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.