal Engineering, Vol. 21, 2001.
124
Dowining, R.C., “Refrigerant Equations”, ASHRAE Transactions, Vol. 80, Part II,
1974.
Edmister, W. C., “Applied Hydrocarbon Thermodynamics”, Hydrocarbon Processing,
Volume 47, No. 9, 1968.
Gabrielli, C. and Vamling, L., “Changes in optimal design of a dry-expansion
evaporator when replacing R22 with R407C”, International Journal of refrigeration,
Volume 21, Issue 7, 1998.
Gabriellii, C. and Vamling, L., “Drop-in replacement of R22 in heat pumps used for
district heating – influence of equipment and property limitations”, International Journal
of Refrigeration, Vol. 24, issue 7, July 2001.
Genetron 407C Product Brochure, Allied Signal Chemicals, G525-013, 1/95 5K, 1995.
Goto, M., Inoue, N. and Ishiwatari, N., “Condensation and evaporation heat transfer of
R410A inside internally grooved horizontal tubes”, International Journal of
refrigeration, Volume 24, Issue 7, 2001.
Gungor, K. E., and Winterton, R. H. S., “A genral correlation of flow boiling in tubes
and annuli”, International Journal of Heat and Mass Transfer, Volume 29, 1986.
Heating, Ventilating and Air-Conditioning (HVAC) Systems and Equipment, ASHRAE
Handbook, 2004.
Huber, M.L. and Ely, J.F.,” A predictive extended corresponding states model for pure
and mixed refrigerants including an equation of state for R134a”. Int. J. Refrigeration
17: 18-31, 1994.
Huber, M.L., Friend, D.G. and Ely, J.F,.” Prediction of the thermal conductivity of
refrigerants and refrigerant mixtures”, Fluid Phase Equilibria 80: 249-261, 1992.
125
Johansson, A. and P. Lundqvist, “A method to estimate the circulated composition in
refrigeration and heat pump systems using zeotropic refrigerant mixtures”, International
Journal of Refrigeration, July 2000.
Kandlikar, S. G., “A general correlation for saturated two phase flow boiling heat
transfer inside horizontal and vertical tubes”, Journal of Heat Transfer, volume 112,
1989.
Kattan, N., Thomas, J. R. and Favrat, D., “Flow boiling in horizontal tubes: part 3 –
development of a new heat transfer model based on flow pattern”, transactions of the
ASME, Volume 120, 1998.
Kim M. S., Ph.D., and D. A. Didion, D. Eng., P.E., “Simulation of Isothermal and
Adiabatic Leak Processes of Zeotropic Refrigerant Mixtures”, International Journal of
Refrigeration, Vol. 1, issue 1, January 1995.
Kruse H., “Alternative refrigerants – Synthetic and Natural Fluids”, International
Conference, Romania, Institute of refrigeration, 1996
Lemmon, E.W., “A generalized model for the prediction of the thermodynamic
properties of mixtures including vapor-liquid equilibrium”, Ph.D. thesis, University of
Idaho, Moscow, ID, 1996.
Lemmon, E.W. and Jacobsen, R.T., “A generalized model for the thermodynamic
properties of mixtures”, Int. J. Thermophysics 20: 825-835, 1999.
McLinden, M.O., Klein, S.A. and Perkins, R.A., “An extended corresponding states
model for the thermal conductivity of refrigerants and refrigerant mixtures”. Int. J.
Refrigeration 23: 43, 2000.
Man-Hoe Kim, “Performance evaluation of R22 alternative mixtures in a breadboard
heat pump with pure cross-flow condenser and counter-flow evaporator”, Energy, 27,
167-181, 2002.
126
Min Soo Kim, Ph.D., David A. Didion, D. Eng., PE, “Evulation of a Leak/Recharge
Process of Refrigerant Mixtures”, HVAC&R Research, Vol. 1, No 3, July 1995.
Olivier Zurcher and Thomas John R., and Daniel Farvat, “In-Tube Boiling of R407C
and R407C/Oil Mixtures Part II: Plain Tube Results and Predictions”, HVAC&R
Research, Vol. 4, No 4, October 1998.
Reid, R.C. and Prausnitz, J.M., Poling, B.E., “The Properties of Gases”, 4th edition, Mc
Graw Hill, 1987.
Report, “The Montreal Protocol on substances that deplete the ozone layer”,
Department of Environment and Heritage, Australian Government, October 2004.
Samuel M. Sami and Daniel E. Desjardins, “Performance analysis of new alternatives to
HCFC-22 inside air/refrigerant-enhanced surface tubing”, International Journal of
Energy research, 24:759-768, 2000.
Shah, M., “Chart Correlation of Saturated Boiling Heat Transfer: Equations and Further
Study”, ASHRAE transactions, Volume 88, 1982.
Shao, David W. and Granryd, Eric G., “Flow Pattern, Heat Transfer and Pressure Drop
in Flow Condensation Part II: Zeotropic Refrigerant Mixtures (NARMs)”, HVAC&R
Research, Volume 6, No. 2, April 2000.
Soave, G, “Rigorous and Simplified Procedures for determining the pure component
Parameters in the Redlich-Kwong-Soave Equation of State”, Chemical Engineering
Science, Volume 35, 1980.
Thomas, J. R., “Boiling of new refrigerants: a state of art review”, International Journal
of Refrigeration, Volume 19, No. 7, 1996.
127
Tillner-Roth, R., Baehr, H.D., “An international standard equation of state for the
thermodynamic properties of 1,1,1,2-tetrafluoroethane (HFC-134a) for temperatures
from 170 K to 455 K at pressures up to 70 MPa”, Journal of Phys. Chem. Ref. Data 26,
1994.
Tillner-Roth, R., Yokozeki, A. , “An international standard equation of state for
difluoromethane (R-32) for temperatures from the triple point at 136.4 K to 435 K at
pressures up to 70 MPa”, Journal of Phys. Chem. Ref. Data 26, 1997.
Tolouee, C., M., Behnia, M. and Leonardi, E., “Computer simulation of mixed
refrigerant cycles”, Journal of Australian Refrigeration, Air-Conditioning and Heating
(AIRAH), May 1991.
United Nations Environment Program (UNEP), “Protecting Ozone layer”, volume 1,
Refrigerants, 2001.