Citation
Study of Cooling Production with a Combined Power and Cooling Thermodynamic Cycle

Material Information

Title:
Study of Cooling Production with a Combined Power and Cooling Thermodynamic Cycle
Copyright Date:
2008

Subjects

Subjects / Keywords:
Ammonia ( jstor )
Boiling ( jstor )
Cooling ( jstor )
Inlet temperature ( jstor )
Inlets ( jstor )
Liquids ( jstor )
Rectifiers ( jstor )
Turbines ( jstor )
Vapors ( jstor )
Working fluids ( jstor )
City of Gainesville ( local )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Embargo Date:
12/18/2004

Downloads

This item has the following downloads:


Full Text












STUDY OF COOLING PRODUCTION WITH A COMBINED POWER AND
COOLING THERMODYNAMIC CYCLE















By

CHRISTOPHER MARTIN


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2004

































Copyright 2004

by

Christopher Martin















ACKNOWLEDGMENTS

I would like to express my appreciation to those people who supported this work

and provided me with the encouragement to pursue it. First I would like to thank my

advisor, Dr. D. Yogi Goswami, for his teaching and providing me with this opportunity.

Additionally, I would also like to thank Dr. Skip Ingley, Dr. William Lear, Dr. S. A.

Sherif, and Dr. Samim Anghaie for serving on my advisory committee. Their time and

consideration are appreciated. Special thanks are also extended to the editorial staff of

the Solar Energy and Energy Conversion Laboratory (SEECL), Barbara Graham and

Allyson Haskell. Also, the advice and humor of Chuck Garretson have been much

appreciated during my time at the SEECL.

There are also many colleagues I would like to thank for their help, consultation,

and camaraderie. Gunnar Tamm and Sanjay Vijayaraghavan have provided excellent

examples that I have tried to follow. I began this process with Nitin Goel and Amit

Vohra, with whom I have become friends. Also I have made friends with the recently-

joined students, Madhukar Mahishi, Shalabh Maroo, and Ben Hettinger.

I would like to also acknowledge the support of my parents, Lonnie and Loretta.

Finally, and most importantly, I want to thank my wife Janell for her unquestioning

support during this work. Without it, I doubt that I would have reached this personal

milestone.
















TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ................................................................................................. iii

LIST OF TABLES ............................... ... ......... .... ... ..... .............. viii

LIST OF FIGURES ................................. ...... ... ................. .x

N O M E N C L A T U R E ............................ ................................................... ................... xiii

ABSTRACT ....... ............. .............. .. ...... .......... .......... xvii

CHAPTER

1 IN TR OD U CTION ............................................... .. ......................... ..

M o tiv atio n ........................................................ ................ .. 2
P ow er-C cooling C oncept................................................................... .......... ..........4
P problem D definition ................. ...... ........................ ...... ........ ......... .... .5
R research O bjectives.......... ................................................................ ........ .... .6

2 BACK GROUND AND REVIEW ..................................................................... .... .7

B ack g rou n d .................................................................................................. .7
ORC D evelopm ent .................................... .......................... ....
A m m onia-W ater C ycles ............................................................. .....................9
P ow er-C ooling C oncept................................................................... ... .................. 10
Prior W ork .............. .. ................ ............................................................... ....... 12
Other Pow er-Cooling Concepts........................................................ ............... 15
C conclusion ...................................................................................................... ....... 17

3 TH EORETICAL STU D Y ................................................. ............................. 18

M o d e l .........................................................................................................1 8
O operating M mechanism s ....................................................................... ..................20
Effect of Boiling Pressure .................................. ........ ................... 20
Effect of M ixture C concentration ................................... .................................... 22
Effect of Boiling Tem perature.......................................................... ...............23
C ooling P rodu action .......... ..... .............................................................. ........ .. ...... .. 24
Exhaust Tem perature............................................................. ............... 24










V ap or F low R ate ............. ... .................................... .................. ............ ....... .. 2 7
R ectification ..................................................................................................... ..... 27
Perform ance M easures............................................. ................... ............... 30
W ork P reduction .............................. ......................... ... ...... .... ...... ...... 30
C ooling P rodu action .......... ............................................................. ....... ........ .. 30
C o n c lu sio n ...................................................................................................... 3 1

4 EXPANDER CONSIDERATIONS ........................................ ....................... 33

W working F luid P properties ........................................ ............................................33
Prelim inary M machine Sizing ................................................ ............ ............... 35
Technology R review .................. ..................................... ................. 37
Dynamic M machines .................. ........................... .. ...... ................. 37
S team T u rb in e s ............................................................................... 3 8
D isplacem ent M achines............................................................ .....................39
A additional C considerations ................................................. ............................. 41
Expansion process .......................................... .. .. ........... .... ....... 43
C conclusion ...................................................................................................... ....... 44

5 EXPERIMENTAL APPROACH ........................................ .......................... 46

S etu p D description ......... .................................................................... ......... ....... 4 6
E x p a n d e r .....................................................................................5 0
R e c tifie r ....................................................................................................5 2
A b sorb er ................................................................... 53
P u m p ............................................................................... 5 4
D ata C o llectio n ............................................................................... 5 5
E x p erim mental M eth o d ........................................................................................... 5 5
E xperim ents Perform ed ...........................................................57
B oiler E x it T em p eratu re ......................................................................................57
B asic Solution Concentration ................................................................... 57
S u p e rh e atin g ................................................................. ...............................5 7
Absorption Temperature.................................... ......... 57
N o zzle F low A rea .......................................................................................... 5 8
R ectificatio n ................................................................5 8
C conclusion ...................................................................................................... ....... 58

6 EXPERIMENTAL RESULTS ..................................................59

C confirm action of Trends ................................................ ................ 59
P re ssu re V ariatio n ......................................................................................... 5 9
Concentration V ariation .............................................. ............... 60
T em p eratu re V ariation ................................................................................... 6 1
A b so rp tio n P ressu re .............................................................................................6 2
R ectifier P penalty .............................................................63
C concept D em onstration ...........................................................64
E x p an der P perform an ce .......................................................................................... 66


V









C conclusion ......................................... ................................................... 7 1

7 DISCUSSION AND CONCLUSIONS ............................................................... 72

Cooling Conditions ............... .............. ......... ............... .... ...... 72
O ptim um R ectification .............................................. .............................. 73
O overall O ptim um C ooling ........................................................ ............... 74
Exhaust Tem perature....................................... ...... ................... ............... 75
Implementation ........................................................... ............... 77
V ap or Q u ality ...............................................................7 8
Rectifier Implementation..................... ........ ............................ 79
E xperim ental O b servations.............................................................. .....................79
Absorption Pressure.............. .......................... .........80
R ectifier Pressure E ffect............................................................... ............... 83
C conclusion ......................................... ................................................... 85

8 RECOMMENDATIONS ............................................................ ...............88

E x p erim mental T estin g .............................................................. ........ .................... 8 8
Practical A application ....................... .................... ................... ........ 90
O R C C om prison ...................................... ............... .... ....... 90
C cooling P rodu action ......... .......................................................... ... .. .... ... ....9 1
C conclusion ........................................................................ .... ......... 92

APPENDIX

A PROPERTY EVALUATION ....................................................... ............... 94

Pure C om ponent Properties ............................................... ............................. 94
L iquid M ixture Properties................................................. .............................. 97
V apor M ixture P properties ........................................ ............................................99
E equilibrium C condition s ..................................................................... ...................99
Com puter Im plem entation ............................................................................ 101
Saturation Tem peratures......................................................... ............... 101
E n th a lp y ................................................... ................ 10 2
E n tro p y ..............................................................................................................1 0 3
S p ecific V o lu m e .............. ................................................ ................. ... .. 10 5

B MODEL FORMULATION........................................................... .................. 106

Therm odynam ic Form ulations....................................................... ............... 106
C om puter Im plem entation ................................................................................. 110
Saturated L iquid Pressure................................................................. ........... .111
Two-Phase Mixture Determination .... .......... ........................................112
Saturated Liquid Concentration..................................................................... 115
Saturated Vapor Concentration ...................................................................... 116
Two-Phase Mixture Enthalpy.....................................................................117









Temperature Determination Using Enthalpy ....................................................118
Isentropic Tem perature D eterm ination.............................................................119
Overall Cycle Calculation ...........................................................................120

C EXPERIM ENTAL DETAILS ........................................................ ............... 127

Instrum ent Settings .................. ..................................... ...... ............ .. 127
D ata A acquisition System ........................................................ ............. 127
Gas Chrom atograph........................................ ........................ ............... 127
Uncertainty of Direct M easurem ents................................................ ............... 128
T em p eratu re ................................................... ............ ................ 12 9
Pressure........... .... ................... ................................ ......... 130
V olu m e F low R ate ............................... ............................ ...... ..................... 130
C o n c en tratio n .............................................................................................. 13 1
S h aft S p eed .................................................................................. ............. 13 2
Uncertainty of Derived Measurements.....................................................133
V ap or C on centration ........................................... ........................................ 133
M ass F low R ates .......... .... .. .................... ......... ........ ........ 134
Power Output .......................... ........ .... ........ 135
E x p an der E fficien cy ........................................... ......................................... 13 5
E quipm ent Specification ................................................ ..................................... 136
Instrum entation .................. .............................. ...... ................. 136
Expander D details ................................... .. .. ...... ...............137

D EXPANDER AIR TESTING ............................................................................140

Experimental Setup.................. ................. ..... .......... 140
T e st R e su lts ..............................................................................................14 3
E x p erim mental D details ....................................................................... ...................14 6
M easurem ent U uncertainties ..................................................... ...... ......... 147
Equipm ent Specification ............................................................................148

REFERENCES ..................................... ................ .............149

BIOGRAPHICAL SKETCH ............................................................. ............... 155
















LIST OF TABLES


Table pge

3-1 Flow identification for the configuration of Figure 3-1 ..........................................18

4-1 Fluid properties for a typical ammonia-water concentration and other power cycle
fluids for isentropic expansion from saturated conditions at 1000 C to
condensation/absorbtion at 350 C ....................................... ........................ 34

4-2 Single stage specific speed calculations versus nominal work output and shaft
sp e e d ............................................................................ 3 6

4-3 Approximate specific speed and specific diameter ranges for efficient (>60%)
single stage expander types [49, 50]. ........................................... ............... 37

4-4 Reported turbine operating parameters and efficiencies for three systems using an
am m onia-w after w working fluid [54-56]. .......................................... ............... 38

4-5 Estimated operating data for the three turbine stages of a Kalina-based bottoming
cy c le [5 3 ] ......................................................................... 3 9

4-6 Reported efficiencies of scroll expanders [18, 20, 62, 63] .....................................40

6-1 Measured decrease of absorption pressure with basic solution concentration .........61

6-2 Measured data indicating effects of absorption temperature...............................63

6-3 Averaged values for rectifier operation .............................................................. 63

6-4 Values for rectifier operation highlighting penalty to work production. ................64

6-5 Averaged conditions for the testing of Figure 6-4. ............... ................. ..........66

7-1 Typical operating characteristics for cooling and work optimized cycles...............75

A-1 Coefficient and reference state values for ammonia and water .............................96

A-2 Reference values for reduced property computation ........................ ............96

A-3 Coefficient values used to compute excess properties...................... ..............98









A-4 Coefficient values for the determination of mixture bubble and dew point
tem peratures. ...................................................................... 100

B-l Flow identification for the configuration of Figure B- ...............................107

C-1 Calibration factors for the thermocouples used in this work ..............................129

C-2 Pressure transducer calibration factors........................................... .................. 130

C-3 Stated uncertainties for pressure transducers. .............................. ......... ...... .130

C-4 Derived measurement uncertainty summary ........................................................ 136

C-5 Detailed descriptions of the instrumentation and equipment used for this work...136

D-l Thermal and torque-based measurement uncertainties. .......................................147

D-2 Summary of the equipment and components used for the air tests...................148
















LIST OF FIGURES


Figure pge

1-1 Ideas for using an ORC to incorporate a renewable element into distributed power
g en eratio n .......................................................... ................ 3

2-1 Schematic of the power-cooling cycle. ............................................ ...............11

2-2 Ammonia-water phase equilibrium diagram highlighting the source of cooling
tem peratures. .........................................................................12

3-1 Power-cooling schematic used for modeling. .................................. ............... 19

3-2 Conceptual relationship between the factors affected by boiling pressure ............21

3-3 Output parameter variation as a function of boiling pressure. ................................22

3-4 Variation of output parameters as a function of basic solution concentration.........23

3-5 Effect of boiler exit temperature on output parameter profiles.............................24

3-6 Computed effect of vapor concentration and inlet temperature on expander exhaust
tem perature...............................................................................................26

3-7 Computed effect of expander efficiency and inlet temperature on expander exhaust
tem perature.................................................................... ........... ....... 26

3-8 Beneficial effect on expander exhaust temperature as a function of increasing
rectification... ....................... .............. ................ 28

3-9 Effect of rectification and rectifier efficiency on work production.........................29

5-1 Schem atic of experim ental setup.................................... ........................... ......... 47

5-2 Photograph of experimental setup ................................................. ............... 48

5-3 M odified turbine used for experimental testing ........................................... .......... 51

5-4 Original and modified absorber configurations. ................... ................... .......... 53

6-1 Measured effect of pressure variation on vapor quantity and concentration. ..........60









6-2 Measured effect of basic solution concentration on vapor production ..................61

6-3 Measured change in vapor flow rate (relative to basic solution flow) due primarily
to changes in boiling tem perature. ........................................ ....................... 62

6-4 Experimental measurement of the expansion of vapor to temperatures below those
at which absorption-condensation is taking place.................................................65

6-5 Expected equilibrium exhaust qualities for conditions similar to those of the
experim mental study ...................... .. ........................ .... ...... .... ........... 68

6-6 Temperature-enthalpy diagram covering the phase change of pure ammonia and a
high concentration ammonia-water mixture. ................................... ..................... 69

6-7 Comparison between the measured no-load power consumption of operation with
com pressed air and am m onia-w ater .............................................. ............... 70

7-1 Maximum effective COP values where the work component is the amount of work
lost due to operation with rectification vs. equivalent conditions with no
rectific atio n .................................................... ................ 7 4

7-2 Maximum overall effective COP values as defined by Equation 7-2 ....................76

7-3 Corresponding exhaust temperatures for the optimum conditions presented in
F figure 7-2 ...........................................................................76

7-4 Design point map showing the relative sensitivity of overall effective COP to vapor
mass flow fraction and exhaust temperature... ................... ......................... 78

7-5 Effect of the minimum and maximum bounds of rectifier operation on effective
C O P values ............................................. ............................ 80

7-6 Computed effect of weak solution storage on basic solution concentration............82

7-7 Computed absorption pressures taking into account the changes of basic solution
concentration compared with measured absorber pressures. ..................................83

7-8 Measured drop in boiling pressure due to rectifier operation. ...............................84

7-9 Amount of the produced vapor that was condensed in the rectifier. This data
corresponds to the results of Figure 7-8...... ....................... ............85

B-l Schematic used for the theoretical modeling. .............................. ......... ...... .107

C-1 View of the assembled rear housing. ............... .... ..... .................. 138

C-2 Exploded view of the rear housing assembly.................................... ............... 139

D- Setup schematic used for the air testing. ...................................... ............... 141









D-2 Rear view of expander with cover rem oved........................................................ 141

D-3 Photograph of generator loading arrangement. ...................................................... 142

D-4 Photograph of gearbox mounted on expander spindle....................................143

D-5 Air testing results comparing the value of power that was computed by the thermal-
based and torque-based measurements. ...................................... ............... 144

D-6 Comparison of the difference between the power measurements of Figure D-5 and
the no-load power measurements, ........................................... ...............145















NOMENCLATURE

A area under mV-time curve

COP coefficient of performance

Cp constant pressure specific heat

D diameter, [m]

dspec specific diameter parameter

G Gibbs free energy

h enthalpy

m, m mass flow rate, [kg/s]

nspec specific speed parameter

ORC organic Rankine cycle

OTEC ocean thermal energy conversion

P pressure, [MPa]

P, pressure ratio

q heat transfer, [kJ/kg]

Q heat transfer, [W], volume flow

Qact actual volume flow rate

Qexit expander exit volume flow rate, [m3/s]

QOnd indicated volume flow rate

R universal gas constant

Rc w ratio of cooling to work











s

T

W

x

y

Ahideal

Subscripts

1stLaw

a

absorber

actual

B

basic new

basic org

boiler

bubble

cal

cool

crit-water

dew

effective

exit

expander


entropy

temperature, [C]

work output/input, [W]

ammonia mass concentration [kg/kg], mixture quality

fraction of working fluid that is stored in separator

isentropic enthalpy change



first law formulation

ammonia component properties

absorber parameter

actual parameter values

reference properties

new basic solution parameter with storage

original basic solution parameter before storage

boiler parameter

saturated liquid properties

calibration parameter values

cooling heat exchanger parameter

water critical point properties

saturated vapor properties

effective value formulation

expander exit condition

expander parameter










float

inlet

m

no rect

overall

pump

r

recovery

s

superheat

v

vr

w

wb

w/cool

wr

with rect

workopt

0

Superscripts

E

g

I

mix


float properties

expander inlet condition

mixture properties

conditions with no rectifier operation

overall comparison, with cooling production and work optimized

pump parameter

reduced properties

recovery heat exchanger properties

strong solution, isentropic end state

superheater parameter

vapor

rectified vapor

weak solution, water component properties

weak solution from boiler

conditions with cooling production

weak solution from rectifier

conditions with rectifier in operation

conditions optimized for work output

reference state properties



excess properties

gas phase properties

liquid phase properties

parameters of mixing










Greek

q efficiency

v specific volume

p density

co angular velocity, [rad/s]















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

STUDY OF COOLING PRODUCTION WITH A COMBINED POWER AND
COOLING THERMODYNAMIC CYCLE

By

Christopher Martin

December 2004

Chair: D. Y. Goswami
Major Department: Mechanical and Aerospace Engineering

This work is an investigation of a novel concept to produce power and cooling with

the energy contained in low-temperature (< 2000 C), thermal resources. These resources

can be obtained from non-concentrating solar thermal energy, low-grade geothermal

resources, and a near infinite variety of waste heat sources. The concept under

investigation uses thermal energy in a low-temperature boiler to partially boil an

ammonia-water working fluid mixture. This produces an ammonia rich vapor that drives

an expander. The expander's output is mechanical power; however, under certain

operating conditions its exhaust can be cold enough to use for cooling. This possibility is

the focus of the present study.

An analytical study is presented which identifies expander efficiency, expander

inlet conditions, and exhaust pressure as the factors determining exhaust temperature.

Estimated expander efficiencies are based on a consideration of the operating conditions

and a review of current technology. Preferred inlet conditions are identified; however,









they are linked to the overall operation of the cycle, as is absorption pressure. An optimal

balance between vapor generation and expander exhaust temperature is found for cooling

production.

Purifying the vapor is shown to enhance cooling production, but it penalizes work

output. A new coefficient of performance is defined as the ratio of the cooling gained to

the work output lost and is used to determine the optimal purification. Additionally,

another performance coefficient is defined and used to judge the overall value of cooling

produced.

An experimental study is presented that verifies the predicted trends. Furthermore,

a measurement of sub-ambient exhaust temperatures is provided that demonstrates the

key concept of this cycle. It is concluded that with improved expander performance,

practical power and cooling production can be achieved with this concept. Deviations

between measured and simulated performance are discussed as they relate to improving

future modeling and system design efforts.


xviii














CHAPTER 1
INTRODUCTION

The conversion of thermal energy into mechanical work is a fundamental task of

mechanical engineering. Performing it cleanly, cheaply and efficiently all influence the

eventual conversion scheme. Rankine-based cycles enjoy widespread usage and are

particularly suited for low resource temperatures since their operation can approximate

that of a Carnot engine.

Many adaptations and modifications have been made to the basic Rankine cycle in

order to extract the most energy from heat sources such as geothermal wells, solar

thermal energy, and waste heat streams. A relatively recent cycle has been proposed in

which thermal energy is used to produce work and to generate a sub-ambient temperature

stream that is suitable for cooling applications [1]. It has been the focus of theoretical

and experimental investigation [2-5]; however, until this work, there has not been a

complete, experimental implementation of this power-cooling cycle. Therefore, this

study is an investigation, both theoretical and experimental, into the distinguishing

feature of this concept, which is cooling production.

The cycle is a combination of Rankine power production and absorption

refrigeration cycles, and is unique in power production cycles because it exploits the

temperature drop across an expander to the point of being able to obtain useful cooling.

Optimization of system parameters, working fluid selection, and preliminary

experimentation with the cycle have been performed [2-5]. What this work provides is

an experimental proof-of-concept that demonstrates the key feature of this cycle. Also









included is a discussion of the parameters affecting cooling production and a method to

quantify its production.

Certainly this work advances the development of this power-cooling cycle, but

more importantly, research in the field of heat recovery is an active step in moving

society toward a sustainable energy policy. This concept belongs to the broader class of

low temperature, Rankine based cycles which have been shown to be one of the most

effective means for utilizing low temperature resources. They have been applied to the

production of mechanical power using heat from solar, geothermal, and waste heat from

topping power cycles and industrial processes. Despite their wide range of possible

applications, these systems have found limited success in practice. It is hoped that this

work will aid any resurgence in today's energy market.

Motivation

The wide-ranging motivation for this work comes from the possible applications

for this category of low temperature, Rankine based cycles. Being simply a heat engine

with the potential for good second law efficiencies, the possible recovery applications are

limited only by the economics of the situation. In the future, the economics may be more

favorable to devices that can produce power without additional resource consumption.

When considering the future of world electricity production, the only apparent

certainty is that generation will be done by more diverse means than it is currently [6]. It

appears that the paradigm of a few, large, centralized power producers is becoming more

conducive to adding more, smaller, distributed generators. There are many reasons for

this; key among them are to increase the reliability of the electrical system by promoting

diversity, provide cleaner energy by incorporating more renewables, and simply to

increase capacity to meet additional demand.









The distributed generation trend will open opportunities in two key ways. First, by

adding smaller distributed generators, the mechanics of connecting to the grid will no

longer be prohibitively complex, but will become more routine. Second, with on-site

generation the opportunities to recover and use waste heat resources will make economic

sense. In fact, the U. S. Department of Energy expects the utilization of waste heat alone

to provide a significant source of pollution-free energy in the coming decades [7].

Viewed in this way, the use of a low-temperature power cycle is one of the many

possible distributed technologies that could connect to the grid or be used to recover

thermal resources. They can be used on a small scale to convert renewable energy

sources, use conventional fuels efficiently, or conserve energy by recovering waste heat

from energy-intensive processes, Figure 1-1. Ultimately they would have positive

impacts on overall energy conversion efficiency and could be used to incorporate

renewable energy sources.




Topping
Cycle

Solar
Thermal
I Waste Heat


Power
Conventional Adaptable
Fuel : ORC
Combustion


Figure 1-1. Ideas for using an ORC to incorporate a renewable element into distributed
power generation. Efficient use of multiple energy sources would require a
highly adaptable heat engine and, of course, any configuration would have to
be economically viable.









Power-Cooling Concept

This work is not directly aimed at reducing the cost of this technology. Rather it is

directed at improving the underlying science to make it more versatile and thus more

attractive for implementation. Mechanical power is one useful form of energy, the

generation of low temperatures for cooling or refrigeration is another. The cycle under

study in this work was intended to explore the feasibility of using thermal resources to

simultaneously produce these two useful outputs.

Put simply, the configuration of this power-cooling cycle allows the vapor passing

through the turbine to be expanded to below ambient temperatures. Cooling can then be

obtained by sensible heating of the turbine exhaust. A more detailed explanation of this

process follows in Chapter 2, but here it suffices to say that the use of a working fluid

mixture, ammonia-water, is the key to this process. Just as in conventional aqua-

ammonia absorption cooling, absorption-condensation is also used here to regenerate the

working fluid. This eliminates the expansion temperature restriction which is in place

when pure condensation is used.

The power-cooling cycle has the obvious advantage of two useful outputs, but it

has other attributes that make it an attractive energy conversion option. The first of these

characteristics is that the cycle uses a binary working fluid that has a variable boiling

temperature at constant pressure. This avoids heat exchange "pinch point" problems that

pure component working fluids experience due to their constant phase change

temperature at constant pressure. In addition, turbine designs for ammonia-water are

reasonably sized for large power outputs when compared to the more traditional organic

working fluid choices.









Problem Definition

The distinguishing feature of this cycle, compared to other power cycles and even

those in the developing class of combined power and cooling cycles, is the method in

which cooling is produced. In other power cycles the working fluid is regenerated by

pure condensation, rather than absorption-condensation which is used here; this limits the

minimum turbine exhaust temperature to roughly the temperature at which condensation

is taking place. When considering other combined power-cooling cycles, cooling is

typically produced in the same manner as a conventional absorption system, that is,

condensation and throttling of the refrigerant. Here in this cycle, vapor is expanded

through a turbine to produce power and because of the advantage of absorption-

condensation, it can be expanded to sub-ambient temperatures.

While the method of cooling production is the key feature of this cycle, until this

work it has not been experimentally investigated. What has been experimentally

investigated are the underlying boiling and absorptions processes [4]. For those

experiments a turbine was not implemented; its performance was simulated with an

expansion valve and a heat exchanger. Coupled with the lack of experimentation, the

question of implementing cooling production with this concept has not been treated in

any depth.

Discussion of the possible uses of this cycle have suggested the utilization of solar

thermal, geothermal, or waste heat resources. However, a proper use for the potential

cooling output has not been put forward, possibly because the specific nature of an

application will be determined by the characteristics of cooling production. There has not

been a thorough discussion of the trends of cooling production.









Research Objectives

In response to the deficits mentioned in the previous section, the objectives of this

work are to experimentally implement a turbine for power production and identify the

factors important for cooling production and investigate them analytically and

experimentally. Analytically, the study will identify the conditions favorable for cooling

production, estimate performance using available expander technologies, and quantify

cooling production in terms of energy consumption. In addition, this work will

experimentally investigate the concepts key for cooling production and document design

and operating experience for use with future modeling or implementation efforts.














CHAPTER 2
BACKGROUND AND REVIEW

The purpose of this chapter is to introduce the concept of this cycle in the context

of both low temperature thermodynamic power cycles and conventional cooling cycles.

In addition, to accurately provide the context for this work, a review of previous effort

into this concept is presented.

As an overview, the power-cooling cycle is best described as a compromise

between a conventional aqua-ammonia absorption system and a Kalina-type power

generation cycle. It is a continuation of the evolution of binary mixture Rankine cycles

but makes use of the cooling effect possible due to the working fluid concentration

change. As for previous work on this cycle, numerous theoretical studies have been

produced and initial experimentation has begun. From a review of that work, this study

is shown to be the first experimental confirmation of the power-cooling cycle's key

concept and to provide initial consideration for system operation.

Background

The thermodynamic conversion of low temperature resources into mechanical

power traces its roots to at least the beginning of the industrial revolution. Utilizing solar

thermal energy to pump water was the impetus and this work continued haphazardly until

the early decades of the twentieth century when it was interrupted by World War I and

the discovery of a new resource, oil and gas [8]. Modern research into low temperature

power conversion surfaced again when the panacea of cheap coal power was beginning to

break down and energy alternatives were sought in the decades following World War II.









The application was utilizing liquid-producing, geothermal fields where flash boiling is

not suitable [9]. Additional interest came during the 1970's oil crisis in using these low

temperature engines for solar thermal energy and heat recovery applications. The

common description for these systems is organic Rankine cycle (ORC) engines, because

many of the working fluids are organic hydrocarbons or refrigerants.

ORC Development

Intense research of non-geothermal ORC use took place in this country during the

early 1970's through the early 1980's. ORC heat engines were reconsidered for utilizing

solar resources and conserving other resources by recovering energy from waste heat.

Seemingly no application was overlooked as a few innovative examples illustrate. In one

an ORC was integrated with a large truck engine to recover heat from the exhaust and

save on fuel costs [10] and in another application the idea of replacing the automobile

internal combustion engine with an ORC system was explored [11].

Mechanical cooling systems were one of the more productive research areas that

dealt with the conversion of solar thermal energy. A significant amount of the published

literature regarding ORC conversion of solar thermal energy comes from this and related

work [12]. The concept started as an alternative to solar-driven, absorption, air-

conditioning cycles which have a limited coefficient of performance. Essentially,

mechanical work produced by a solar-driven ORC would be used to drive vapor-

compression air conditioning equipment, with the potential of a higher COP than

absorption equipment [12]. These projects produced many successful prototype units

( e.g. [13] ) and led to a feeling of technical maturity for the low-temperature, small-

scale, conversion of solar thermal energy [14]. As for ORC technology today, it has

found some niche successes in geothermal utilization, biomass utilization, some industrial









heat recovery, and cathodic protection of pipelines, as judged by a few manufacturer's

portfolios.

More recent research in the area has largely taken place internationally ( e.g. [15-

18] ), with much interest being placed on expander implementation. Two approaches

have been noted: one is to develop and design systems around high-speed

turbomachinery with a shaft integral generator and circulation pump [19], thus reducing

costs by simpler design, and the other, more recent idea is to adapt mass-produced

(cheap) displacement compressors for use as reasonably efficient expanders [18, 20, 21].

Ammonia-Water Cycles

While much of the related material on ORC systems is intended for small scale

application or implementation (especially solar-driven units), the lineage that the power-

cooling cycle is derived from was initially intended for utility-scale bottoming cycle duty.

The first study of an absorption based power cycle was performed by Maloney and

Robertson [22] who concluded no significant advantage to the configuration. Several

decades later, Kalina [23] reintroduced the idea of an ammonia-water power cycle as a

superior bottoming cycle option over steam Rankine cycles. Some independent studies

have been performed [24, 25] that concede some advantage of the Kalina cycle under

certain conditions.

The key advantage of the ammonia-water working fluid is its boiling temperature

glide, which allows a better thermal match with sensible heat sources and reduces heat

transfer related irreversibilities. This same advantage, however, could be a problem

during the condensation phase of the cycle in which the condensation temperature glide

could cause a thermal mismatch with the heat rejection fluid and an increase in heat

transfer irreversibility. The solution employed is to vary the concentration of the working









fluid so that the fluid passing through the turbine is of different composition than that

being condensed in the condenser. In fact, by taking advantage of the chemical affinity

of ammonia and water, the condensation process can be replaced by absorption-

condensation.

Investigation of these power cycle configurations has become a new specialty in

engineering thermodynamics, and the power-cooling cycle of this work is a product of

this research area. As a result of this diversified interest, ammonia-water based power

cycles have been proposed for solar utilization, geothermal, ocean thermal energy

conversion, and other forms of heat recovery.

Power-Cooling Concept

While it was the interest brought about by Kalina's proposal that led to the

introduction of the power-cooling cycle, it is somewhat ironic that the original suggestion

for its implementation is more similar to the original Maloney-Robertson implementation

[1, 2]. Figure 2-1 is a schematic of the power-cooling cycle. Aside from the operating

parameters, the key difference between the cycle of Figure 2-1 and the Maloney-

Robertson cycle is the addition of a vapor rectifier following the boiler.

Referring to Figure 2-1, basic solution fluid is drawn from the absorber and

pumped to high pressure via the solution pump. Before entering the boiler, the basic

solution recovers heat from the returning weak solution in the recovery heat exchanger.

In the boiler, the basic solution is partially boiled to produce a two-phase mixture; a

liquid, which is relatively weak in ammonia, and a vapor with a high concentration of

ammonia. This two-phase mixture is separated and the weak liquid is throttled back to

the absorber. The vapor's ammonia concentration is increased by cooling and condensate

separation in the rectifier. Heat can be added in the superheater as the vapor proceeds to









the expander, where energy is extracted from the high-pressure vapor as it is throttled to

the system low-pressure. The vapor rejoins the weak liquid in the absorber where, with

heat rejection, the basic solution is regenerated.

Heat Out Heat In


Figure 2-1. Schematic of the power-cooling cycle.

In this configuration, the vapor temperature exiting the expander can be

significantly below ambient conditions and cooling can be obtained by sensibly heating

the expander exhaust. The temperature drop possible across the expander is due to the

fact that the working fluid is a binary mixture, and at constant pressure the condensing

temperature of an ammonia rich vapor can be below the saturation temperature for a

lower concentration liquid. This is best illustrated with a binary mixture, phase

equilibrium diagram, as shown in Figure 2-2. The low concentration saturated liquid







12


state represents the basic solution exiting the absorber, while the high concentration

vapor is typical of the expander exhaust conditions. This shows how it is possible for the

vapor to be expanded to a temperature below that at which absorption is taking place.

According to the equilibrium diagram, to maximize this temperature difference the basic

solution should be low in ammonia concentration and the vapor should be high. Also,

partial condensation of the expander exhaust would cause an additional decrease in vapor

temperature. This is entirely possible since ammonia becomes saturated upon expansion.



120 Vapor Pressure = 0.203 MPa


100
Two-Phase
80

Liquid
3 60

E 40 Basic solution
Sin absorber

20 Expander
exhaust
0


-20
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Ammonia Mass Fraction

Figure 2-2. Ammonia-water phase equilibrium diagram highlighting the source of
cooling temperatures.

Prior Work

Since the proposal of the idea by Goswami a theoretical and experimental

investigation has been under way by a group at the University of Florida. Initial

investigations were performed theoretically and they focused on procuring reliable









property data for the ammonia-water mixture [26] and identifying operating trends [27,

28]. Later studies concluded that the cycle could be optimized for work or cooling

outputs and even for efficiency. Optimization studies began to appear, optimizing on the

basis of various efficiency definitions, minimum cooling temperature, working fluid

combination, and system configuration. Also, an experimental study was described by

Tamm and Goswami [4] which generally verified the expected boiling and absorption

processes.

Goswami and Xu [27] presented the first theoretical analysis of the power-cooling

cycle. Turbine inlet temperatures of 400 500 K were considered along with absorption

temperatures of 280 320 K. Cooling production suffered with increased turbine inlet

and absorption temperatures, and benefited with increased boiler pressure. Many of the

operating trends of importance in this work were introduced here.

Optimization studies began to appear following this work, which identified the

balance of effects that dictate cycle operation. Lu and Goswami [2] optimized the ideal

cycle conditions using various objectives, work output, cooling output, first and second

law efficiencies. All operating parameters, efficiencies, power/cooling output, etc., were

found to decrease with increasing heat rejection temperatures. At high heat source

temperatures, 440 K, no cooling was possible at conditions optimized for second law

efficiency. A contrast between work optimized and cooling optimized cases was

provided. Important differences in the cooling optimized case versus the work optimized

one were higher vapor concentration, lower turbine inlet temperature, low vaporization

fraction (16.5 % vs. 91.2 %), and a lower basic solution concentration. Minimum cooling









temperatures were also optimized [29], and a minimum turbine exhaust temperature of

205 K was identified under the assumptions considered.

The question of appropriate efficiency expressions for the cycle was tackled by

Vijayaraghavan and Goswami [30]. The conditions obtained from an optimization study

were found to be heavily influenced by the weight given to the cooling output. Some

expressions simply added the outputs of power and cooling, which gives an overestimate

of system performance, or cooling was weighted by an ideal COP value computed for

equivalent temperature limits, which tends to underestimate the value of cooling. They

[30] introduced a satisfactory second law efficiency definition based upon ideal Lorenz

cycle performance which accounts for sensible heat addition and rejection behavior.

However, they concede that ultimately the value of work and cooling will be decided by

the end application [30].

Both first and second law efficiency analyses were performed for the cycle [31,

32]. A second law efficiency of 65.8 % was determined, using the definition of [30], for

the idealized model considered. The largest source of irreversibility was found to be the

absorber at all conditions considered; while at higher heat source temperatures the

rectifier also contributed significantly.

Less-than-ideal modeling began with Tamm et al. [33, 34], in preparation for the

initial experimental studies [4]. The largest deviation from idealized simulations was due

to the non-isentropic performance of the turbine. This relates well to the findings of Badr

et al. [35], who identified the expander isentropic efficiency as the single-most influential

factor affecting overall ORC engine performance. Initial experimentation was reported

[36]; however, turbine operation was simulated by an expansion valve and a heat









exchanger. General boiling condition trends were demonstrated, for example vapor mass

flow fraction, vapor concentration, and boiler heat transfer. Vapor production was less

than expected and improvements to the setup were identified and implemented.

Performance of the new configuration, still having a simulated turbine, was also reported

[4]. Vapor production and absorption processes were shown to work experimentally,

however still with some deviations.

An independent study of the power-cooling concept has been provided by Vidal et

al. [37], who also noted the significant impact of non-ideal turbine performance on

cooling production. Vidal et al. also reported poor cooling production at higher ambient

conditions.

Other Power-Cooling Concepts

The development of the power-cooling cycle under investigation in this work has

been presented as it relates to other power production cycles. However, there is now a

small class of combined power and cooling cycles, especially since the proposal by

Goswami [1]. Differentiation of this concept from others in the literature is now

presented.

Oliveira et al. [38] presented experimental performance of an ORC-based,

combined power-cooling system which used an ejector placed in-parallel to the turbine

for cooling production. Ejector cooling has been an academic topic for solar thermal-

powered cooling, for example [39, 40]. The implementation and operation of an ejector

cooling system is quite simple and rugged; however, its COP tends to be low and in this

combined case it siphons away high pressure vapor directly from the turbine that could

have been used to produce power.









Considering integrated, ammonia-water cycles, Erickson et al. [41] present the

most intuitive. The proposal is essentially an absorption cycle, with advanced thermal

coupling between the absorber and generator, with a turbine placed in-parallel to the

condenser and evaporator. So vapor is produced and, depending on the outputs desired,

split between expansion in a turbine or condensation and throttling in the refrigeration

path. Integration comes from the common components, for example the absorber,

generator, and feed pumps. However, the mechanism of cooling is the same as that for

an aqua-ammonia absorption system. The very pure ammonia vapor is condensed at high

pressure and throttled to the absorption pressure where flash boiling and evaporation take

place.

The concept of parallel paths for power and refrigeration production has been

incorporated with the thermal-matching concepts of a Kalina cycle by a research group at

Waseda University [42]. Unlike the proposal by Erickson et al., however, only the

working fluid is shared between the two systems. The power production and

refrigeration cycles can be driven independently, but it was found that more power could

be produced by sharing the working fluid [42]. Therefore, cooling in this case is also

produced in the same manner as with an aqua-ammonia absorption system.

A more thorough integration of power and refrigeration production has been

recently proposed by Zhang et al. [43]. In this configuration the ammonia-water basic

solution is separated into a high concentration ammonia vapor and a relatively weak

solution liquid in a device similar in operation to a distillation column. The vapor is

condensed and throttled to produce cooling while the weak solution liquid is vaporized

and superheated, then expanded in a turbine for power production. The streams are then









cooled and rejoined in an absorber. The authors claim a 28% increase in exergy

efficiency of this arrangement over separate steam Rankine and aqua-ammonia

absorption systems [43]. As with the other concepts, cooling is produced the same way

as with a aqua-ammonia absorption system.

Conclusion

As compared to other power and cooling concepts, the distinguishing feature of this

cycle is the method in which cooling is produced. Its configuration is most similar to that

of an aqua-ammonia absorption system; however, instead of using condensation and

throttling for cooling production an expander is used to extract energy from the vapor--to

the point that cooling can be obtained from the exhaust. As compared to the absorption

cycle, the trade-off for work production is reduced cooling since no latent heat is

involved. The remainder of this work will discuss the opposite situation, the penalty to

power cycle operation due to combined cooling production.














CHAPTER 3
THEORETICAL STUDY

In this chapter a model of the system is presented and used to simulate the steady

state performance of the power-cooling cycle. With this model, a straightforward

parametric study is carried out which identifies the essential operating mechanisms

affecting cooling production. These results are used to design the experiments discussed

in Chapters 5 and 6 and again to extrapolate the data used for the final conclusions.

Model

The model used for this work is based upon the schematic of Figure 3-1 which has

subtle differences from the one in Figure 2-1 to be more representative of the

experimental system. Table 3-1 contains the identifying information for the working

fluid streams in Figure 3-1.

Table 3-1. Flow identification for the configuration of Figure 3-1.
Identifier/
Subscript Description
s Basic (strong) solution flow from absorber through boiler
v Vapor flow produced from partial vaporization in boiler
vr Rectified vapor passing through turbine and cooling heat exchanger
w Weak (in ammonia) solution liquid returning to absorber
wr Weak condensate formed in rectifier
wb Weak liquid produced from partial vaporization in boiler

For the purposes intended, it was adequate to use first order approximations for

each component, conservation of mass and energy, so detailed component modeling was

not included. The complete formulations that were used in the computations, as well as

the subroutines themselves, can be found in Appendix B; however, the key points are

summarized as follows.

































Solution I
Pump
Figure 3-1. Power-cooling schematic used for modeling.

* The boiling conditions are completely specified, i.e. boiling temperature, pressure,
and basic solution concentration are provided as inputs. This means that the quality
at boiler exit is allowed to be determined in accordance.

* The system low pressure is dictated by the basic solution concentration and the
minimum absorption temperature, both of which are specified.

* Isentropic efficiencies are assumed for the pump and expander while effectiveness
values are used for heat exchangers.

* The degree of rectification is determined by specifying the rectifier exit
temperature. Similarly, the superheater exit temperature is also specified.

In addition to the specifications above, which are needed to determine the steady

state conditions, the following stipulations were enforced to avoid computational

problems and/or make the scenarios closer to reality.

* The minimum absorption temperature considered was 250 C with most attention
given to 350 C cases.









* Vapor rectification was limited by either the specified rectifier exit temperature or
an ammonia mass fraction of 0.999, whichever was encountered first. The
minimum rectification temperature considered was 350 C.

* The minimum amount of vapor leaving the rectifier that was allowed was 5 % of
the basic solution flow rate.

* The quantity of cooling produced (if any) was calculated as the energy needed to
heat the expander exhaust from the exhaust temperature to 150 C.

Thermodynamic property data for the ammonia-water working fluid is essential for

this type of modeling. The correlations used are based on those presented by Xu and

Goswami [26] which are a combination of the Gibbs free energy method for mixture

properties and empirical equations of bubble and dew point temperatures for phase

equilibrium. Details of the complete correlations and their implementation into C++ can

be found in Appendix A.

Operating Mechanisms

Early in the theoretical investigation of this cycle it was determined that the system

could be optimized for various outputs [27]. In this section, simulated data is used to

illustrate these optimums and the balance of effects that causes them. Boiling conditions

are considered first and then the effects of heat rejection conditions.

First consider the independent effects of the parameters at the heart of the power-

cooling cycle, the boiling conditions. These effects are common to all binary mixture

power cycles; however, it will be shown that they have added significance for cooling

production.

Effect of Boiling Pressure

The boiling pressure in the power cycle is regulated by the rate of vapor production

and the rate at which vapor is released through the restriction imposed by the expander.

For a binary mixture working fluid at constant temperature and having constant










composition boiling takes place over a range of pressures from the saturated liquid state

to the saturated vapor state. At the upper extreme boiling pressure is limited by the

corresponding saturation pressure, above which no vapor is produced. The lower

pressure extreme is bound by the system low pressure or the absorption-condensation

pressure. Depending on the conditions, the working fluid may or may not be fully

vaporized at the lower pressure extreme.

Figure 3-2 is a graphical representation of the mechanisms of variable pressure

boiling. As can be seen the mass flow rate of vapor changes inversely with pressure

ratio. Also, a quantity like the work output (assuming constant efficiency expander),

which is dependent on both the pressure ratio and amount of vapor flow, contains a

maximum within the boiling region. Work production is bounded by a unity pressure

ratio at low boiling pressure and zero vapor flow at the highest boiling pressure.




Vapor Mass
Flow

Pratio





E Absorption P Boiler P too
higher than high, no
boiler P vaporization


T Work
Output


Pabsorber Pbubble
Boiling Pressure [MPa]

Figure 3-2. Conceptual relationship between the factors affected by boiling pressure.










As a further example, Figure 3-3 presents computed results also for the variation of

boiling pressure. The relative position of the maxima for work production, first law

efficiency, and cooling production are shown. Similar to work output, cooling also has a

maximum, which is limited at low pressures by higher turbine exhaust temperatures and

bounded at higher pressures by the low production of vapor.


25 8

Work Output 1st Law Efficiency
S7
20



S15/ / \ \
,--j


0 4
0) / / -- \ (D
1 Cooling Output o
10 3
o 3 / / \


2
5

Boiler Temperature: 80 C
Strong Concentration: 40%
0 0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Boiling Pressure [MPa]

Figure 3-3. Output parameter variation as a function of boiling pressure.

Effect of Mixture Concentration

Mixture concentration directly changes the saturation temperatures and temperature

glide of the working fluid. Similar to the boiling pressure, it has upper and lower bounds

which are illustrated in Figure 3-4. The lowest possible concentration corresponds to the

saturation concentration for the given temperature and pressure. Concentrations lower

than this will not boil at the specified temperature. Also, the absorption pressure is

determined in part by the basic concentration, therefore, the upper limit of concentration










is reached when the corresponding absorber saturation pressure equals the boiling

pressure. Similar to that with pressure variation, work production has a maximum which

is governed by the extremes of zero vapor production and a pressure ratio of one.


8
25
Work Output
1st Law Efficiency 7

20
,) 6

S5
o0 15 Boiler Temperature: 80 C
SBoiling Pressure: 0.7 MPa O
4
0) I \ \ \
0
S10 \ \ \ 3 W

0 Cooling Output 2

5
1

0 0
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Basic Solution Concentration

Figure 3-4. Variation of output parameters as a function of basic solution concentration.

Effect of Boiling Temperature

While the effects of basic solution concentration and boiling pressure are intended

for optimization of the system, boiling temperature is considered largely dependent on

the heat resource. However, adaptability to such changes has been identified as an

important evaluation criteria due to the inherent heat source variability associated with

solar, waste heat [44], and even geothermal resources [45]. So the effect of boiling

temperature is also considered here. Assuming that the temperature is above that needed

for vaporization, there is no convenient upper bound on boiling temperature. The effect










of interest here is shown in Figure 3-5, which is to essentially shift the trends that were

presented in Figure 3-3.


45
Work Output 14
40 Cooling Output
1st Law Eff
12
35 Basic Concentration: 40%

30 -10
0)
25
\3 -8

o 6
10 /





80C / 2
15 \
4
10

sS 100 C
S 60 C Boiling T
0 0
0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5
Boiling Pressure [MPa]

Figure 3-5. Effect of boiler exit temperature on output parameter profiles.

Cooling Production

As explained in Chapter 2, the difference in concentration between the vapor and

basic solution streams allows the expander exhaust to be below ambient temperatures.

Sensible heating of this exhaust provides the combined cooling output. Therefore, when

considering cooling production the exhaust temperature and quantity of vapor both factor

into the total amount. Initially, the sensitivity of each parameter is examined individually

then total cooling production is discussed.

Exhaust Temperature

Some comments regarding minimizing the expander exhaust temperature were

discussed in relation to the binary phase diagram of Chapter 2, Figure 2-2. A reiteration









of those conclusions is that the basic solution should be low in ammonia concentration

and the vapor should be high. The total absorption pressure should be low. Also, partial

condensation of the expander exhaust would cause an additional decrease in vapor

temperature. To investigate the effect of expander inlet vapor conditions on exit

temperatures, consider the entropy of the working fluid at the expander exit.

Minimization of the exhaust temperature also implies a minimization of the vapor

entropy at expander exhaust, assuming constant exit pressure. From this consideration an

efficient expander is an obvious feature for low temperatures, but even an ideal device

would only maintain the vapor entropy from inlet to exhaust. Therefore, expander inlet

conditions should be considered.

For the binary vapor mixture of ammonia-water, entropy decreases with increasing

pressure, increasing concentration, and decreasing temperature. The limit of these

conditions, while still maintaining vapor, would be saturated, pure ammonia. As an

aside, aqua-ammonia absorption cooling cycles further reduce the throttle inlet entropy

by condensing the vapor to a liquid.

Figures 3-6 and 3-7 show the effects of inlet temperature, vapor composition, and

expander efficiency on exhaust temperature. The intuitive effect of decreasing inlet

temperature is shown in both figures. Figure 3-6 highlights the sensitivity to vapor

composition, while Figure 3-7 reiterates the benefit of good expander efficiency. The

effect of exhaust pressure has not been shown, however, lower temperatures are

encountered with lower exhaust pressures.











15

Pmlet/Pexit = 0.516 MPa/0.208 MPa
expander = 70%

10
J Xvr = 0.995

E

5 0.997
wo

X

0




-5
37 39 41 43 45 47 49 51 53 55
Expander Inlet Temp. [C]

Figure 3-6. Computed effect of vapor concentration and inlet temperature on expander
exhaust temperature.




35 Pniet/Pexit = 0.516 MPa/0.208 MPa
Xvr= 0.995 expander =

30 %


5 70 %
S15
W 70%
o


100 % -


37 39 41 43 45 47 49 51 53 55
Expander Inlet Temp. [C]

Figure 3-7. Computed effect of expander efficiency and inlet temperature on expander
exhaust temperature.









Relating this to the variations in boiling conditions mentioned previously indicates

conditions preferable for cool exhaust temperatures. The conditions closest to low

temperature, high pressure, and high concentration ammonia vapor occur at the limit

where boiling is just beginning. Then assuming expander efficiency and exhaust pressure

remain constant, the lowest exhaust temperatures will occur at this leading edge and will

increase as the fraction of vaporization of the basic solution increases.

Vapor Flow Rate

Vapor flow rate was discussed as it was affected by boiling and concentration

changes. It was shown to vary from a maximum of 100% of the basic solution flow rate

at low pressures and high concentrations to a minimum where boiling just starts at high

pressures and low concentrations. From the standpoint of cooling production, much like

that for work production, the two determining factors vary inversely with each other as a

function of pressure. This creates a maximum for cooling production as shown in

Figures 3-3 through 3-5.

Rectification

This section discusses conditioning the vapor stream before it enters the expander

to enhance cooling production. This conditioning involves cooling and condensate

separation to increase the vapor concentration. It is commonly employed in aqua-

ammonia absorption refrigeration systems to prevent water build-up in the evaporator and

it is termed rectification [46]. In the power-cooling cycle its purpose is to change the

temperature and concentration of vapor to values more suitable for cool exhaust.

For sufficiently low absorption pressures, rectification may not be necessary for

cooling production; however, this typically requires unreasonably cool heat rejection










(absorption) temperatures. For the temperatures considered in this work, 250 and 350 C,

some rectification is typically needed to reach exhaust temperatures below 150 C.

The effect of rectification on expander exhaust temperatures is twofold. First it

lowers the expander inlet temperature of the vapor, which beneficially lowers exhaust

temperatures. Second it increases the vapor concentration so that expansion can take

place to lower temperatures before the dew point is encountered. The improvements to

vapor concentration and expander exit temperature with rectifier exit temperature are

shown in Figure 3-8.


80 1 -

70 /T"*\- 0.98
Concentration 0.9
60)

E 50 PB/PA = 0.5/0.2 MPa 0.94
S m,r/m = 0.392
0.92 a
5 40 expander = 30 %

30
S0.88 0
a 20
\ -0.86 C.
w1 10 ------7-----------\----- 0.84
10
0.84
Exhaust Temp. 0
0 0.82

-10 0.8
0 20 40 60 80 100 120
Rectifier Exit Temp. [C]

Figure 3-8. Beneficial effect on expander exhaust temperature as a function of increasing
rectification (decreasing rectifier exit temperature).

A word regarding the operation of the rectifier is appropriate here. There are many

physical setups that can be used to purify the vapor, with some being more efficient, in

terms of purified vapor flow, than others. For this work, upper and lower limits to










rectifier efficiency are considered. The upper bound is the theoretical maximum of

rectified vapor that can be produced as determined from a mass balance of the rectifier.

This could be implemented with a direct contact, counter-flow heat exchanger where

additional ammonia is scavenged from the counter-flowing condensate. The lower

bound, which represents the arrangement of the experimental setup and the computer

model, is the flow rate that occurs with simple cooling of the vapor and condensate

separation. No attempt to recover ammonia from the condensate is made. The effects of

this vapor-production efficiency are considered next.

The improvements to cooling from rectification do not come without cost,

however, and in cases where it is used, some potential work is sacrificed for cooling.

Figure 3-9 shows an example of this sacrifice, which is a plot of normalized work output

versus rectifier exit temperature.




0.9

0.8 Maximum Vapor
Production
0.7
0
S0.6

o 0.5 PB/PA = 0.5/0.2 MPa
S/ m,/ms = 0.3
| 0.4 Texpander = 30 %
0 Minimum Vapor
Z Production
0.3

0.2
0 20 40 60 80 100 120
Rectifier Exit Temp. [C]

Figure 3-9. Effect of rectification and rectifier efficiency on work production.









The penalty to work production is caused by two factors, the decline in available

energy from cooling and the reduction in mass flow rate due to condensate formation and

separation. The combined effects of both items on normalized work output is shown in

Figure 3-9 for the upper and lower production rates for the rectifier.

As evident from Figure 3-9, at minimal amounts of rectification the difference

between rectifier performance is small. However, with increasing amounts of

rectification the difference is severe and it appears that investment in a more efficient

device may be warranted.

Performance Measures

This section describes the efficiency parameters used to evaluate the relative

performance of the power-cooling system.

Work Production

A measure of performance is needed to compare the relative efficiency of

producing work with equivalent heat source conditions and cycle configurations, and also

to identify conditions for maximum work production for a given set of heat source/sink

conditions. For this purpose, a first law efficiency formulation is adequate, Equation 3-1.

(JWp W
stLaw (expander pump (3-1)
( Qboiler + Qsuperheat )

Cooling Production

There is some element of personal choice involved in including cooling in an

efficiency definition. This comes from the options of converting cooling to equivalent

work terms. Other work has discussed the merits of adding work and cooling directly or

weighting cooling with a COP based on ideal cycle performance [30]. In this chapter it

has been shown that cooling and work optimums generally do not coincide, that is to









produce cooling some sacrifice in work has to be made. Based on this observation, an

effective COP can be defined as the ratio of the cooling produced to the work that could

have been produced, but was avoided to generate cooling. In general terms the concept

can be written as follows.

CO e Cooling Gained (3-2)
efese Work Lost

This term has been called the effective COP since cooling and work are only

indirectly related, in other words there is no device directly producing cooling with the

work that is given up. This definition is simply a way to determine the effectiveness, in

terms of energy, of cooling production with this cycle.

It was mentioned that some rectification is needed to produce any cooling with the

absorption temperatures considered. Recalling also that rectification diminishes work

production by the mechanisms of reduced mass flow and available energy, then some

work is inevitably lost when cooling is produced. Therefore, a more specific effective

COP can be defined based on the need for rectification, and is shown as Equation 3-3.

Qool
COPeve wthrcool e (3-3)
effec-tve ^ ^
(Wnoirect with rect

Conclusion

This chapter has shown the operating mechanisms affecting cycle operation. These

mechanisms determine the relative amounts of work and cooling production by affecting

the balance of vapor production, expander pressure ratio, and expander exhaust

temperature. Expander exhaust temperature is sensitive to the sensitive to the inlet vapor

conditions (pressure, temperature, and concentration), exhaust pressure, and expander

efficiency. Considering the preferred inlet conditions, the ultimate effect of the partial






32


boiling and rectification process would be to separate ammonia vapor from liquid water

rather than generate high pressure vapor for power production. The trends identified in

this chapter are used to guide the experimental study described next chapter. Also, the

implications that can be extrapolated from these trends are discussed during the

conclusion of this work.














CHAPTER 4
EXPANDER CONSIDERATIONS

As shown in the previous chapter, efficient operation of the expander is an obvious

requirement for cooling production within the power-cooling cycle. The purpose of this

chapter is to review the considerations for expander application in the power-cooling

cycle. An evaluation of the ammonia-water working fluid properties is given and

comparisons are made with other power cycle working fluids. These properties are

linked to design considerations for various machine types and data from the literature is

used to base estimations on the expected performance of expanders for this application.

Working Fluid Properties

In this section the thermophysical properties of the ammonia water working fluid

are considered as they relate to expander design. Ammonia is seen to behave more like

steam rather than the organic fluids that are typically used in low temperature Rankine

conversion systems. Expanders are treated in two groups, one being dynamic machines,

those that convert the fluid's energy to velocity and create shaft power by a momentum

transfer, and the other being displacement devices, where the working fluid is confined

and allowed to expand against a moving boundary.

Table 4-1 is a comparison of fluid properties for other power cycle working fluids

as well as ammonia-water for a hypothetical, isentropic expansion. The most significant

difference between ammonia-water and the typical ORC fluids is the large isentropic

enthalpy drop of ammonia-water. This corresponds to a higher ideal jet velocity which

has an impact on dynamic turbine design. Considering steam's characteristics, it and









ammonia-water are similar with respect to enthalpy drop and jet velocity. This is to be

expected because of the close molecular weights of both fluids.

Table 4-1. Fluid properties for a typical ammonia-water concentration and other power
cycle fluids for isentropic expansion from saturated conditions at 1000 C to
condensation/absorption at 35 C. A basic solution concentration of 0.40 was
assumed for the ammonia-water data.
Ammonia- Isobutane HCFC-123
Fluid Characteristics Water, 0.99 Steam [47] [48]
Molecular Weight 17 18 58 153
Isentropic Enthalpy Drop, kJ/kg 327.8 418.0 56.75 29.77
Ideal Jet Velocity, m/s 809.7 914.3 336.9 244.0
Volumetric Expansion Ratio 11.5 13.1 4.91 5.96
Exhaust Quality 0.837 0.873 1 1

Since work production with a turbine is a momentum transfer process, the relative

velocities of the fluid stream and rotating blades are critical design parameters and they

can be used to characterize the operation of a turbine. The velocity ratio, which is the

ratio of rotor tip tangential speed to working fluid ideal jet velocity has preferred values

for differing flow arrangements, axial, radial, etc. The preferred value of this parameter

will relate the design parameters of rotor diameter, rotor inlet area, and rotational speed to

the isentropic enthalpy drop across the turbine. For steam and ammonia, with their high

ideal jet velocities, these requirements result in a choice between extremely high rotor

speeds with small diameters or impractically small inlet flow passages with enlarged

rotor diameters. Designs for heavier organic fluids result in efficient geometries at small

sizes, that is reasonable shaft speeds and inlet flow areas even with small diameter rotors.

In fact, the advantage of using organic fluids to design efficient, small turbomachinery is

a well-known feature of ORC engines [49]. Multiple stages or partial admission

operation is the traditional solution for steam turbines; however, multiple stages add to

the cost and partial admission operation places limits on maximum efficiency.









As for positive displacement expanders, the important parameter is the expansion

ratio. For these machines, the higher enthalpy drop of ammonia manifests itself, not as

high velocities, but increased pressures. Referring back to Table 4-1 indicates that the

ammonia-water fluid has a significantly higher expansion ratio than the presented organic

fluids.

Preliminary Machine Sizing

Given the information from the previous section regarding the expansion properties

of the ammonia-water working fluid, a similarity analysis can be used to strengthen the

generalizations mentioned in a more quantified manner.

For this discussion it will be helpful to use the similarity parameters of specific

speed, nspec, and specific diameter, dspec for a single stage unit. They are defined as:


nspec o- 7/ t (4-1)

and

DA/l ,
dspec = D (4-2)
Qexit

where co is the shaft speed in rad/s, Ahideal is the isentropic enthalpy change across the

turbine(J/kg), and Qexit is the volume flow rate exiting the turbine (m3/s). The term D (m)

represents diameter and is defined differently for different turbine types, but it is a

characteristic dimension that indicates size of the unit. Two additional similarity

parameters are needed to fully describe the performance of geometrically similar

machines, typically machine Reynolds number and Mach number. However, these

additional parameters have only secondary effects and are typically neglected for

similarity analysis [50].









With some general operating conditions from the simulation, a range of typical

specific speed values can be computed and used to identify possible expander designs.

For example, referring to Figure 3-3 the peak efficiency value occurs near a boiling

pressure of 0.7 MPa, this operating condition was used to compute a range of specific

speed values and is presented in Table 4-2. By definition, specific speed is influenced by

the capacity of the device through the exit volume flow rate. This effect is accounted for

by considering two nominal output values in the table.

Table 4-2. Single stage specific speed calculations versus nominal work output and shaft
speed.
Shaft Speed
Ideal Output 5000 rpm 20,000 rpm 60,000 rpm
5 kW 0.0084 0.033 0.10
30 kW 0.021 0.082 0.25

To provide some indication of suitable expanders, specific speed and diameter

ranges taken from suitable references [49, 50] are presented in Table 4-3. What can be

gathered is that partial admission axial turbines could match well with cycle conditions

over a wide range. Full admission devices, both axial and radial inflow, would be

suitable for only very high speeds or large power outputs. However, while partial

admission operation may be better than full admission devices, by reducing clearance and

secondary flow losses, other loss mechanisms appear and the net result is that partial

admission operation will always be less efficient than optimum full admission devices

[51]. For cases of low speeds and/or small outputs a positive-displacement reciprocating

device may give good performance. Note that Table 4-3 is not inclusive of all potential

candidates, other types, for example rotary vane and scroll expanders, could also be

considered as will be shown. Also multiple stages could be used to divide the fluid's

energy content among separate stages and thus change the resultant specific speed values.









Table 4-3. Approximate specific speed and specific diameter ranges for efficient (>60%)
single stage expander types [49, 50].
Expander Type Approximate nspec range Approximate dspec range
Partial admission axial 0.008-0.1 10-50
Full admission axial 0.1-8 1-15
Radial inflow 0.1-1 2-10
Reciprocating piston 0.00002-0.008 14-70
Rotary piston 0.015-0.4 1.5-10

Technology Review

With some background into the expansion situation present in the power-cooling

cycle, a review is now presented which highlights solutions of similar applications,

proposed designs, and expected performance values.

Dynamic Machines

As indicated in the previous section, single stage, full admission turbines are more

appropriate for larger outputs, which is confirmed by examples in the literature. Direct

examples of turbines operating with ammonia or ammonia-water are found primarily in

two research areas: Kalina cycle research and its derivatives and closed loop systems for

ocean thermal energy conversion (OTEC). The Kalina-based research is more relevant

and is discussed here. Closed loop OTEC systems typically employ ammonia as the

working fluid, however, the limited temperature drop being exploited results in turbine

pressure ratios of 1.4-1.5, for example [52], which is lower than the preferred conditions

for the power-cooling cycle. Additionally, some discussion of steam turbine

implementation will also be relevant.

One of the many advantages emphasized by Kalina cycle supporters is the fact that

no new turbomachinery needs to be developed, rather, conventional steam equipment can

be employed successfully due to the fluid dynamic similarities of ammonia and water

[53]. Not surprisingly most examples available in the literature have employed steam









equipment. Some information comes from the few Kalina-based cycles that have been

constructed. The first implementation of this technology was a 3 MW heat recovery pilot

plant at Canoga Park, CA [54]. A subsequent application was utilization of a geothermal

resource in Husavik, Iceland [55], where the nominal output was 2 MW. Finally, a team

at Waseda University in Japan is experimentally investigating a derivative of the Kalina

cycle technology which they term the Waseda Ammonia-Water Mixture Turbine System

(W-MTS) [56]. Table 4-4 is a summary of the pertinent operating features for these

turbines. Other installations have been reported, however, operational data is limited.

Table 4-4. Reported turbine operating parameters and efficiencies for three systems using
an ammonia-water working fluid [54-56].
X T,nlet Pmlet Pexit Exit q
Description [kg/kg] [oC] [MPa] [MPa] Quality Size [%]
Canoga Park 0.70 514 11.03 0.192 1 3.7 MW 90.1
Husavik, Iceland 0.95 121 2.72 0.534 0.946 2 MW 60
Waseda Univ. 0.62 162 1.5 0.385 0.990 60 kW 40

Steam Turbines

It has been previously mentioned that conventional steam equipment has been

proposed and used for operation with ammonia-water mixtures. This facilitates turbine

design for large systems [53], and lowers the investment cost for smaller installations--

where an off-the-shelf device would be used [57]. This section presents typical

efficiencies for available steam turbines using the assumption that they will compare well

to performance with ammonia-water. Over the range of 50 kW to approximately 10 MW

the appropriate turbine choices range from small single stage machines to larger

multistage units with a corresponding efficiency range of 50-80% [58, 59]. Larger output

machines, 20-100 MW, report slightly better efficiencies, to +80% [60]. However, with

tailored designs, better efficiencies are expected, as evidenced by the designs for a









Kalina-based, gas turbine bottoming cycle [53]. High efficiencies, +90%, were estimated

but no verification is available. The computed operating parameters are presented in

Table 4-5. Note that superheating is used to avoid condensed flow at the exhaust.

Table 4-5. Estimated operating data for the three turbine stages of a Kalina-based
bottoming cycle [53].
X Tnlet Pinlet Pexit Exit q
Description [kg/kg] [oC] [MPa] [MPa] Quality Size [%]
HighP 0.817 565.6 19.4 3.45 1 105 MW 94
Intermediate P 0.817 511.8 3.11 0.606 1 96 MW 96.5
Low P 0.817 141.8 0.572 0.281 0.998 23 MW 89.5

Displacement Machines

In displacement machines the high specific energy of the nearly pure ammonia

working fluid is not necessarily a disadvantage for machine design since it is manifested

as higher operating pressures. Within this class are two basic distinctions: reciprocating

and rotary. Reciprocating machines produce linear motion with parts such as pistons or

diaphragms and require valves to alternately open and close for operation. Rotary

devices create expanding chambers through the geometry of one or more rotating

members. Valve operation with these devices is much simpler, usually the rotating

components are used to cover and uncover the inlet and exhaust ports. Reciprocating

machines can be constructed to minimize leakage in the expansion chamber; however,

the need for valve actuation hinders efficiency. Because of their simpler valve

construction, rotary machines are generally preferred even though leakage around the

rotating members that comprise the expansion chambers is a significant limit to

efficiency.

Displacement machines have generally been relegated to small power outputs,

being replaced in the larger sizes by gas and steam turbines for power production









applications. Badr et al. [61] evaluated available expanders on the basis of efficiency and

operation for small scale steam expansion. Reciprocating devices were judged to be less

reliable and efficient than rotary devices for reasons similar to those already mentioned.

The rotary machines evaluated were Wankel, screw, and vane expanders, which have

published isentropic efficiencies (with steam operation) of 13-25%, 25-40%, and 65-80%

respectively [61]. However, a machine type not in common use during the survey by

Badr et al. was the scroll expander. Since the scroll's introduction, primarily as

compressors in air conditioning and refrigeration systems, they have been proposed for

use as expanders in small-scale Rankine systems [18, 20, 21]. There are some examples

of their use in the literature; however, none appear to mention ammonia and very few

with steam. Reported maximum efficiencies for R123 and R134a working fluids are

approximately 67% [18], with compressed air values to 73% [20] have been seen, and

34% is a reported efficiency for non-lubricated steam operation [62]. Table 4-6

summarizes the operational parameters for examples found in the literature.

Table 4-6. Reported efficiencies of scroll expanders [18, 20, 62, 63].
Description Fluid Vol. Ratio Output "
Kane Topping HCFC 123 2.3 5 kW 67 %
Kane Bottom HFC 134a 2.3 8 kW 67 %
Oomori HCFC 123 2 400 W ~ 50 %
Smith A/C Air -2 120 W 74 %
Smith Ref. Air ~4 < 500 W 72 %
Kim Steam-no oil 4.6 15 kW 34 %

On the other hand, displacement devices have an added complication, lubrication.

Lubrication, usually in the form of oil mixed with the working fluid, serves two purposes:

it reduces the friction between sliding surfaces within the device, and more importantly it

enhances the sealing action of the expansion chambers [64]. The high viscosity of oil

compared to vapor working fluid prevents it from being quickly driven through small









gaps created by close running parts. This essentially traps the working fluid for a short

time creating a well-fitted seal. Therefore, to be effective in sealing, the oil must be

applied from the high pressure side of the clearance. In compressors this is not difficult

because the oil can be throttled from the high pressure side to the compressor inlet. For

expander duty, some work must be expended to inject the oil into the high pressure inlet

side [20]. A second problem comes at the expander exhaust when the oil needs to be

separated from the working fluid. It has been reported that oil lubricated steam

expanders form an emulsion in the exhaust stream that was difficult to separate [61].

Also, the condition of the oil itself, quantity and temperature, have their own impacts on

expander performance. Temperature affects the oil viscosity and it must be balanced

between the extremes of being too thin, and not able to seal the working fluid, versus

being too thick, where viscous friction losses are excessive. Badr et al. [64] found that

oil temperature alone affected rotary vane expander isentropic efficiency by

approximately 4%.

Additional Considerations

The previous sections discussed the relative merits and considerations for differing

expander types, this section considers some of the more practical design issues involved

with expander implementation.

When specifying or modifying an expander the corrosive nature of ammonia will

need to be considered. The most critical issue is that copper and copper alloys are

severely attacked by ammonia and will not survive service in an environment directly

contacting ammonia. Another issue is that ammonia mildly attacks certain seal materials

and even some lubricants. Most notably these are Viton and silicone.









Depending on the design, the expander shaft may protrude from the housing and a

seal will be needed. This is a concern for any closed thermodynamic system; however, it

is particularly acute here due to the objectionable properties of ammonia-water. In large

installations, for example the Husavik Kalina cycle, conventional seals pressurized with

nitrogen are used to prevent leakage [57]. Also, the use of labyrinth seals with packing

kept under slight vacuum have been proposed for large turbines [53]. Any ammonia

leakage is diluted in water and is either recycled or used as fertilizer [53, 55].

In smaller systems these complicated sealing systems are not feasible and simple

contacting shaft seals will likely be employed. For the experimental part of this work

graphite-impregnated PTFE shaft seals have been used with some success. Given a

proper design it should be possible to specify a seal with an acceptable lifetime.

However, wear is inevitable so the best solution would be a hermetically sealed device,

very similar in construction to the compressor units of household refrigerators and small

air conditioning units. One such proposal is the use of a high-speed (30-150 krpm)

turbomachine with an integrally mounted generator and feed pump [65]. Similar ideas

are proposed for small (10-200 kW) combustion turbines to also reduce costs by

eliminating the need for a reduction gearbox [66]. The problem is of course corrosion of

the copper windings that will be in an enclosed generator. Two solutions have been

encountered; one was employed in a General Electric-designed ORC which was a

magnetically coupled shaft coupling which transmitted torque through a housing wall

[12]. Another, used in microturbine designs, is a magnetically coupled generator with

shaft-mounted, rotating magnets inside the turbine housing and a surrounding, but

external, set of field windings [66].









A potential concern is the increase in expander exhaust pressure due to the

implementation of the cooling heat exchanger. It has been shown that exhaust

temperatures are sensitive to exhaust pressure, with lower pressures preferred for cooling

production. "Off-the-shelf" heat exchangers may introduce a more severe pressure drop

since their economical designs typically trade performance for compactness (cost). The

same situation has been faced by ORC designers in the past, where regenerators were

used to preheat the liquid boiler feed with the turbine exhaust. Low pressure drop

designs were produced, for example the custom unit described by Batton and Barber [67],

and similar devices would be needed for this application.

Expansion process

Compared to steam, ammonia wets more readily upon expansion. In fact, the work

by Kremmer and Okurounmu [68], which is a study of the condensation process during

rapid expansion, used ammonia because of its relatively high rate of nucleation and its

rapid approach to saturation conditions, as compared to other candidate fluids. This

implies that super-saturation of ammonia will occur to a lesser extent than with steam.

The net effect will be more condensation with ammonia than that compared to equivalent

steam exhaust qualities.

Working fluid condensation could hinder dynamic machine operation because of

the momentum lost to the condensed droplets. These droplets are also responsible for

erosion damage and can limit the useful life of the equipment. For example, the

ammonia-water turbine used at the Husavik geothermal station in Iceland required

rebuilding due to condensate flow from an inadequate separator design [57].

Conditions for the expectation of maximum condensation are described in

Chapter 7. As for contending with the issue, similar conditions are encountered in the









utilization of geothermal steam that is produced by flash boiling. The solution used for

those conditions is to construct blades with erosion and corrosion resistant materials and

to incorporate condensate-draining channels into the blades [69]. Condensation is

typically not a design issue for displacement machines and may actually be beneficial for

sealing purposes.

Conclusion

Based on the information in this chapter, some general divisions can be drawn.

One is on the effect of scale and machine type on isentropic efficiency. For small

outputs, positive displacement devices are preferred, and with some development into the

promising machine types efficiencies of +70% appear possible. The chief design

parameter for these devices is the isentropic volume ratio of the working fluid. At the

other end of the spectrum, large output devices, turbines are dominant and efficiencies to

95% have been demonstrated at the high end of this range and 90% estimated at the lower

end. The mid-range is a different matter, turbines are generally more economical;

however, based on comparable steam performance, efficiencies range from 50 to 80%.

With sufficient flow rates, successful turbine designs can be formulated for nearly any

pressure ratio large to small.

The other separation comes from the thermo-physical properties of the nearly pure

ammonia working fluid. For the purposes of cooling production saturated inlet

conditions are preferred, which leads to condensation forming in the exhaust. Ammonia

wets more readily than steam upon expansion, so the exit quality must be monitored to

avoid damage to the turbine. At least for turbines, this requirement could essentially limit

the minimum exhaust temperature to the corresponding dew point temperature. To avoid

condensed exhaust, similar measures as those used for steam equipment may be needed,






45


for example superheating the inlet stream. The disadvantage is that this would move the

inlet vapor conditions further from the preferred ones outlined in Chapter 3. This would

undoubtedly degrade any cooling production.














CHAPTER 5
EXPERIMENTAL APPROACH

An experimental study of system operation was carried out to demonstrate sub-

ambient exhaust temperatures, confirm expected trends in exhaust temperature and vapor

production, and identify inadequacies in the theoretical modeling. This chapter contains

a description of the experimental setup, general procedures used for experiments, and an

outline of the experiments performed.

Setup Description

The experimental setup used for this work is detailed in this section. The system is

based on the setup used by Tamm [70]. For this work Tamm's boiling-absorption loop

was modified and expanded. Primary additions to the experimental setup include a

rectifier to condition the vapor and a turbine to extract work from the fluid. In Tamm's

experiments the high pressure and temperature vapor was throttled to the low-pressure

side without performing work [70] because a turbine was not used. Instead, a heat

exchanger was used to remove heat from the vapor, which was intended to

thermodynamically simulate the effect of a turbine [70]. In addition to adding a turbine,

modifications to Tamm's circulation loop were also made to improve performance.

The basic schematic of the current experimental system is shown in Figure 5-1,

which is intentionally similar to the modeled system of Figure 3-1. Key differences

between the two are that there is currently no experimental cooling heat exchanger and

the rectifier condensate drains directly to the absorber without mixing with the boiler

weak solution. As shown in the next chapter, experimental cooling production was










minimal so a heat exchanger following the expander was not judged to be necessary for

these tests. The change in plumbing for the rectifier condensate is a result of a repair

modification to the experimental setup. Figure 5-1 also shows the location of

instrumentation in the experimental system. Figure 5-2 is a photograph of the

experimental system.


T T

Cooling Rectifier Condensate


T S T 1U ILS Superheater


Separator



SExpander
Heat M
Source < P
Boiler


Recovery Heat
SExchanger

Key Throttle- -
Key
( Thermocouple Absorber Cooling
Flow Measurement Water
Fluid Sampling Port Solution F
Pressure Transducer Pump (

Figure 5-1. Schematic of experimental setup.

Heated water is used as the heat source for the experiment. It is heated in an

electric water heater which uses phase change material for thermal storage. This water

heater is controlled by an adjustable thermostat. A second, conventional water heater is

used solely as a hot water storage tank. A centrifugal pump is used to circulate water









between the two tanks as well as to the boiling heat exchanger of the power-cooling

system. This arrangement is used for convenience and control of the heat source.


tlgure --2. Photograph ot experimental setup.

The hot water pump circulates the heat source through one side of the boiling heat

exchanger. This heat exchanger is composed of two individual flat plate heat exchangers.

This construction of heat exchanger is also used for the recovery heat exchanger in which

the cool strong solution recovers heat from the weak liquid exiting the separator. A

rotary vane pump, driven by an electric motor, serves as the solution pump and is used to

draw strong solution from the absorber and pump it through these heat exchangers into

the separator, refer to Figure 5-1.

Once the two-phase mixture enters the separator it is separated by gravity into its

liquid and vapor components. The separator itself is simply an empty tank, it contains no









baffles or special equipment. The weak liquid drains from the bottom of the separator

and is pressure-driven through the recovery heat exchanger through a throttling valve into

the absorber. The vapor rises out of the top of the separator and depending on whether

rectification is performed, either enters the bottom of the rectifier or is directed through a

bypass line to the top of the rectifier. The rectifier is the combination of another flat-

plate heat exchanger and a packed-bed entrainment separator. Cooling for the rectifier is

provided by the circulation of chilled water and ethylene-glycol mixture, the same fluid

used to cool the absorber. Any condensate draining from the rectifier passes through a

sight glass and is throttled through a valve back to the absorber. The sight glass is used

to ensure that only liquid is being throttled back to prevent short-circuiting of the vapor.

The vapor that leaves the top of the rectifier is routed through a superheating

section and on to the turbine. Superheating is achieved by a variable- temperature

heating tape that has been wrapped around a portion of the vapor plumbing. There is also

a throttling valve parallel to the turbine so the vapor could be bypassed. From the turbine

exhaust the vapor is routed to the absorber where it is bubbled into a pool of basic

solution liquid through a tube perforated with small holes. Absorber heat exchanger

arrangement is presented in a subsequent section.

As for the absorber cooling source, a vapor-compression water chiller is used to

cool a 50/50 mixture of ethylene-glycol and water. A centrifugal pump circulates fluid

between a storage tank (approximately 110 gallon) and the chiller. The chiller has an

internal, adjustable thermostat that maintains storage tank temperature. A separate

centrifugal pump is used to circulate fluid from the storage tank through the heat

exchangers in the absorber (yet another pump is used to circulate fluid to the rectifier).









As with the heat source, a chiller is used to provide a range of convenient and

controllable heat rejection conditions.

Construction of the system itself is primarily with stainless steel tubing connected

with high quality compression fittings. In some areas, mainly where larger diameters are

needed, the construction is with threaded black iron pipe and fittings. Specific changes to

the original system [70] are described in the following sections and complete

experimental details are provided in Appendix B.

Expander

The expander for the experimental system has to conform to the parameters set out

in Chapters 3 and 4, but it has the additional constraint of operating with low flow rates

produced in the experimental system. Initial testing was performed with a rotary vane

compressor, which was modified to operate as an expander. Performance with this

particular device was poor, apparently due to internal running friction since no oil

lubrication was used. Testing of this device was abandoned in favor of the dynamic

machine described next.

From the specific speed considerations mentioned in Chapter 4, partial admission

turbines can be well suited to the conditions generated by the power-cooling cycle. An

off-the-shelf turbine that could be adapted for partial admission operation was sought.

The device that was found is a single-stage, radial inflow turbine originally for use in an

air-cycle cooling system. It was suitable for this application because of its small size and

because it could be configured to operate in a partial admission configuration. The stator

is made up of a ring of individual nozzles, and these could be individually blocked off,

with an epoxy in this case, so as to operate with a low flow rate in partial admission

configuration. The nozzles have a convergent section only. Furthermore, all components









were aluminum or steel so it would be compatible with the ammonia working fluid. The

original housing on the other hand, could not be used since it was vented to the

atmosphere and the protruding shafts were sealed only with a labyrinth seal. A new leak-

proof rear housing was constructed to mount and enclose the turbine spindle, Figure 5-3.

Also the bearings were replaced since the original ones used a bronze ball retainer that

was quickly consumed by the ammonia working fluid.























Figure 5-3. Modified turbine used for experimental testing.

Loading of the turbine is done in a simple, non-contact manner because of the high

shaft speeds encountered (20,000 60,000 rpm). The turbine drives an aluminum disc

that is mounted inside the housing on the turbine shaft. A strong permanent magnet is

used to induce a counter-current in the rotating disc. This creates a reaction torque that

the turbine must overcome and is proportional to shaft speed. The magnitude of this

counter-torque is controlled by changing the distance between the magnet and the disk.

Of course this device produces no mechanical work, all of the energy is turned to heat









and dissipated (< 300 W). While this method has the disadvantages of not producing

mechanical work and potentially adding heat to the spindle of the turbine, it eliminates

the need to incorporate a shaft seal since the field lines of the magnet can penetrate the

walls of the aluminum turbine housing. This makes the experimental apparatus easier to

construct and maintain. Bearing life may be shortened, however, that is not a concern for

this work. On the other hand, heat transfer from the hot disk to the (ideally cold) turbine

exhaust may prove to negatively affect results. Discussion relating to this point is

provided in the subsequent chapters.

According to Figure 5.1, the inlet temperatures and pressures of the turbine are

measured with appropriate transducers. The turbine was also outfitted for shaft speed

measurements. The sensor is based around an infrared emitter-receiver pair, that registers

a pulse when a hole in the turbine shaft aligns between them. Two pulses occur for each

revolution of the shaft, therefore, the frequency of this signal is proportional to the shaft

speed.

Rectifier

The significant impact of rectification on cooling production was identified in the

theoretical modeling of Chapter 3. This prompted the implementation of a rectifier in the

experimental setup. Construction is very simple, cooling and condensation take place in

a flat-plate brazed heat exchanger and condensate separation occurs in a vertical section

of 2-inch pipe filled with 14 inch Berl saddle packing. The packing captures any

entrained liquid droplets. The column was originally intended to function as a direct

contact heat exchanger; however, performance in that mode of operation was not

satisfactory so the external condenser was added.










Absorber

Modifications to the absorber were directed at correcting two issues: increasing the

heat rejection capacity and directly cooling the exiting strong solution. The initial

configuration of the absorber is shown in Figure 5-4. As can be seen, heat rejection was

taking place only as the weak solution was dripping over the heat exchanger stack. No

cooling was taking place in the liquid pool where a significant amount of heat addition

was occurring due to the bubbling in of the vapor. This resulted in performance problems

such as pressure and temperature build-up in the absorber and served to hinder pump

performance. To alleviate these issues a heat exchange coil was designed and placed

below the existing heat exchanger stack in the liquid pool. A new bubbling tube was

fabricated to accommodate the heat exchanger coil, as shown in Figure 5-4.

Original absorber Modified absorber
configuration configuration


Weak In -= Weak In
Heat
Exchanger
Stack









-Shelf
Liquid Level Coil Assembly

< Vapor In : Vapor In
Bubbling Tube


SStrong Out Bubbling TubeStrong Out


Figure 5-4. Original and modified absorber configurations.









The coils are made of 9.5 mm OD aluminum tubing, and the entire assembly is

made of 3 separate coils connected in parallel. Each of the 3 coils is made of

approximately 3.4 m of tubing, so the entire coil assembly contains approximately 10.2 m

of tubing. Furthermore, this coil assembly is in series with the upper heat exchanger

stack, with the coolant passing through the coil assembly before the stack.

Pump

In previous work, cavitation problems caused by pumping strong solution very near

saturation, were a hindrance to system operation [70]. Several measures were taken to

alleviate those problems for this work. For instance the piping connecting the outlet of

the absorber to the inlet of the pump was simplified, which was comprised of reduced

length and number of bends, and it was increased in diameter to match the exit fitting of

the absorber. In addition, the diaphragm pump was replaced. From an applications

standpoint, a diaphragm pump should have been an appropriate choice for the power-

cooling cycle. However, the old pump, which was removed from other equipment,

appeared to have a minimum inlet pressure requirement and did not pump well, or at all,

at low absorber pressures. This was particularly troubling at startup when the absorber

was cool. A rotary vane pump was chosen as a replacement partially because it could

handle a liquid-vapor mix but also because it was compatible with a drive already in the

lab. Service was adequate, but far from ideal. Perhaps this is not surprising since Barber

and Prigmore advised of poor pump performance when giving design guidelines for ORC

engines [49]. For estimation purposes Barber and Prigmore used a pump efficiency of

40%.









Data Collection

This section describes the general equipment used for data collection, further

details can be found in Appendix B. Most of the collected data is recorded with a

computer-interfaced collection system. All temperatures are measured with T-type

thermocouples and pressures are detected with pressure transducers. Measurements are

saved to a PC through the appropriate interface cards and data acquisition software. Two

instrument types are used for measuring flow rates: the vapor flow into the turbine is

measured with a turbine-type flow meter that provides a signal whose frequency is

proportional to flow rate, and the strong, weak, heat source fluid, and coolant flows are

measured with float-type rotameters and are recorded manually.

Provisions are in place to sample the working fluid at points of interest in the cycle.

Syringe sampling ports were placed on the strong, weak, rectifier vapor inlet, and the

rectifier vapor outlet lines, see Figure 5-1. These correspond to four of the five different

concentrations that are present during steady state operation of the cycle, Figure 3-1. The

procedure for liquid samples is to sample the working fluid with a syringe and then

determine concentration with a gas chromatograph (GC) analysis. This procedure works

well with liquid samples, but not as well with the saturated vapor samples. The vapor

condenses easily in the syringe thus causing errors in the GC analysis. Therefore, vapor

concentrations are determined from property relations using the locally measured

temperature and pressure and using the assumption of saturated vapor at the separator and

rectifier exits.

Experimental Method

For each set of conditions to be tested, a standard test routine was established and is

described here. The first step in establishing stable system operation was to arrange the









parameters that cannot be adjusted during operation, for example the number of open

turbine nozzles. Next, the heat source and heat rejection subsystems were started and

allowed to circulate until temperatures stabilized; then, circulation of the basic solution

was started. The heat source flow rate was controlled to maintain the desired boiling

temperature of the two phase mixture leaving the boiler. With basic solution flow

established the weak solution flow from the separator to the absorber was controlled so as

to maintain the desired level of solution in the absorber. Vapor flow was regulated only

by the nozzle restriction of the turbine. As the solution in the absorber was heated by the

returning liquid and vapor streams, the absorber coolant flow was adjusted to maintain

the desired absorber pool temperature. When the rectifier was active, vapor flow was

diverted through the vapor heat exchanger and the packed bed entrainment separator.

Vapor coolant flow was adjusted to maintain the desired rectifier exit temperature. The

feedback and adjustments mentioned were performed manually. Fortunately, during

testing pseudo-steady-state operating conditions were encountered and the adjustments

were quite manageable.

With the system operating at a specified set of conditions data acquisition could

begin. For each set of conditions several individual measurements were made, usually

eight. The only parameter that was changed during these measurements was the loading

of the turbine. This was done to find the shaft speed where optimum efficiency occurred.

A period of five minutes was judged to be adequate between adjustments to the turbine

loading. Therefore, for each set of operating conditions, testing lasted approximately 40-

45 minutes.









Experiments Performed

Experiments were designed to demonstrate sub-ambient expander exhaust

temperatures and to isolate the important trends identified during the theoretical analysis.

The effective techniques that were available and the variations performed with them are

discussed in this section.

Boiler Exit Temperature

As mentioned in the previous section, the mixture temperature exiting the boiling

heat exchanger was controlled by the flow and temperature of the heat source fluid.

Nominal exit temperatures of 600, 800, and 950 C were considered.

Basic Solution Concentration

The basic solution concentration was varied over a fairly narrow range by

monitoring the solution level in the absorber. General effects were to increase vapor

production and concentration, however, at the expense of absorption pressure. Three

absorber levels were tested which resulted in basic solution concentrations of

approximately 0.381, 0.396, and 0.414.

Superheating

Heating of the vapor before it entered the turbine was performed with an electrical

resistance heating "tape" which was wrapped around a section of the vapor piping. The

electrical tape heating was controlled with a variable transformer. Given the arrangement

only a small amount of superheating was possible, approximately 50 100 C.

Absorption Temperature

Similar to the boiler exit temperature, the absorber liquid pool temperature was

controlled by varying the chilled fluid temperature and flow rate. Considered

temperatures were nominally 250 and 350 C.









Nozzle Flow Area

The effect of nozzle flow area was to change the flow restriction imposed by the

turbine. This directly controlled the coupled effects of boiling pressure and vapor

concentration. For a discernable change the number of open nozzles tested were one and

four.

Rectification

As with the theoretical modeling, the amount of rectification in the experimental

setup was determined by the rectifier exit temperature. This in turn was controlled by the

coolant flow rate through the condensing heat exchanger. Only two conditions were

tested, either no rectification was performed or the exit temperature was nominally set for

350 C.

Conclusion

This chapter has presented an overview of the experimental system that was used

for this work. As shown in the concluding chapters, the data from these experimental

tests can be used to provide general verification of the trends identified in Chapter 3 and a

moderate demonstration of the power-cooling cycle's key concept. Also, unexpected

behavior of this real system adds to the collective design experience.














CHAPTER 6
EXPERIMENTAL RESULTS

In the previous chapter the techniques and types of experiments were described.

This chapter presents the results of those tests as they relate to the operating trends

identified in the theoretical analysis. Overall, the anticipated trends have been confirmed

by these results. However, there were unexpected deviations, and these included

problematic turbine performance and other equipment limitations that obscured some

effects.

Confirmation of Trends

The results based on system modeling have been informative regarding the

characteristics and potential of the power-cooling cycle. Confirmation of this expected

performance is needed in order to place confidence in the derived conclusions. This

section describes the experimental results as they relate to the operating mechanisms

under discussion.

Pressure Variation

The boiling pressure is the first parameter to be considered. Its effect is to control

the vaporization of the basic solution in the boiler. As pressure decreases more vapor is

formed and its concentration drops--ultimately until the saturation pressure is reached and

all of the basic solution has vaporized. These effects have been isolated experimentally

and are presented in Figure 6-1.







60



1 0.35

0.95 0.3
Simulated Vapor
Concentration
0.9 (+/-0.00067 kg/kg) 0.25 -

8 0.85 ,,
0 0.2
0 0
0.8 LL
0E Simulated Vapor Mass Flow 0.15
E 0.75
< 0..
o Measured Data -0.1
U 0.7
> 04 nozzle xv --
01 nozzle xv
0.65 4 nozzle mass flow 005
o 1 nozzle mass flow
0.6 1 0
0.2 0.3 0.4 0.5 0.6 0.7 0.8
Boiling Pressure [MPa]


Figure 6-1. Measured effect of pressure variation on vapor quantity and concentration.
Simulated data superimposed to extend observed trend.

To obtain Figure 6-1 the number of open nozzles in the turbine was changed from 1

to 4. The effect of increasing the vapor mass flow rate and decreasing the vapor

concentration is clearly shown. Simulated results are superimposed on the figure and are

used to extend the trend to the limits of partial vaporization.

Concentration Variation

Basic solution concentration variations have similar ultimate effects as those of the

boiling pressure. That is the vapor mass flow fraction will increase with the basic

solution concentration. However, with fixed boiling temperature and pressure, the vapor

concentration remains constant. These trends are seen in Figure 6-2, which also has

simulated results superimposed. Since the liquid level in the absorber was used to vary

the basic solution concentration, the span of concentrations was limited to a somewhat







61


narrow range. As seen in Figure 6-2 the measured vapor flow rate is consistently lower

than that expected from the equilibrium model indicating some inefficiencies in the vapor

production process.


1 0.35

0.9
/ 0.3
0.8 Simulated Vapor Concentration

E 0.7 0.25o
0.7
a 0.6
o 0.2
O 0
.c 0.5- -- l
o Simulated Vapor Mass Flow / 0.15
E 0.15 |
E 0.4
E 7

0 0.3 0.1 C

0.2 / 0 Measured Concentration,
+/-0.00067 kg/kg 0.05
O Measured Mass Flow
0.1

0 1 0
0.2 0.25 0.3 0.35 0.4 0.45 0.5
Basic Solution Concentration

Figure 6-2. Measured effect of basic solution concentration on vapor production.

The other effect of basic solution concentration is to reduce the expander pressure

ratio by the increase of absorption pressure with increasing concentration. The relevant

data are presented in Table 6-1.

Table 6-1. Measured decrease of absorption pressure with basic solution concentration.
Basic Concentration Ave. Absorption Temp. Ave. Absorption Press
0.414 34.80 C 0.250 MPa
0.396 34.70 C 0.237 MPa
0.381 34.40 C 0.225 MPa

Temperature Variation

Temperature is the last boiling condition to be considered. It also changes the

degree of vaporization and thus the amount and concentration of vapor that is formed.










The simulated effect of temperature change has been shown in Chapter 3, however, in the

experimental system there is some coupling between temperature and pressure because of

the fixed turbine nozzle restriction, that is the pressure is not held constant. Figure 6-3 is

a plot of the change to vapor mass flow fraction (mass flow vapor/mass flow basic

solution) with boiling temperature. The expected trend of an increase in vapor flow with

temperature is shown, but it is somewhat curtailed due to the associated rise in boiling

pressure, also shown. Agreement with equilibrium modeling is reasonable except for the

highest temperature case where an unusually high concentration reading is the suspect.


1 0.3

0 Pressure (+/-2.24 kPa) 0
0.9 0
Vapor Flow Fraction/ -0.25

-0.8 0
Simulated vapor mass 0.2
Sflow fractions
S0.7
3~~ \0
0.15 uL
0.6

o 0.1
m 0.5 ,


0.4 0.05
Average pressure values
'' used in simulation
0.3 1 1 0
50 55 60 65 70 75 80 85 90 95 100
Boiling Temperature [C]

Figure 6-3. Measured change in vapor flow rate (relative to basic solution flow) due
primarily to changes in boiling temperature. The effect is counteracted by the
indicated rise in boiling pressure.

Absorption Pressure

The effect of absorption pressure on cooling production was shown in Chapter 2

with the aid of an ammonia-water binary phase diagram. As it pertains to that diagram,









increasing or decreasing absorption pressures shift the two-phase envelope either up or

down respectively. Increases to pressure would then increase the dew point of the turbine

exhaust, which is an initial indicator of the minimum temperature possible.

Experimentally measured turbine inlet and outlet pressures and temperatures are

presented in Table 6-2. As shown with the sample data, the lower exhaust pressure

corresponding to the lower absorption temperature allows the vapor to be expanded to

lower temperatures.

Table 6-2. Measured data indicating effects of absorption temperature.
Turbine Parameter 250 C Absorption T 350 C Absorption T
Inlet Pressure [MPa] 0.329 0.329
Exhaust Pressure [MPa] 0.157 0.217
Inlet Temperature [C] 56.0 55.5
Exhaust Temperature [C] 42.0 46.1

Rectifier Penalty

Rectifying the vapor before it enters the expander was shown to have a positive

effect on exhaust temperatures and cooling production to a certain extent. The benefits to

vapor temperature and concentration from rectifier operation is presented in Table 6-3.

Also indicated in the table is the concomitant loss of vapor due to condensation.

Naturally, the cases with the lowest inlet concentration have more of the vapor condensed

to reach essentially the same ending concentration.

Table 6-3. Averaged values for rectifier operation.
Nominal Boiling
Rectifier Parameter Temp. = 600 C 800 C 950 C
Inlet Temp. [oC] 58.7 78.1 94.4
Exit Temp. [oC] 34.0 36.2 33.9
Inlet Concentration 0.968 0.927 0.854
Exit Concentration 0.993 0.994 0.996
Normalized
Inlet Mass Flow 1 1 1
Normalized
Exit Mass Flow 0.953 0.866 0.695










Another way to view the losses associated with rectification is to consider the work

that was sacrificed for lower exhaust temperatures. Table 6-4 presents such information

for a couple of the cases in Table 6-3. The measured work out is the work output, per

kilogram of basic solution flow, based on experimentally measured conditions at the

point of maximum turbine efficiency. The computed work out is the estimated work that

could have been produced with the vapor not going through the rectifier but straight to

the turbine. For this calculation, the turbine's efficiency was assumed to be the same as

the maximum measured value. The data in Table 6-4 clearly show the detrimental effect

of excessive rectification on work production.

Table 6-4. Values for rectifier operation highlighting penalty to work production.
Parameter Nominal Boiling T = 600 C 950 C
Measured Exhaust Temp. 25.90 C 22.30 C
Max Measured Work Out 563 J/kg 1660 J/kg
Computed Work w/no Rect. 612 J/kg 2600 J/kg
Computed Drop in Work 7.9 % 36 %

Concept Demonstration

The realities of the experimental setup required a compromise in the testing plan.

Based on experimental measurements, the turbine, even with a single nozzle, is slightly

oversized for the experimental setup. The result is that the boiling pressure falls to a

value that is below the optimum identified in the theoretical analysis. The consequences

come in the form of reduced pressure ratios and vapor concentrations, both effects

degrade cooling production. To compensate, boiling temperatures were increased to

compromise between sufficient vapor flow rate for turbine operation and sufficient

pressure to allow for a high degree of rectification without sub-ambient condensing

temperatures. As can be concluded from the previous theoretical analysis, this incurred









much loss due to the rectifier, but the resulting vapor mass flow and turbine inlet pressure

was higher than could be achieved at lower temperatures. Efficiency was essentially

traded for more suitable vapor conditions.

Figure 6-4 shows the measured temperatures across the turbine in relation to the

measured absorption temperature and Table 6-5 presents the parameters for this testing.

Successive stages of rectifying and superheating enabled the production of vapor with

0.993 concentration and temperatures ranging from the vapor saturation temperature,

approximately 390 C according to Figure 6-4, up to the useful limit of the superheater.



rTexpander = 20 %
33
I t rt i25
3 0 -----


27


224


21 O Expander Exhaust
STemp
SmAbsorption Temp.
18


15
35 38 41 44 47 50 53
Expander Inlet Temp. [C]

Figure 6-4. Experimental measurement of the expansion of vapor to temperatures below
those at which absorption-condensation is taking place.

Obviously the minimum exhaust temperatures of Figure 6-4 are not suitable for a

cooling load, however, it is a clear measurement of the power-cooling cycle concept as it

was explained in Chapter 2. In Chapter 2 the power-cooling cycle was contrasted with









pure working fluid Rankine cycle operation, where it is impossible to expand the vapor to

a temperature below that at which condensation is taking place. While not a dramatic

demonstration, Figure 6-4 clearly shows expansion of the vapor to temperatures well

below the absorption-condensation temperature.

Table 6-5. Averaged conditions for the testing of Figure 6-4.
Expander Inlet Pressure: 0.516 MPa
Expander Exit Pressure: 0.208 MPa
Vapor Flow Rate: 0.00299 kg/s
Vapor Concentration: 0.993 kg/kg
Rectifier Inlet Temp.: 83.40 C
Absorber Temp.: 31.40 C

Superimposed with the data points of Figure 6-4 are lines of simulated performance

for several isentropic efficiencies. At the lower inlet temperatures the experimental data

is approximated by an expansion process with 20% isentropic efficiency. At higher inlet

temperatures the experimental data drifts away from the 20% line and appears to improve

in efficiency. This is an unexpected deviation in the measurements and its possible

source is discussed further in the next section.

Expander Performance

Some difficulties with the thermodynamic performance of the turbine were

encountered. First, the efficiency with the ammonia-water working fluid was lower than

the anticipated efficiency obtained from air testing the turbine. Second, some of the

results based on thermodynamic measurements seem to indicate that the turbine

efficiency is sensitive to inlet conditions. This section is a summary of the analysis into

these phenomena and the conclusions regarding the turbine performance measurements.

Initial testing with the turbine was performed with compressed air as the working

fluid since it was simple to control and leaks were not a problem. Details of these tests as

they relate to this work are provided in Appendix D. Based on this air testing the









expected turbine efficiencies were in the range 20-30%. The optimum ratio of ideal jet

velocity to rotor tip speed was approximately 0.3. In order to maintain a similar ratio

when testing with the ammonia-water mixture it would be necessary for the rotor speed to

increase because of the higher ideal jet velocity for ammonia-water. However, this was

not observed, possibly due to the reduced mass flow rate of ammonia-water as compared

to the tests with air. This likely caused additional incidence losses and resulted in the

lower efficiency.

Using thermodynamic measurements the turbine efficiency appeared to vary and

seemingly worsened as cooler exhaust conditions were approached. A few thoughts on

the measured performance are given here. In general, the observations from the

experimental testing followed these trends: the expander exhaust consistently expanded

to a point at or near the dew point for the measured exhaust pressure and estimated vapor

concentration. This resulted in good indicated performance when the inlet temperatures

were significantly higher than the exhaust dew point and poor indicated performance

when the inlet temperatures were not significantly higher, for example the cases with

rectification. A few possible explanations are discussed below.

The expander is a partial admission dynamic turbine which was not designed to

expand a two-phase working fluid. If enough flow were to condense it would alter the

momentum transfer in the turbine and there would be a corresponding drop in efficiency.

However, when the amount of condensation is examined for the experimental conditions,

the concomitant effect on efficiency should be small. For example, Figure 6-5 presents

the expected expander exit quality values for simulated conditions similar to those of the

experimental testing. As can be seen, even for an isentropic expansion, the minimum







68


expected exit quality does not go below 0.97 for these conditions. This amount of

condensation would result in a minor expected penalty, 0.95 or higher [50]. Unless there

were substantially more condensed flow than expected, it does not appear to explain

expander performance.



Slexpander --
0.995 30 %


0.99
70 %
X 0.985


100 %
S0.98


0.975 Pniet/Pexit = 0.516 MPa/0.208 MPa
Xvr= 0.993

0.97 1111
35 37 39 41 43 45 47 49 51 53 55
Expander Inlet Temp. [C]

Figure 6-5. Expected equilibrium exhaust qualities for conditions similar to those of the
experimental study. The exit quality is expected to be above 97% even for an
isentropic device.

Another possible explanation could come from errors in the thermodynamic power

measurements. For pure component fluids near saturation conditions, the sensitivity of

temperature to enthalpy changes is poor due to their isothermal phase change. This could

introduce significant error in temperature-based enthalpy measurements. For the

ammonia-water binary mixture and the range of condensation which is being considered,

however, condensation does not take place isothermally, even for a very high ammonia

concentration vapor. Figure 6-6 is a temperature-enthalpy diagram for two fluids, pure







69


ammonia and a high concentration ammonia-water mixture, at a pressure typical of the

exhaust conditions for the expander. The curves are similar except near the saturated

vapor-two phase mixture boundary, where condensation begins for the mixture in the

equivalent sensible heating range for the pure fluid. Figure 6-6 indicates that expansions

resulting in qualities of 0.97 or greater fall within a range where the sensitivity of

temperature to enthalpy changes is good. Furthermore, as Figure 6-6 also shows, if the

near-isothermal phase change region were being encountered, the temperature would be

drastically lower than what has been measured.


35
Pressure = 0.208 MPa
Mixture Vapor Quality = 1
25


15
15 0.99-
2- 0.993 Ammonia Concentration
S5 -- Pure Ammonia




0.97
-15


-25
-200 0 200 400 600 800 1000 1200 1400
Enthalpy [kJ/kg]

Figure 6-6. Temperature-enthalpy diagram covering the phase change of pure ammonia
and a high concentration ammonia-water mixture. For mixture qualities
above 97% the sensitivity of enthalpy to temperature appears good.

Direct measurements of power output were attempted and are described in

Appendix D. The mechanism worked while testing with air, however, it gave

inconclusive results with ammonia-water testing. Due to this, air testing results were










used to estimate the no-load power needed to drive the expander. These values can then

be compared to the no-load results from the ammonia-water testing, which is shown in

Figure 6-7.


120
0 0
oAmmonia-water no-load test points 80 C 95 C
100 labeled with the nominal expander
inlet temperature

80


60
No-load power consumption
a- based on air testing 35 C
40
40 60 C T

20 T
S35 C 35 C

035C
29000 31000 33000 35000 37000 39000 41000 43000
Shaft Speed [rpm]

Figure 6-7. Comparison between the measured no-load power consumption of operation
with compressed air (solid line) and ammonia-water (individual points).

As can be seen in Figure 6-7, the air testing results indicate an approximate no-load

power consumption of 10 to 25 W over the shaft speeds considered. On the other hand,

the ammonia-water test points vary substantially. For instances where the expander inlet

temperatures are nominally 600 C or lower (the multiple 350 C readings are all cases

where rectification was used) the no-load power consumption is in the range of 10 to 42

W, and at inlet temperatures of 800 C and above the consumption is greater than 110 W.

While these results are not conclusive in themselves (because not all of the test conditions

were equivalent) they do imply that the power output is greatly over-estimated for cases









with high expander inlet temperatures. This suggest that an unaccounted heat transfer

from the hot inlet fluid may be skewing the energy balance of the expander. Because of

this, those particular experimental results have not been incorporated into this work.

Admittedly, this expander was not the best choice for this size of system, as

evidenced by the recommendations for expanders given in Chapter 4. It was, however, a

choice between relative performance among tested devices and ease of adaptability to the

ammonia-water working fluid. Had the anomalies just discussed not occurred and the

efficiency was within the anticipated range of 20-30%, exhaust temperatures 5-100 C

cooler could have been expected.

Conclusion

A demonstration of the key concept of the power-cooling cycle has been provided

and the trends important to cooling production have been verified. However, the small

scale of the experiment complicated the testing conditions so full agreement was not

achieved, neither was a truly convincing example of combined power and cooling

outputs. These complications were more evident in the turbine than in the other

components where performance was poor and some readings were erroneous. Another

issue is the fact that even with the minimum flow of the turbine, a single open nozzle, it

regulated pressure to a lower level than that considered optimum by the theoretical

analysis. On the other hand, the experimental setup was not designed to be an economic

success. Rather, it is a test-bed for exploring operating issues with the power-cooling

cycle and some observations from it are included in the conclusions next chapter.














CHAPTER 7
DISCUSSION AND CONCLUSIONS

The conclusions of this work can be broadly divided into two categories: those that

are derived from a theoretical analysis of the power-cooling cycle thermodynamics and

those that result from the deviations encountered during the experimental study. Based

on this information, the following conclusions regarding cooling production have been

formed.

In general, cooling production with this cycle is counter-productive to work output.

This is a direct consequence of the need to reduce the entropy of the vapor in order to

expand it to low temperatures. The effective COP parameter, introduced previously, is

used to quantify the trade-off of work and cooling and to select favorable cooling

conditions. Characteristics of these optimum conditions are explored with regards to

system implementation and operation. As for the experimental results, they have been

presented in Chapter 6 to verify many of the operating mechanisms of the power-cooling

cycle. However, while these mechanisms are an aid for system design and evaluation,

experimental testing indicates that they will have only secondary effects on the operation

of a real system. The primary effects, which were largely unaccounted for, are described

and their impact on system performance evaluated.

Cooling Conditions

During the discussion of the operating mechanisms of the power-cooling cycle it

was noted that cooling production has a maximum value for a given heat source

temperature. This maximum is a result of the balance between vapor mass flow rate and









minimum temperature from the exhaust. In this section, the balance of conditions are

quantified by using the effective COP parameter introduced in Chapter 3. Initially, the

optimum amount of rectification is determined, then the overall energy advantage of the

power-cooling cycle is evaluated by comparison of cooling-optimized and work-

optimized systems.

Expander choice and expected performance has been discussed in Chapter 4. The

general conclusion was that the expected efficiency increases with power output, ranging

from 60-70% for multi-kW displacement machines to +90% for multi-MW dynamic

turbines. Turbines also cover the mid-output range with widely varying efficiencies, 60-

90%, where much of the variation depends on whether stock steam turbines are used or

custom design takes place. For the remaining simulation results, efficiencies have been

chosen to place bounds on the anticipated performance.

Optimum Rectification

For the heat rejection temperatures considered in this work, some rectification is

necessary to achieve any practical cooling, so the determination of optimum cooling

conditions is largely an issue of determining the optimum amount of rectification. The

effective COP value presented in Chapter 3 is used to determine this balance, it is

repeated here as Equation 7-1.


ool
COP -Q001 (7-1)
no rect with rect

The optimization strategy given a boiling temperature, absorption temperature, and

expander efficiency, is to chose the boiling pressure, rectification temperature, and basic

solution concentration that maximize the effective COP value. This results in the most










energy effective trade-off of work and cooling production given the specified inputs.

Figure 7-1 is a plot of the maximum effective COP as a function of boiling temperature.


5



4.5



4
0
90%

3.5

Expander Efficiency = 60 %

3



2.5
60 65 70 75 80 85 90 95 100
Boiling Temp. [C]

Figure 7-1. Maximum effective COP values where the work component is the amount of
work lost due to operation with rectification vs. equivalent conditions with no
rectification.

The results of the maximum effective COP appear quite good. In the best cases,

the work that is lost due to rectification penalties is effectively traded for 4-5 times the

quantity of cooling. This is better performance than most conventional cooling systems,

however, it is not quite the complete picture of energy efficiency since work production

can also be optimized.

Overall Optimum Cooling

For a given heat source temperature, work optimization results in the system

configuration evolving toward a pure component working fluid Rankine cycle, that is a

high ammonia concentration in the basic solution. General optimum characteristics for









the power-cooling cycle have already been discussed and a brief comparison of the two is

given in Table 7-1.

Table 7-1. Typical operating characteristics for cooling and work optimized cycles.
Optimization Vapor Mass Flow Fraction Basic Solution Concentration
Cooling 5-10% approx. 0.50
Work +90 % approx. 0.95

For a more stringent evaluation, cooling production can be assessed by the amount

of lost work that could be obtained from a work-optimized system. When considering

this scenario the overall effective COP has the following formulation.


Q R
C=OP coa =W (7-2)
uloerall W -W <. /
work opt w/cool work opt
K 7wcool )

Where the subscripts work opt and w/cool refer to parameters for the work

optimized case and those with cooling production, respectively. The rightmost

formulation represents the actual implementation with dimensionless parameters. The

term Rc/w is the ratio of cooling to work.

Figure 7-2 presents the maximum values of the overall, effective COP as a function

of boiling temperature. Based solely on energy considerations and the assumptions

inherent to Figure 7-2, at the best conditions nearly equal amounts of work must be

forfeited for the amount of cooling gained. The ratio of cooling to net work output is also

presented in Figure 7-2 to provide the relative magnitudes of each.

Exhaust Temperature

The exhaust temperature will determine suitable cooling applications, e.g. space

conditioning vs. refrigeration. The corresponding temperatures for the optimized cooling

cases of Figure 7-2 are presented in Figure 7-3.













1.2




1


"5
0_
-S
0.8



0.6 z
0
0

0
0.4 o




0.2


0.5 --- 0
60 65 70 75 80 85 90 95 100
Boiling Temp. [C]

Figure 7-2. Maximum overall effective COP values as defined by Equation 7-2.


15


5



\ \ Expander Efficiency = 90 %
0
X
-5







-5 60 %


-10
60 65 70 75 80 85 90 95 100

Boiling Temp. [C]

Figure 7-3. Corresponding exhaust temperatures for the optimum conditions presented in
Figure 7-2.









However, these values are complicated by numerically maximizing the effective

COP values and do not provide a complete picture of the exhaust temperatures attainable.

Consideration of the relationship between overall effective COP and other system

parameters provides a better indication of achievable exhaust temperatures. Figure 7-4

attempts to show this relationship by plotting results of constant temperature operation in

a boiling pressure-basic solution concentration plane. Three parameters are plotted in

Figure 7-4: the overall effective COP value, the expander exhaust temperature, and the

vapor mass flow fraction (mvr/ms). The maximum effective COP value is centered within

the 1.08 isoline at approximately a pressure of 1.03 MPa and a basic solution

concentration of roughly 0.46. The corresponding exhaust temperature and vapor flow

fraction is -4.5 C and 7.8%, respectively. Figure 7-4 shows that the effective COP is

much more sensitive to the vapor flow fraction than the exhaust temperature. Therefore,

with a mild penalty to effective COP values, a large range of exhaust temperatures could

be accessed. For example, while operating within the 1.08 effective COP isoline, the

exhaust temperatures could vary between approximately -7 C and -2.3 C. Furthermore,

if the effective COP were diminished by 7.5% (the 1.0 isoline), the range of possible

exhaust temperatures would be between roughly -12 C to 5.60 C.

Implementation

The preceding results identify preferred conditions for cooling production. In this

section some of the implications of these conditions are considered as they relate to

physical implementation of the power-cooling cycle.












1.7. 5%
1*"; t" ,"10% / l


089=Ehu Tm. [
0 1.4


Sm s-4.5' ..Oth .-/ overall
Y) w wt a mil pa Effective
J_ 1.1o




v0.8 '.sf.l-- 9f =de Exhaust Temp. [C] low i



0.5
0.4 0.44 0.48 0.52 0.56 0.6
Basic Solution Concentration

Figure 7-4. Design point map showing the relative sensitivity of overall effective COP to
vapor mass flow fraction and exhaust temperature. Sensitivity to mass flow is
high, while with a mild penalty to effective COP a wide range of temperatures
could be expected. Boiling temperature of 800 C.

Vapor Quality

Considering the optimum cooling conditions, certain regions may cause exit quality

concerns. Combinations of high boiling temperatures, high expander efficiencies, and

effective cooling production result in relatively low equilibrium exhaust qualities, 90-

95%. This is due to the fact that at these conditions the entropy of the vapor is low

enough that it can be expanded to conditions that have significant condensation.

It was concluded in Chapter 4 that dynamic machines, turbines, were the better

expander choice for medium to large work outputs. However, these machines are

sensitive to vapor quality, in terms of both operating efficiency and erosion damage.









From examination of steam turbine performance, the effect of vapor quality on

machine efficiency is mild, for example an exit quality of 90% would result in an

expected efficiency penalty of 5-10% [50]. On the other hand, erosion of flow path

components is a significant concern since it could require machine rebuilding or

replacement. For these cases cooling may have to be curtailed, either by limiting

rectification or superheating, to avoid unacceptable conditions.

Rectifier Implementation

It was pointed out in Chapter 3 that the physical setup of the rectification process

can have an effect on the penalties to work production because it can affect the quantity

of vapor that continues on to the expander. Here, these effects are reconsidered in light

of the conclusions regarding optimum rectification.

In Chapter 3 upper and lower bounds were placed on the efficiency of the

rectification process. It was pointed out that with high amounts of rectification, lower

rectifier exit temperatures, the difference between the two becomes excessive while there

is not a significant difference with minimal rectification.

The effects of minimum and maximum rectifier efficiencies on the optimum overall

effective COP is shown in Figure 7-5. From the figure it can be seen that compared to

the minimum performance case, an approximate 20% improvement could be obtained for

most conditions. This benefit will be weighed against the added complexity of a direct

contact heat exchanger.

Experimental Observations

Correlation of the experimental results to simulated data revealed unexpected and

unaccounted trends in system operation. Many of these deviations overshadowed the










basic operating mechanisms that were studied analytically. This section explains these

deviations and any impact on system design.


1.4 50

COP with max. possible
1.2 rectified vapor production
1.2 40



a
O 1 30 V

COP with minimum o
rectified vapor production -
S0.8 "20
O



0.6 10

% Increase

0.4 11 0
60 65 70 75 80 85 90 95 100
Boiling Temp. [C]

Figure 7-5. Effect of the minimum and maximum bounds of rectifier operation on
effective COP values.

Absorption Pressure

The experimentally measured absorption pressure was higher than that predicted by

using saturation properties at the basic solution concentration and the measured

temperature. This is not surprising since previous experimental investigation [70]

showed similar results. However, part of the explanation is an unexpected coupling

between the boiling temperature and basic solution concentration. Basic solution

concentration increased with temperature, thus increasing absorption pressure, due to the

storage of weak solution in the separator.









At steady state operation all of the ammonia and water leaving the absorber are

returned to it so that the concentration of the solution in the absorber remains fairly

constant. It would be expected then, even at different operating conditions, that the

concentration remains at its original value. However, in the experimental setup there is

some stored liquid in the separator that causes this scenario to be out of balance. The

liquid is retained to form a vapor seal in the separator and amounts to approximately 25%

of the total working fluid mass. This is weak solution liquid, which is a larger percentage

water than the basic solution, so by basically removing it from the working fluid, the

basic solution concentration inevitably increases. Furthermore, the weak solution

concentration is dependent on the boiling temperature and this results in the coupling

between boiling temperature and basic solution concentration. A formulation of this

effect is given as Equation 7-3.


Xb sC new Xc ong YXweak (7-3)
1-y

Where the new basic solution concentration, Xbasic new, is dependent on the original

concentration, Xbasic orig, the current weak solution concentration, xweak, and the fraction of

the total mass of working fluid that is stored, y. The effect is plotted in Figure 7-6 as a

function of both boiling temperature and stored fraction. Also included in the figure are

experimental measurements of the basic solution concentration at the approximate

amount of storage.

As shown in Figure 7-6, the change in basic solution concentration is more

pronounced with higher boiling temperatures and stored fractions. Accounting for this

effect when determining absorption pressure, Figure 7-7, explains much of the difference










when simply assuming a constant concentration, 0.35 for this case. The remainder of the

difference is attributed to non-equilibrium performance of the absorber.


c 0.45
0
U 80 C


0.4
60 C


0.35



0.3
0 10 20 30 40 50 60
Stored Fraction [%]

Figure 7-6. Computed effect of weak solution storage on basic solution concentration.
Experimentally measured basic concentrations also shown.

The crucial effect of absorption pressure on cooling production has been

mentioned, lower pressures are certainly preferred. Therefore, if not taken into account,

an undue pressure rise caused by this effect could limit cooling production. The obvious

solution, according to Figure 7-6, is to minimize the stored fraction.




Full Text

PAGE 1

STUDY OF COOLING PRODUCTION WITH A COMBINED POWER AND COOLING THERMODYNAMIC CYCLE By CHRISTOPHER MARTIN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2004

PAGE 2

Copyright 2004 by Christopher Martin

PAGE 3

iii ACKNOWLEDGMENTS I would like to express my appreciation to those peopl e who supported this work and provided me with the enc ouragement to pursue it. Firs t I would like to thank my advisor, Dr. D. Yogi Goswami, for his teach ing and providing me with this opportunity. Additionally, I would also like to thank Dr. Skip Ingley, Dr. William Lear, Dr. S. A. Sherif, and Dr. Samim Anghaie for serving on my advisory committee. Their time and consideration are appreciated. Special thanks are also extended to the editorial staff of the Solar Energy and Energy Conversion La boratory (SEECL), Barbara Graham and Allyson Haskell. Also, the advice and hu mor of Chuck Garretson have been much appreciated during my time at the SEECL. There are also many colleagues I would lik e to thank for their help, consultation, and camaraderie. Gunnar Tamm and Sanjay Vijayaraghavan have provided excellent examples that I have tried to follow. I began this process with Nitin Goel and Amit Vohra, with whom I have become friends. Also I have made friends with the recentlyjoined students, Madhukar Mahishi, Shalabh Maroo, and Ben Hettinger. I would like to also acknowledge the suppor t of my parents, Lonnie and Loretta. Finally, and most importantly, I want to thank my wife Janell for her unquestioning support during this work. Without it, I doubt that I would have reached this personal milestone.

PAGE 4

iv TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iii LIST OF TABLES...........................................................................................................viii LIST OF FIGURES.............................................................................................................x NOMENCLATURE........................................................................................................xiii ABSTRACT....................................................................................................................xvi i CHAPTER 1 INTRODUCTION........................................................................................................1 Motivation..................................................................................................................... 2 Power-Cooling Concept................................................................................................4 Problem Definition.......................................................................................................5 Research Objectives......................................................................................................6 2 BACKGROUND AND REVIEW................................................................................7 Background...................................................................................................................7 ORC Development................................................................................................8 Ammonia-Water Cycles........................................................................................9 Power-Cooling Concept..............................................................................................10 Prior Work..................................................................................................................12 Other Power-Cooling Concepts..................................................................................15 Conclusion..................................................................................................................17 3 THEORETICAL STUDY..........................................................................................18 Model.......................................................................................................................... 18 Operating Mechanisms...............................................................................................20 Effect of Boiling Pressure...................................................................................20 Effect of Mixture Concentration.........................................................................22 Effect of Boiling Temperature.............................................................................23 Cooling Production.....................................................................................................24 Exhaust Temperature...........................................................................................24

PAGE 5

v Vapor Flow Rate..................................................................................................27 Rectification................................................................................................................27 Performance Measures................................................................................................30 Work Production.................................................................................................30 Cooling Production..............................................................................................30 Conclusion..................................................................................................................31 4 EXPANDER CONSIDERATIONS...........................................................................33 Working Fluid Properties...........................................................................................33 Preliminary Machine Sizing.......................................................................................35 Technology Review....................................................................................................37 Dynamic Machines..............................................................................................37 Steam Turbines....................................................................................................38 Displacement Machines.......................................................................................39 Additional Considerations..........................................................................................41 Expansion process......................................................................................................43 Conclusion..................................................................................................................44 5 EXPERIMENTAL APPROACH...............................................................................46 Setup Description........................................................................................................46 Expander..............................................................................................................50 Rectifier...............................................................................................................52 Absorber..............................................................................................................53 Pump....................................................................................................................54 Data Collection....................................................................................................55 Experimental Method.................................................................................................55 Experiments Performed..............................................................................................57 Boiler Exit Temperature......................................................................................57 Basic Solution Concentration..............................................................................57 Superheating........................................................................................................57 Absorption Temperature......................................................................................57 Nozzle Flow Area................................................................................................58 Rectification........................................................................................................58 Conclusion..................................................................................................................58 6 EXPERIMENTAL RESULTS...................................................................................59 Confirmation of Trends..............................................................................................59 Pressure Variation...............................................................................................59 Concentration Variation......................................................................................60 Temperature Variation.........................................................................................61 Absorption Pressure.............................................................................................62 Rectifier Penalty..................................................................................................63 Concept Demonstration..............................................................................................64 Expander Performance................................................................................................66

PAGE 6

vi Conclusion..................................................................................................................71 7 DISCUSSION AND CONCLUSIONS......................................................................72 Cooling Conditions.....................................................................................................72 Optimum Rectification........................................................................................73 Overall Optimum Cooling...................................................................................74 Exhaust Temperature...........................................................................................75 Implementation...........................................................................................................77 Vapor Quality......................................................................................................78 Rectifier Implementation.....................................................................................79 Experimental Observations.........................................................................................79 Absorption Pressure.............................................................................................80 Rectifier Pressure Effect......................................................................................83 Conclusion..................................................................................................................85 8 RECOMMENDATIONS............................................................................................88 Experimental Testing..................................................................................................88 Practical Application..................................................................................................90 ORC Comparison................................................................................................90 Cooling Production..............................................................................................91 Conclusion..................................................................................................................92 APPENDIX A PROPERTY EVALUATION.....................................................................................94 Pure Component Properties........................................................................................94 Liquid Mixture Properties...........................................................................................97 Vapor Mixture Properties...........................................................................................99 Equilibrium Conditions..............................................................................................99 Computer Implementation........................................................................................101 Saturation Temperatures....................................................................................101 Enthalpy.............................................................................................................102 Entropy..............................................................................................................103 Specific Volume................................................................................................105 B MODEL FORMULATION......................................................................................106 Thermodynamic Formulations..................................................................................106 Computer Implementation........................................................................................110 Saturated Liquid Pressure..................................................................................111 Two-Phase Mixture Determination...................................................................112 Saturated Liquid Concentration.........................................................................115 Saturated Vapor Concentration.........................................................................116 Two-Phase Mixture Enthalpy............................................................................117

PAGE 7

vii Temperature Determination Using Enthalpy.....................................................118 Isentropic Temperature Determination..............................................................119 Overall Cycle Calculation.................................................................................120 C EXPERIMENTAL DETAILS..................................................................................127 Instrument Settings...................................................................................................127 Data Acquisition System...................................................................................127 Gas Chromatograph...........................................................................................127 Uncertainty of Direct Measurements........................................................................128 Temperature.......................................................................................................129 Pressure..............................................................................................................130 Volume Flow Rate.............................................................................................130 Concentration....................................................................................................131 Shaft Speed........................................................................................................132 Uncertainty of Derived Measurements.....................................................................133 Vapor Concentration.........................................................................................133 Mass Flow Rates................................................................................................134 Power Output.....................................................................................................135 Expander Efficiency..........................................................................................135 Equipment Specification...........................................................................................136 Instrumentation..................................................................................................136 Expander Details...............................................................................................137 D EXPANDER AIR TESTING...................................................................................140 Experimental Setup...................................................................................................140 Test Results...............................................................................................................143 Experimental Details................................................................................................146 Measurement Uncertainties...............................................................................147 Equipment Specification...................................................................................148 REFERENCES................................................................................................................149 BIOGRAPHICAL SKETCH...........................................................................................155

PAGE 8

viii LIST OF TABLES Table page 3-1 Flow identification for the configuration of Figure 3-1...........................................18 4-1 Fluid properties for a ty pical ammonia-water concentr ation and other power cycle fluids for isentropic expansion from saturated conditions at 100 C to condensation/absorbtion at 35 C............................................................................34 4-2 Single stage specific speed calculation s versus nominal work output and shaft speed........................................................................................................................36 4-3 Approximate specific speed and specif ic diameter ranges for efficient (>60%) single stage expander types [49, 50].......................................................................37 4-4 Reported turbine operating parameters and efficiencies for three systems using an ammonia-water working fluid [54-56].....................................................................38 4-5 Estimated operating data for the three turbine stages of a Kalina-based bottoming cycle [53]. .................................................................................................................39 4-6 Reported efficiencies of sc roll expanders [18, 20, 62, 63].......................................40 6-1 Measured decrease of absorption pressu re with basic solution concentration.........61 6-2 Measured data indicating eff ects of absorption temperature....................................63 6-3 Averaged values fo r rectifier operation....................................................................63 6-4 Values for rectifier operation high lighting penalty to work production..................64 6-5 Averaged conditions for the testing of Figure 6-4...................................................66 7-1 Typical operating characteristics for cooling and work optimized cycles...............75 A-1 Coefficient and reference stat e values for ammonia and water................................96 A-2 Reference values for reduced property computation................................................96 A-3 Coefficient values used to compute excess properties.............................................98

PAGE 9

ix A-4 Coefficient values for the determ ination of mixture bubble and dew point temperatures...........................................................................................................100 B-1 Flow identification for the configuration of Figure B-1........................................107 C-1 Calibration factors for the ther mocouples used in this work..................................129 C-2 Pressure transducer calibration factors...................................................................130 C-3 Stated uncertainties for pressure transducers.........................................................130 C-4 Derived measurement uncertainty summary..........................................................136 C-5 Detailed descriptions of the instrument ation and equipment used for this work...136 D-1 Thermal and torque-based measurement uncertainties..........................................147 D-2 Summary of the equipment and co mponents used for the air tests........................148

PAGE 10

x LIST OF FIGURES Figure page 1-1 Ideas for using an ORC to incorporate a renewable element in to distributed power generation...................................................................................................................3 2-1 Schematic of the power-cooling cycle.....................................................................11 2-2 Ammonia-water phase equilibrium diag ram highlighting the source of cooling temperatures.............................................................................................................12 3-1 Power-cooling schematic used for modeling...........................................................19 3-2 Conceptual relationship between the factors affected by boiling pressure..............21 3-3 Output parameter variation as a function of boiling pressure..................................22 3-4 Variation of output parameters as a f unction of basic solution concentration.........23 3-5 Effect of boiler exit temperat ure on output parameter profiles................................24 3-6 Computed effect of vapo r concentration and inlet te mperature on expander exhaust temperature...............................................................................................................26 3-7 Computed effect of expander efficiency and inlet temperature on expander exhaust temperature...............................................................................................................26 3-8 Beneficial effect on expander exhaust temperature as a function of increasing rectification...........................................................................................................28 3-9 Effect of rectification and r ectifier efficiency on work production..........................29 5-1 Schematic of experimental setup..............................................................................47 5-2 Photograph of e xperimental setup............................................................................48 5-3 Modified turbine used for experimental testing.......................................................51 5-4 Original and modified absorber configurations.......................................................53 6-1 Measured effect of pressure variat ion on vapor quantity and concentration...........60

PAGE 11

xi 6-2 Measured effect of basic soluti on concentration on vapor production....................61 6-3 Measured change in vapor flow rate (relative to basic solution flow) due primarily to changes in boiling temperature............................................................................62 6-4 Experimental measurement of the expans ion of vapor to temperatures below those at which absorption-condensation is taking place....................................................65 6-5 Expected equilibrium exhaust qualities for conditions similar to those of the experimental study...................................................................................................68 6-6 Temperature-enthalpy diagram covering the phase change of pure ammonia and a high concentration amm onia-water mixture............................................................69 6-7 Comparison between the measured no-l oad power consumption of operation with compressed air and ammonia-water.........................................................................70 7-1 Maximum effective COP values where th e work component is the amount of work lost due to operation with rectificat ion vs. equivalent conditions with no rectification...............................................................................................................74 7-2 Maximum overall effective COP values as defined by Equation 7-2......................76 7-3 Corresponding exhaust temperatures fo r the optimum conditions presented in Figure 7-2.................................................................................................................76 7-4 Design point map showing the relative sens itivity of overall eff ective COP to vapor mass flow fraction and exhaust temperature............................................................78 7-5 Effect of the minimum and maximum bounds of rectifier operation on effective COP values...............................................................................................................80 7-6 Computed effect of weak solution storage on basic solution concentration............82 7-7 Computed absorption pressures taking in to account the change s of basic solution concentration compared with measured absorber pressures....................................83 7-8 Measured drop in boiling pre ssure due to rectifier operation..................................84 7-9 Amount of the produced vapor that was c ondensed in the rectif ier. This data corresponds to the results of Figure 7-8...................................................................85 B-1 Schematic used for the theoretical modeling.........................................................107 C-1 View of the assembled rear housing......................................................................138 C-2 Exploded view of the rear housing assembly.........................................................139 D-1 Setup schematic used for the air testing.................................................................141

PAGE 12

xii D-2 Rear view of expander with cover removed...........................................................141 D-3 Photograph of generator loading arrangement.......................................................142 D-4 Photograph of gearbox m ounted on expander spindle...........................................143 D-5 Air testing results comparing the value of power that was computed by the thermalbased and torque-based measurements..................................................................144 D-6 Comparison of the difference between th e power measurements of Figure D-5 and the no-load power measurements, .....................................................................145

PAGE 13

xiii NOMENCLATURE A area under mV-time curve COP coefficient of performance cp constant pressure specific heat D diameter, [m] dspec specific diameter parameter G Gibbs free energy h enthalpy m m mass flow rate, [kg/s] nspec specific speed parameter ORC organic Rankine cycle OTEC ocean thermal energy conversion P pressure, [MPa] Pr pressure ratio q heat transfer, [kJ/kg] Q heat transfer, [W], volume flow Qact actual volume flow rate Qexit expander exit volume flow rate, [m3/s] Qind indicated volume flow rate R universal gas constant Rc/w ratio of cooling to work

PAGE 14

xiv s entropy T temperature, [C] W work output/input, [W] x ammonia mass concentrati on [kg/kg], mixture quality y fraction of working fluid th at is stored in separator hideal isentropic enthalpy change Subscripts 1stLaw first law formulation a ammonia component properties absorber absorber parameter actual actual parameter values B reference properties basic new new basic solution parameter with storage basic org original basic solution parameter before storage boiler boiler parameter bubble saturated liquid properties cal calibration parameter values cool cooling heat exchanger parameter crit-water water crit ical point properties dew saturated vapor properties effective effective value formulation exit expander exit condition expander expander parameter

PAGE 15

xv float float properties inlet expander inlet condition m mixture properties no rect conditions with no rectifier operation overall overall comparison, with c ooling production and work optimized pump pump parameter r reduced properties recovery recovery heat exchanger properties s strong solution, isentropic end state superheat superheater parameter v vapor vr rectified vapor w weak solution, water component properties wb weak solution from boiler w/cool conditions w ith cooling production wr weak solution from rectifier with rect conditions with rectifier in operation workopt conditions optimized for work output 0 reference state properties Superscripts E excess properties g gas phase properties l liquid phase properties mix parameters of mixing

PAGE 16

xvi Greek efficiency specific volume density angular velocity, [rad/s]

PAGE 17

xvii Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy STUDY OF COOLING PRODUCTION WITH A COMBINED POWER AND COOLING THERMODYNAMIC CYCLE By Christopher Martin December 2004 Chair: D. Y. Goswami Major Department: Mechanic al and Aerospace Engineering This work is an investiga tion of a novel concept to pr oduce power and cooling with the energy contained in low-temperature (< 200 C), thermal resources. These resources can be obtained from non-concentrating so lar thermal energy, low-grade geothermal resources, and a near infini te variety of waste heat sources. The concept under investigation uses thermal energy in a lo w-temperature boiler to partially boil an ammonia-water working fluid mixture. This pr oduces an ammonia rich vapor that drives an expander. The expanders output is mechanical power; however, under certain operating conditions its exhaust can be cold enoug h to use for cooling. This possibility is the focus of the present study. An analytical study is presented which identifies expander efficiency, expander inlet conditions, and exhaust pressure as the factors determining exhaust temperature. Estimated expander efficiencies are based on a consideration of the operating conditions and a review of current tec hnology. Preferred inlet conditio ns are identified; however,

PAGE 18

xviii they are linked to the overall operation of the cy cle, as is absorption pressure. An optimal balance between vapor generation and expande r exhaust temperature is found for cooling production. Purifying the vapor is shown to enhan ce cooling production, but it penalizes work output. A new coefficient of performance is de fined as the ratio of the cooling gained to the work output lost and is used to dete rmine the optimal purification. Additionally, another performance coefficient is defined a nd used to judge the overall value of cooling produced. An experimental study is pres ented that verifies the predicted trends. Furthermore, a measurement of sub-ambient exhaust temper atures is provided that demonstrates the key concept of this cycle. It is concl uded that with improved expander performance, practical power and cooling produ ction can be achieved with this concept. Deviations between measured and simulated performance ar e discussed as they relate to improving future modeling and system design efforts.

PAGE 19

1 CHAPTER 1 INTRODUCTION The conversion of thermal energy into mech anical work is a fundamental task of mechanical engineering. Performing it clean ly, cheaply and efficiently all influence the eventual conversion scheme. Rankine-based cycles enjoy widespread usage and are particularly suited for low resource temperat ures since their operation can approximate that of a Carnot engine. Many adaptations and modificat ions have been made to the basic Rankine cycle in order to extract the most energy from heat sources such as geothermal wells, solar thermal energy, and waste heat streams. A re latively recent cycle has been proposed in which thermal energy is used to produce work and to generate a sub-ambient temperature stream that is suitable for cooling applications [1]. It has been the focus of theoretical and experimental investigation [2-5]; howev er, until this work, there has not been a complete, experimental implementation of th is power-cooling cycle. Therefore, this study is an investigation, bot h theoretical and experiment al, into the distinguishing feature of this concept, which is cooling production. The cycle is a combination of Ranki ne power production and absorption refrigeration cycles, and is unique in power production cycles because it exploits the temperature drop across an expande r to the point of being able to obtain useful cooling. Optimization of system parameters, wo rking fluid selection, and preliminary experimentation with the cycle have been perf ormed [2-5]. What this work provides is an experimental proof-of-concept that demonstr ates the key feature of this cycle. Also

PAGE 20

2 included is a discussion of the parameters affecting cooling production and a method to quantify its production. Certainly this work advances the devel opment of this power-cooling cycle, but more importantly, research in the field of heat recovery is an active step in moving society toward a sustainable energy policy. Th is concept belongs to the broader class of low temperature, Rankine based cycles which have been shown to be one of the most effective means for utilizing low temperature resources. They have been applied to the production of mechanical power using heat from solar, geothermal, and waste heat from topping power cycles and industr ial processes. Despite th eir wide range of possible applications, these systems have found limited success in practice. It is hoped that this work will aid any resurgence in todays energy market. Motivation The wide-ranging motivation for this work comes from the possible applications for this category of low temperature, Rankine based cycles. Being simply a heat engine with the potential for good second law efficien cies, the possible recove ry applications are limited only by the economics of the situation. In the future, the economics may be more favorable to devices that can produce power without additional resource consumption. When considering the future of world electricity production, the only apparent certainty is that generation will be done by more diverse means than it is currently [6]. It appears that the paradigm of a few, large, centralized power producers is becoming more conducive to adding more, smaller, distributed generators. There are many reasons for this; key among them are to increase the reliabi lity of the electrical system by promoting diversity, provide cleaner energy by incorpor ating more renewables, and simply to increase capacity to meet additional demand.

PAGE 21

3 The distributed generation trend will open opportunities in two key ways. First, by adding smaller distributed generators, the m echanics of connecting to the grid will no longer be prohibitively complex, but will beco me more routine. Second, with on-site generation the opportunities to recover and use waste heat resources will make economic sense. In fact, the U. S. Department of En ergy expects the utilization of waste heat alone to provide a significant source of pollutionfree energy in the coming decades [7]. Viewed in this way, the use of a low-te mperature power cycle is one of the many possible distributed technologies that could connect to the gr id or be used to recover thermal resources. They can be used on a small scale to convert renewable energy sources, use conventional fuels efficiently, or conserve energy by recovering waste heat from energy-intensive processes, Figure 1-1. Ultimately they would have positive impacts on overall energy conversion effici ency and could be us ed to incorporate renewable energy sources. Adaptable ORC Conventional Fuel Topping Cycle Solar Thermal Waste Heat Combustion Power Figure 1-1. Ideas for using an ORC to incor porate a renewable element into distributed power generation. Efficient use of multiple energy sources would require a highly adaptable heat engine and, of c ourse, any configuration would have to be economically viable.

PAGE 22

4 Power-Cooling Concept This work is not directly aimed at reduci ng the cost of this te chnology. Rather it is directed at improving the underlying science to make it more versatile and thus more attractive for implementation. Mechanical power is one useful form of energy, the generation of low temperatures for cooling or refrigeration is anot her. The cycle under study in this work was intended to explore th e feasibility of using thermal resources to simultaneously produce these two useful outputs. Put simply, the configuration of this po wer-cooling cycle allo ws the vapor passing through the turbine to be expanded to below am bient temperatures. Cooling can then be obtained by sensible heating of the turbine exha ust. A more detailed explanation of this process follows in Chapter 2, but here it suffi ces to say that the use of a working fluid mixture, ammonia-water, is the key to this process. Just as in conventional aquaammonia absorption cooling, abso rption-condensation is also us ed here to regenerate the working fluid. This eliminates the expansi on temperature restriction which is in place when pure condensation is used. The power-cooling cycle has the obvious a dvantage of two useful outputs, but it has other attributes that make it an attractiv e energy conversion option. The first of these characteristics is that the cycle uses a bina ry working fluid that has a variable boiling temperature at constant pressure. This avoi ds heat exchange pin ch point problems that pure component working fluids experience due to their constant phase change temperature at constant pressure. In a ddition, turbine designs for ammonia-water are reasonably sized for large power outputs when compared to the more traditional organic working fluid choices.

PAGE 23

5 Problem Definition The distinguishing feature of this cycle, compared to other power cycles and even those in the developing class of combined power and cooling cycles, is the method in which cooling is produced. In other power cycles the working fluid is regenerated by pure condensation, rather than ab sorption-condensation which is used here; this limits the minimum turbine exhaust temperature to rough ly the temperature at which condensation is taking place. When considering other combined power-cooling cycles, cooling is typically produced in the same manner as a conventional absorp tion system, that is, condensation and throttling of th e refrigerant. Here in th is cycle, vapor is expanded through a turbine to produce power and b ecause of the advantage of absorptioncondensation, it can be expanded to sub-ambient temperatures. While the method of cooling production is th e key feature of this cycle, until this work it has not been experimentally inves tigated. What has been experimentally investigated are the underlyi ng boiling and absorptions pr ocesses [4]. For those experiments a turbine was not implemented; its performance was simulated with an expansion valve and a heat exchanger. Coupl ed with the lack of experimentation, the question of implementing cooling production with this concept has not been treated in any depth. Discussion of the possible uses of this cycl e have suggested the utilization of solar thermal, geothermal, or waste heat resour ces. However, a proper use for the potential cooling output has not been put forward, po ssibly because the specific nature of an application will be determined by the characte ristics of cooling production. There has not been a thorough discussion of th e trends of cooling production.

PAGE 24

6 Research Objectives In response to the deficits mentioned in th e previous section, the objectives of this work are to experimentally implement a tu rbine for power production and identify the factors important for cooling production and investigate them analytically and experimentally. Analytically, the study will identify the conditions favorable for cooling production, estimate performance using availa ble expander technologies, and quantify cooling production in terms of energy c onsumption. In addition, this work will experimentally investigate the concepts ke y for cooling production and document design and operating experience for use with future modeling or implementation efforts.

PAGE 25

7 CHAPTER 2 BACKGROUND AND REVIEW The purpose of this chapter is to introduce the concept of this cycle in the context of both low temperature thermodynamic power cycles and conventional cooling cycles. In addition, to accurately provide the context fo r this work, a review of previous effort into this concept is presented. As an overview, the power-cooling cycle is best described as a compromise between a conventional aqua-ammonia abso rption system and a Kalina-type power generation cycle. It is a continuation of th e evolution of binary mixture Rankine cycles but makes use of the cooling effect possibl e due to the working fluid concentration change. As for previous work on this cycl e, numerous theoretical studies have been produced and initial experimentation has begun. From a review of that work, this study is shown to be the first experimental c onfirmation of the power-cooling cycles key concept and to provide initial c onsideration for system operation. Background The thermodynamic conversion of low temperature resources into mechanical power traces its roots to at least the beginning of the industr ial revolution. Utilizing solar thermal energy to pump water was the impetu s and this work continued haphazardly until the early decades of the twentieth century when it was interrupted by World War I and the discovery of a new resource, oil and gas [8 ]. Modern research into low temperature power conversion surfaced again when the pa nacea of cheap coal power was beginning to break down and energy alternatives were sought in the decades follo wing World War II.

PAGE 26

8 The application was utilizing liquid-producing, geothermal fields where flash boiling is not suitable [9]. Additional interest came during the 1970s oil crisis in using these low temperature engines for solar thermal ener gy and heat recovery applications. The common description for these systems is orga nic Rankine cycle (ORC) engines, because many of the working fluids are orga nic hydrocarbons or refrigerants. ORC Development Intense research of non-geothermal ORC use took place in this country during the early 1970s through the early 1980 s. ORC heat engines were reconsidered for utilizing solar resources and conserving other resource s by recovering energy from waste heat. Seemingly no application was overlooked as a few innovative examples illustrate. In one an ORC was integrated with a large truck engine to recove r heat from the exhaust and save on fuel costs [10] and in another app lication the idea of replacing the automobile internal combustion engine with an ORC system was explored [11]. Mechanical cooling systems were one of the more productive research areas that dealt with the conversion of solar thermal energy. A significant amount of the published literature regarding ORC convers ion of solar thermal energy comes from this and related work [12]. The concept started as an alternative to solar-driven, absorption, airconditioning cycles which have a limited co efficient of performance. Essentially, mechanical work produced by a solar-driv en ORC would be used to drive vaporcompression air conditioning equipment, w ith the potential of a higher COP than absorption equipment [12]. These projects produced many successful prototype units ( e.g. [13] ) and led to a fee ling of technical maturity fo r the low-temperature, smallscale, conversion of solar thermal energy [14]. As fo r ORC technology today, it has found some niche successes in geothermal utili zation, biomass utilization, some industrial

PAGE 27

9 heat recovery, and cathodic protection of pipelines, as judged by a few manufacturers portfolios. More recent research in the area has larg ely taken place intern ationally ( e.g. [1518] ), with much interest being placed on expander implementation. Two approaches have been noted: one is to deve lop and design systems around high-speed turbomachinery with a shaft integral generato r and circulation pump [19], thus reducing costs by simpler design, and the other, more recent idea is to adapt mass-produced (cheap) displacement compressors for use as reasonably efficient expanders [18, 20, 21]. Ammonia-Water Cycles While much of the related material on ORC systems is intended for small scale application or implementation (especially sola r-driven units), the li neage that the powercooling cycle is derived from was initially intended for utility-scale bottoming cycle duty. The first study of an absorption based power cycle was performed by Maloney and Robertson [22] who concluded no significant advantage to the conf iguration. Several decades later, Kalina [23] reintroduced the idea of an ammonia-water power cycle as a superior bottoming cycle opti on over steam Rankine cycles. Some independent studies have been performed [24, 25] that concede some advantage of the Kalina cycle under certain conditions. The key advantage of the ammonia-water working fluid is its boiling temperature glide, which allows a better thermal match w ith sensible heat sources and reduces heat transfer related irreversibilities. This sa me advantage, however, could be a problem during the condensation phase of the cycle in which the c ondensation temperature glide could cause a thermal mismatch with the heat rejection fluid and an increase in heat transfer irreversibility. The so lution employed is to vary th e concentration of the working

PAGE 28

10 fluid so that the fluid passing through the tu rbine is of different composition than that being condensed in the condenser. In fact, by taking advantage of the chemical affinity of ammonia and water, the condensation process can be replaced by absorptioncondensation. Investigation of these power cycle configurations has become a new specialty in engineering thermodynamics, and the power-coo ling cycle of this work is a product of this research area. As a result of this di versified interest, ammonia-water based power cycles have been proposed for solar util ization, geothermal, ocean thermal energy conversion, and other forms of heat recovery. Power-Cooling Concept While it was the interest brought about by Kalinas proposal that led to the introduction of the power-cooling cycle, it is somewhat ironic that the original suggestion for its implementation is more similar to the original Maloney-Robertson implementation [1, 2]. Figure 2-1 is a schematic of the pow er-cooling cycle. Aside from the operating parameters, the key difference between th e cycle of Figure 2-1 and the MaloneyRobertson cycle is the addition of a vapor rectifier following the boiler. Referring to Figure 2-1, basi c solution fluid is drawn from the absorber and pumped to high pressure via the solution pum p. Before entering the boiler, the basic solution recovers heat from the returning weak solution in the recovery heat exchanger. In the boiler, the basic solution is partia lly boiled to produce a two-phase mixture; a liquid, which is relatively weak in ammonia, and a vapor with a high concentration of ammonia. This two-phase mixture is separate d and the weak liquid is throttled back to the absorber. The vapors ammonia concentr ation is increased by cooling and condensate separation in the rectifier. H eat can be added in the superhea ter as the vapor proceeds to

PAGE 29

11 the expander, where energy is extracted from th e high-pressure vapor as it is throttled to the system low-pressure. The vapor rejoins th e weak liquid in the absorber where, with heat rejection, the basic solution is regenerated. Rectifier Boiler Solution Pump Heat Out Refrigeration Heat Exchanger Throttle Recovery Heat Exchanger Absorber Superheater Expander Heat In Figure 2-1. Schematic of the power-cooling cycle. In this configuration, the vapor temp erature exiting the expander can be significantly below ambient conditions and coo ling can be obtained by sensibly heating the expander exhaust. The temperature dr op possible across the e xpander is due to the fact that the working fluid is a binary mixture, and at c onstant pressure the condensing temperature of an ammonia rich vapor can be below the saturation temperature for a lower concentration liquid. This is best illustrated with a binary mixture, phase equilibrium diagram, as shown in Figure 22. The low concentration saturated liquid

PAGE 30

12 state represents the basic solution exiting the absorber, while the high concentration vapor is typical of the expande r exhaust conditions. This sh ows how it is possible for the vapor to be expanded to a temperature below that at which absorpti on is taking place. According to the equilibrium diagram, to maximize this temperature difference the basic solution should be low in ammonia concentr ation and the vapor should be high. Also, partial condensation of the expander exhaust wo uld cause an additiona l decrease in vapor temperature. This is entirely possible si nce ammonia becomes saturated upon expansion. -20 0 20 40 60 80 100 120 00.10.20.30.40.50.60.70.80.91Ammonia Mass FractionTemperature [C]Pressure = 0.203 MPa Vapor Liquid Two-Phase Basic solution in absorber Expande r exhaust Figure 2-2. Ammonia-water phase equilibri um diagram highlighting the source of cooling temperatures. Prior Work Since the proposal of the idea by Gosw ami a theoretical and experimental investigation has been under way by a group at the University of Florida. Initial investigations were performed theoretically and they focused on procuring reliable

PAGE 31

13 property data for the ammonia-water mixtur e [26] and identifying operating trends [27, 28]. Later studies concluded that the cycle could be optim ized for work or cooling outputs and even for efficiency. Optimizati on studies began to appear, optimizing on the basis of various efficiency definitions, minimum cooling temper ature, working fluid combination, and system configuration. Als o, an experimental study was described by Tamm and Goswami [4] which ge nerally verified the expected boiling and absorption processes. Goswami and Xu [27] presented the first theoretical analysis of the power-cooling cycle. Turbine inlet temperatures of 400 500 K were considered along with absorption temperatures of 280 320 K. Cooling producti on suffered with increased turbine inlet and absorption temperatures, and benefited w ith increased boiler pr essure. Many of the operating trends of importance in this work were introduced here. Optimization studies began to appear following this work, which identified the balance of effects that dictat e cycle operation. Lu and Goswami [2] optimized the ideal cycle conditions using various objectives, work output, cooling output, first and second law efficiencies. All operating parameters, efficiencies, power/cooli ng output, etc., were found to decrease with increas ing heat rejection temperat ures. At high heat source temperatures, 440 K, no cooling was possibl e at conditions optimized for second law efficiency. A contrast be tween work optimized and cooling optimized cases was provided. Important differences in the cooli ng optimized case versus the work optimized one were higher vapor concentration, lower turbine inlet temperature, low vaporization fraction (16.5 % vs. 91.2 %), a nd a lower basic solution conc entration. Minimum cooling

PAGE 32

14 temperatures were also optimized [29], a nd a minimum turbine exhaust temperature of 205 K was identified under the assumptions considered. The question of appropriate efficiency expressions for the cycle was tackled by Vijayaraghavan and Goswami [30]. The condi tions obtained from an optimization study were found to be heavily influenced by the we ight given to the cooling output. Some expressions simply added the outputs of pow er and cooling, which gives an overestimate of system performance, or cooling was wei ghted by an ideal COP value computed for equivalent temperature limits, which tends to underestimate the valu e of cooling. They [30] introduced a satisfactory second law efficiency defini tion based upon ideal Lorenz cycle performance which accounts for sensible heat addition and rejection behavior. However, they concede that ultimately the va lue of work and cooling will be decided by the end application [30]. Both first and second law efficiency anal yses were performed for the cycle [31, 32]. A second law efficiency of 65.8 % was de termined, using the definition of [30], for the idealized model considered. The largest s ource of irreversibility was found to be the absorber at all conditions considered; while at higher heat source temperatures the rectifier also cont ributed significantly. Less-than-ideal modeling began with Tamm et al. [33, 34], in preparation for the initial experimental studies [4]. The larges t deviation from idealized simulations was due to the non-isentropic performance of the turbine. This relates well to the findings of Badr et al. [35], who identified the expander isentropic efficiency as the single-most influential factor affecting overall ORC e ngine performance. Initial experimentation was reported [36]; however, turbine operation was simu lated by an expansion valve and a heat

PAGE 33

15 exchanger. General boiling condition trends were demonstr ated, for example vapor mass flow fraction, vapor concentra tion, and boiler heat transfer Vapor production was less than expected and improvements to the se tup were identified and implemented. Performance of the new configur ation, still having a simulated turbine, was also reported [4]. Vapor production and absorption proce sses were shown to work experimentally, however still with some deviations. An independent study of the power-cooling concept has been provided by Vidal et al. [37], who also noted the significant im pact of non-ideal tu rbine performance on cooling production. Vidal et al. also reported poor cooli ng production at higher ambient conditions. Other Power-Cooling Concepts The development of the power-cooling cycle under investigation in this work has been presented as it relates to other powe r production cycles. However, there is now a small class of combined power and cooli ng cycles, especially since the proposal by Goswami [1]. Differentiation of this con cept from others in the literature is now presented. Oliveira et al. [38] presented experi mental performance of an ORC-based, combined power-cooling system which used an ejector placed in-par allel to the turbine for cooling production. Ejector cooling has been an academic topic for solar thermalpowered cooling, for example [39, 40]. The implementation and operation of an ejector cooling system is quite simple and rugged; how ever, its COP tends to be low and in this combined case it siphons away high pressure va por directly from the turbine that could have been used to produce power.

PAGE 34

16 Considering integrated, ammonia-water cy cles, Erickson et al. [41] present the most intuitive. The proposal is essentially an absorption cycle, with advanced thermal coupling between the absorber and generator, with a turbine placed in-parallel to the condenser and evaporator. So vapor is produced and, depending on the outputs desired, split between expansion in a turbine or conde nsation and throttling in the refrigeration path. Integration comes from the common components, for example the absorber, generator, and feed pumps. However, the m echanism of cooling is the same as that for an aqua-ammonia absorption system. The very pure ammonia vapor is condensed at high pressure and throttled to the absorption pre ssure where flash boiling and evaporation take place. The concept of parallel paths for power and refrigeration pr oduction has been incorporated with the thermal-matching concep ts of a Kalina cycle by a research group at Waseda University [42]. Unlike the pr oposal by Erickson et al., however, only the working fluid is shared between the two systems. The power production and refrigeration cycles can be driven independently, but it wa s found that more power could be produced by sharing the working fluid [42]. Therefore, cooling in this case is also produced in the same manner as with an aqua-ammonia absorption system. A more thorough integration of power and refrigeration production has been recently proposed by Zhang et al. [43]. In this configuration the ammonia-water basic solution is separated into a high concentration ammonia vapor and a relatively weak solution liquid in a device similar in operati on to a distillation column. The vapor is condensed and throttled to produce cooling wh ile the weak solution liquid is vaporized and superheated, then expanded in a turbine for power production. The streams are then

PAGE 35

17 cooled and rejoined in an absorber. Th e authors claim a 28% increase in exergy efficiency of this arrangement over se parate steam Rankine and aqua-ammonia absorption systems [43]. As with the other concepts, cooling is produced the same way as with a aqua-ammonia absorption system. Conclusion As compared to other power and cooling c oncepts, the distinguishing feature of this cycle is the method in which cooling is produced. Its configuration is most similar to that of an aqua-ammonia absorption system; how ever, instead of using condensation and throttling for cooling production an expander is used to extract energy from the vapor--to the point that cooling can be obtained from th e exhaust. As compared to the absorption cycle, the trade-off for work production is reduced cooling since no latent heat is involved. The remainder of this work will discuss the opposite situation, the penalty to power cycle operation due to combined cooling production.

PAGE 36

18 CHAPTER 3 THEORETICAL STUDY In this chapter a model of the system is presented and used to simulate the steady state performance of the powe r-cooling cycle. With this model, a straightforward parametric study is carried out which iden tifies the essential operating mechanisms affecting cooling production. Th ese results are used to design the experiments discussed in Chapters 5 and 6 and again to extrapolat e the data used for the final conclusions. Model The model used for this work is based upon the schematic of Figure 3-1 which has subtle differences from the one in Figure 2-1 to be more representative of the experimental system. Table 3-1 contains the identifying information for the working fluid streams in Figure 3-1. Table 3-1. Flow identification fo r the configuration of Figure 3-1. Identifier/ Subscript Description s Basic (strong) solution flow from absorber through boiler v Vapor flow produced from pa rtial vaporization in boiler vr Rectified vapor passing through tu rbine and cooling heat exchanger w Weak (in ammonia) solution liquid returning to absorber wr Weak condensate formed in rectifier wb Weak liquid produced from pa rtial vaporization in boiler For the purposes intended, it was adequate to use first order approximations for each component, conservation of mass and energy, so detailed component modeling was not included. The complete formulations that were used in the computations, as well as the subroutines themselves, can be found in Appendix B; however, the key points are summarized as follows.

PAGE 37

19 Heat Source Recovery Heat Exchanger Throttle Boiler Absorber Coolant Separator Coolant Rectifier Expander vr Solution Pump v wb w wr s Superheater Heat Source Cooling Heat Exchanger Cooled Fluid Figure 3-1. Power-cooling schematic used for modeling. The boiling conditions are completely specifi ed, i.e. boiling temperature, pressure, and basic solution concentrati on are provided as inputs. Th is means that the quality at boiler exit is allowed to be determined in accordance. The system low pressure is dictated by the basic solution concentration and the minimum absorption temperature, both of which are specified. Isentropic efficiencies are assumed for the pump and expander while effectiveness values are used for heat exchangers. The degree of rectification is determ ined by specifying the rectifier exit temperature. Similarly, the superheater exit temperature is also specified. In addition to the specifications above, which are needed to determine the steady state conditions, the following stipulations were enforced to avoid computational problems and/or make the s cenarios closer to reality. The minimum absorption temperature consid ered was 25 C with most attention given to 35 C cases.

PAGE 38

20 Vapor rectification was limited by either th e specified rectifier exit temperature or an ammonia mass fraction of 0.999, whiche ver was encountered first. The minimum rectification temperat ure considered was 35 C. The minimum amount of vapor leaving the rectifier that was allowed was 5 % of the basic solution flow rate. The quantity of cooling produced (if any) was calculated as the energy needed to heat the expander exhaust from the exhaust temperature to 15 C. Thermodynamic property data for the ammoni a-water working fluid is essential for this type of modeling. The correlations us ed are based on those presented by Xu and Goswami [26] which are a combination of the Gibbs free energy method for mixture properties and empirical equations of bubbl e and dew point temperatures for phase equilibrium. Details of the complete corre lations and their implementation into C++ can be found in Appendix A. Operating Mechanisms Early in the theoretical investigation of th is cycle it was determined that the system could be optimized for various outputs [27]. In this section, simulated data is used to illustrate these optimums and th e balance of effects that cau ses them. Boiling conditions are considered first and then the e ffects of heat rejection conditions. First consider the independent effects of the parameters at the heart of the powercooling cycle, the boiling condi tions. These effects are common to all binary mixture power cycles; however, it will be shown that they have added significance for cooling production. Effect of Boiling Pressure The boiling pressure in the power cycle is regulated by the rate of vapor production and the rate at which vapor is released th rough the restriction impos ed by the expander. For a binary mixture working fluid at c onstant temperature and having constant

PAGE 39

21 composition boiling takes place over a range of pressures from the saturated liquid state to the saturated vapor state. At the uppe r extreme boiling pressure is limited by the corresponding saturation pre ssure, above which no vapor is produced. The lower pressure extreme is bound by the system low pressure or the absorption-condensation pressure. Depending on the conditions, the working fluid may or may not be fully vaporized at the lower pressure extreme. Figure 3-2 is a graphical re presentation of the mechanis ms of variable pressure boiling. As can be seen the mass flow rate of vapor changes inversely with pressure ratio. Also, a quantity like the work output (assuming constant efficiency expander), which is dependent on both the pressure ra tio and amount of vapor flow, contains a maximum within the boiling region. Work production is bounded by a unity pressure ratio at low boiling pressure and zero va por flow at the highest boiling pressure. 0 10 20 30 40 50 60 -0.100.10.20.30.40.50.60.70.80.911.11.2 Boiling Pressure [MPa]mv, Pr, W-10 -5 0 5 10 Vapor Mass Flow Pratio Work Output Absorption P higher than boiler P Boiler P too high, no vaporization PabsorberPbubble Figure 3-2. Conceptual relationship between the factors affected by boiling pressure.

PAGE 40

22 As a further example, Figure 3-3 presents co mputed results also for the variation of boiling pressure. The relative position of the maxima for work production, first law efficiency, and cooling production are shown. Similar to work output, cooling also has a maximum, which is limited at low pressure s by higher turbine exhaust temperatures and bounded at higher pressures by the low production of vapor. 0 5 10 15 20 25 0.10.20.30.40.50.60.70.80.91Boiling Pressure [MPa]Work/Cooling Output [kJ/kg]0 1 2 3 4 5 6 7 8Efficiency [%]Boiler Temperature: 80 C Strong Concentration: 40%Work Output 1st Law Efficiency Cooling Output Figure 3-3. Output parameter variation as a function of boiling pressure. Effect of Mixture Concentration Mixture concentration directly changes th e saturation temperatures and temperature glide of the working fluid. Similar to th e boiling pressure, it has upper and lower bounds which are illustrated in Figure 3-4. The lowe st possible concentra tion corresponds to the saturation concentration for the given temper ature and pressure. Concentrations lower than this will not boil at the specified temp erature. Also, the absorption pressure is determined in part by the basic concentration, therefore, the upper limit of concentration

PAGE 41

23 is reached when the corresponding absorber saturation pressure equals the boiling pressure. Similar to that with pressure variation, work producti on has a maximum which is governed by the extremes of zero vapor production and a pre ssure ratio of one. 0 5 10 15 20 25 0.20.30.40.50.60.70.80.9Basic Solution ConcentrationWork/Cooling Output [kJ/kg]0 1 2 3 4 5 6 7 8Efficiency [%]Boiler Temperature: 80 C Boiling Pressure: 0.7 MPaWork Output 1st Law Efficiency Cooling Output Figure 3-4. Variation of output parameters as a function of basic solution concentration. Effect of Boiling Temperature While the effects of basic solution concen tration and boiling pressure are intended for optimization of the system, boiling temper ature is considered largely dependent on the heat resource. However, adaptability to such changes has been identified as an important evaluation criteria due to the inhere nt heat source variabil ity associated with solar, waste heat [44], and even geothermal resources [45]. So the effect of boiling temperature is also considered here. Assumi ng that the temperature is above that needed for vaporization, there is no convenient upper bound on boiling temperature. The effect

PAGE 42

24 of interest here is shown in Figure 3-5, which is to essentially shift the trends that were presented in Figure 3-3. 0 5 10 15 20 25 30 35 40 45 0.10.30.50.70.91.11.31.5Boiling Pressure [MPa]Output [kJ/kg]0 2 4 6 8 10 12 14Efficiency [%] Work Output Cooling Output 1st Law EffBasic Concentration: 40% 80 C 60 C 100 C Boiling T Figure 3-5. Effect of boiler exit te mperature on output parameter profiles. Cooling Production As explained in Chapter 2, the difference in concentration between the vapor and basic solution streams allows the expander e xhaust to be below ambient temperatures. Sensible heating of this exha ust provides the combined cooling output. Therefore, when considering cooling production the exhaust te mperature and quantity of vapor both factor into the total amount. Initially, the sensitivity of each parameter is examined individually then total cooling pr oduction is discussed. Exhaust Temperature Some comments regarding mi nimizing the expander exhaust temperature were discussed in relation to the bi nary phase diagram of Chapte r 2, Figure 2-2. A reiteration

PAGE 43

25 of those conclusions is that the basic solution should be low in ammonia concentration and the vapor should be high. Th e total absorption pressure s hould be low. Also, partial condensation of the expander exhaust woul d cause an additional decrease in vapor temperature. To investig ate the effect of expander inlet vapor conditions on exit temperatures, consider the entropy of th e working fluid at the expander exit. Minimization of the exhaust temperature al so implies a minimization of the vapor entropy at expander exhaust, assuming constant exit pressure. From this consideration an efficient expander is an obvious feature for low temperatures, but even an ideal device would only maintain the vapor entropy from in let to exhaust. Therefore, expander inlet conditions should be considered. For the binary vapor mixture of ammoniawater, entropy decreases with increasing pressure, increasing concentration, and decr easing temperature. The limit of these conditions, while still maintaining vapor, woul d be saturated, pure ammonia. As an aside, aqua-ammonia absorption cooling cycles further reduce the throttle inlet entropy by condensing the vapor to a liquid. Figures 3-6 and 3-7 show the effects of inlet temperature, vapor composition, and expander efficiency on exhaust temperature. The intuitive effect of decreasing inlet temperature is shown in both figures. Figure 3-6 highlight s the sensitivity to vapor composition, while Figure 3-7 reiterates th e benefit of good expander efficiency. The effect of exhaust pressure has not been shown, however, lower temperatures are encountered with lowe r exhaust pressures.

PAGE 44

26 -5 0 5 10 15 37394143454749515355Expander Inlet Temp. [C]Expander Exit Temp. [C]Pinlet/Pexit = 0.516 MPa/0.208 MPa expander = 70% 0.995 0.997xvr= Figure 3-6. Computed effect of vapor c oncentration and inlet temperature on expander exhaust temperature. -15 -5 5 15 25 35 37394143454749515355Expander Inlet Temp. [C]Expander Exit Temp. [C]Pinlet/Pexit = 0.516 MPa/0.208 MPa xvr= 0.995 50 % 100 % 70 %expander= 30 % Figure 3-7. Computed effect of expander efficiency and inlet temperature on expander exhaust temperature.

PAGE 45

27 Relating this to the variati ons in boiling conditions men tioned previously indicates conditions preferable for cool exhaust temp eratures. The conditions closest to low temperature, high pressure, and high concen tration ammonia vapor occur at the limit where boiling is just beginning. Then assumi ng expander efficiency and exhaust pressure remain constant, the lowest exhaust temperatur es will occur at this leading edge and will increase as the fraction of vaporization of the basic solution increases. Vapor Flow Rate Vapor flow rate was disc ussed as it was affected by boiling and concentration changes. It was shown to vary from a maxi mum of 100% of the basic solution flow rate at low pressures and high concentrations to a minimum where boiling just starts at high pressures and low concentrations. From th e standpoint of cooling production, much like that for work production, the two determining fa ctors vary inversely wi th each other as a function of pressure. This creates a ma ximum for cooling production as shown in Figures 3-3 through 3-5. Rectification This section discusses conditioning the vapor stream before it enters the expander to enhance cooling production. This c onditioning involves c ooling and condensate separation to increase the vapor concentr ation. It is commonly employed in aquaammonia absorption refrigeration systems to pr event water build-up in the evaporator and it is termed rectification [46]. In the pow er-cooling cycle its purpose is to change the temperature and concentration of vapor to values more suitable for cool exhaust. For sufficiently low absorption pressures, rectification may not be necessary for cooling production; however, th is typically requires unreaso nably cool heat rejection

PAGE 46

28 (absorption) temperatures. For the temperatur es considered in this work, 25 and 35 C, some rectification is typically needed to reach exhaust temperatures below 15 C. The effect of rectification on expander e xhaust temperatures is twofold. First it lowers the expander inlet temperature of the vapor, which beneficially lowers exhaust temperatures. Second it increases the vapor concentration so that expansion can take place to lower temperatures before the dew point is encountered. The improvements to vapor concentration and expander exit temperat ure with rectifier exit temperature are shown in Figure 3-8. -10 0 10 20 30 40 50 60 70 80 020406080100120Rectifier Exit Temp. [C]Expander Exhaust Temp. [C]0.8 0.82 0.84 0.86 0.88 0.9 0.92 0.94 0.96 0.98 1Vapor Concentration [kg/kg] PB/PA = 0.5/0.2 MPa mvr/ms = 0.3expander = 30 % Concentration Exhaust Temp. Figure 3-8. Beneficial effect on expander e xhaust temperature as a function of increasing rectification (decreasing rect ifier exit temperature). A word regarding the operati on of the rectifier is appropriate here. There are many physical setups that can be used to purify th e vapor, with some being more efficient, in terms of purified vapor flow, than others. For this work, upper and lower limits to

PAGE 47

29 rectifier efficiency are considered. Th e upper bound is the theoretical maximum of rectified vapor that can be produced as determ ined from a mass balance of the rectifier. This could be implemented with a direct contact, counter-flow heat exchanger where additional ammonia is scavenged from the counter-flowing condensate. The lower bound, which represents the arrangement of th e experimental setup and the computer model, is the flow rate that occurs with simple cooling of the vapor and condensate separation. No attempt to recover ammonia fr om the condensate is made. The effects of this vapor-production efficien cy are considered next. The improvements to cooling from r ectification do not come without cost, however, and in cases where it is used, some potential work is sacrificed for cooling. Figure 3-9 shows an example of this sacrifice, which is a pl ot of normalized work output versus rectifier exit temperature. 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 020406080100120 Rectifier Exit Temp. [C]Normalized Power Output [kW/kW] Maximum Vapor Production PB/PA = 0.5/0.2 MPa mvr/ms = 0.3expander = 30 % Minimum Vapor Production Figure 3-9. Effect of rec tification and rect ifier efficiency on work production.

PAGE 48

30 The penalty to work production is caused by two factors, the decline in available energy from cooling and the reduction in mass flow rate due to condensate formation and separation. The combined effects of both it ems on normalized work output is shown in Figure 3-9 for the upper and lower pr oduction rates for the rectifier. As evident from Figure 3-9, at minimal amounts of rectification the difference between rectifier performan ce is small. However, with increasing amounts of rectification the difference is severe and it appears that inve stment in a more efficient device may be warranted. Performance Measures This section describes the efficiency parameters used to evaluate the relative performance of the power-cooling system. Work Production A measure of performance is needed to compare the relative efficiency of producing work with equivalent heat source c onditions and cycle conf igurations, and also to identify conditions for maxi mum work production for a give n set of heat source/sink conditions. For this purpose, a first law efficiency formulat ion is adequate, Equation 3-1. 1 expanderpump stLaw boilersuperheatWW QQ (3-1) Cooling Production There is some element of personal choi ce involved in including cooling in an efficiency definition. This comes from the options of converting co oling to equivalent work terms. Other work has discussed the me rits of adding work a nd cooling directly or weighting cooling with a COP based on ideal cycle performan ce [30]. In this chapter it has been shown that cooling and work optimum s generally do not coincide, that is to

PAGE 49

31 produce cooling some sacrifice in work has to be made. Based on this observation, an effective COP can be defined as the ratio of the cooling produced to the work that could have been produced, but was avoided to genera te cooling. In gene ral terms th e concept can be written as follows. effectiveCoolingGained COP WorkLost (3-2) This term has been called the effective COP since cooling and work are only indirectly related, in other words there is no device directly produc ing cooling with the work that is given up. This definition is simp ly a way to determine the effectiveness, in terms of energy, of cooling production with this cycle. It was mentioned that some rectification is needed to produce any cooling with the absorption temperatures considered. Recalling also that rectificat ion diminishes work production by the mechanisms of reduced mass flow and available energy, then some work is inevitably lost when cooling is pr oduced. Therefore, a more specific effective COP can be defined based on the need for r ectification, and is shown as Equation 3-3. cool effective norectwithrectQ COP WW (3-3) Conclusion This chapter has shown the operating mech anisms affecting cycle operation. These mechanisms determine the relative amounts of work and cooling production by affecting the balance of vapor production, expander pressure ratio, a nd expander exhaust temperature. Expander exhaust temperature is sensitive to the sensitive to the inlet vapor conditions (pressure, temperature, and concen tration), exhaust pressure, and expander efficiency. Considering the preferred inlet c onditions, the ultimate effect of the partial

PAGE 50

32 boiling and rectification proce ss would be to separate amm onia vapor from liquid water rather than generate high pressure vapor fo r power production. The trends identified in this chapter are used to guid e the experimental study describe d next chapter. Also, the implications that can be extrapolated fr om these trends are discussed during the conclusion of this work.

PAGE 51

33 CHAPTER 4 EXPANDER CONSIDERATIONS As shown in the previous chapter, effici ent operation of the expander is an obvious requirement for cooling production within the power-cooling cycle. The purpose of this chapter is to review the c onsiderations for expander appl ication in the power-cooling cycle. An evaluation of the ammonia-wa ter working fluid properties is given and comparisons are made with other power cycle working fluids. These properties are linked to design considerations for various mach ine types and data from the literature is used to base estimations on the expected perf ormance of expanders for this application. Working Fluid Properties In this section the thermophysical proper ties of the ammonia water working fluid are considered as they relate to expander design. Ammonia is seen to behave more like steam rather than the organic fluids that are typically used in low temperature Rankine conversion systems. Expanders are treated in two groups, one being dynamic machines, those that convert the fluid s energy to velocity and crea te shaft power by a momentum transfer, and the other being displacement devices, where the working fluid is confined and allowed to expand against a moving boundary. Table 4-1 is a comparison of fluid propert ies for other power cycle working fluids as well as ammonia-water for a hypothetical, isentropic expansion. The most significant difference between ammonia-water and the ty pical ORC fluids is the large isentropic enthalpy drop of ammonia-water. This corres ponds to a higher ideal jet velocity which has an impact on dynamic turbine design. C onsidering steams characteristics, it and

PAGE 52

34 ammonia-water are similar with respect to enth alpy drop and jet velocit y. This is to be expected because of the close mo lecular weights of both fluids. Table 4-1. Fluid properties fo r a typical ammonia-water co ncentration a nd other power cycle fluids for isentropic expansion from saturated conditions at 100 C to condensation/absorption at 35 C. A ba sic solution concentration of 0.40 was assumed for the ammonia-water data. Fluid Characteristics AmmoniaWater, 0.99 Steam Isobutane [47] HCFC-123 [48] Molecular Weight 17 18 58 153 Isentropic Enthalpy Drop, kJ/kg 327.8 418.0 56.75 29.77 Ideal Jet Velocity, m/s 809.7 914.3 336.9 244.0 Volumetric Expansion Ratio 11.5 13.1 4.91 5.96 Exhaust Quality 0.837 0.873 1 1 Since work production with a turbine is a momentum tran sfer process, the relative velocities of the fluid stream and rotating blad es are critical design parameters and they can be used to characterize the operation of a turbine. The velocity ratio, which is the ratio of rotor tip tangential speed to working fluid ideal jet velocity has preferred values for differing flow arrangements, axial, radial, etc. The preferred value of this parameter will relate the design parameters of rotor diam eter, rotor inlet area, and rotational speed to the isentropic enthalpy drop across the turbine. For steam and ammonia, with their high ideal jet velocities, these requirements resu lt in a choice between extremely high rotor speeds with small diameters or impractically small inlet flow passages with enlarged rotor diameters. Designs for heavier organic fl uids result in efficient geometries at small sizes, that is reasonable shaft sp eeds and inlet flow areas even with small diameter rotors. In fact, the advantage of usi ng organic fluids to design efficient, small turbomachinery is a well-known feature of ORC engines [49]. Multiple stages or partial admission operation is the traditional solution for steam turbines; however, multiple stages add to the cost and partial admission operation places limits on maximum efficiency.

PAGE 53

35 As for positive displacement expanders, th e important parameter is the expansion ratio. For these machines, the higher enthal py drop of ammonia manifests itself, not as high velocities, but increased pressures. Refe rring back to Table 4-1 indicates that the ammonia-water fluid has a significantly higher expansion ratio than the presented organic fluids. Preliminary Machine Sizing Given the information from the previous section regarding the expansion properties of the ammonia-water working fluid, a similari ty analysis can be used to strengthen the generalizations mentioned in a more quantified manner. For this discussion it will be helpful to use the similarity parameters of specific speed, nspec, and specific diameter, dspec for a single stage unit. They are defined as: 3 4exit spec idealQ n h (4-1) and 1 4ideal spec exitDh d Q (4-2) where is the shaft speed in rad/s, hideal is the isentropic enth alpy change across the turbine(J/kg), and Qexit is the volume flow rate exiting the turbine (m3/s). The term D (m) represents diameter and is defined differen tly for different turbine types, but it is a characteristic dimension that indicates size of the unit. Two additional similarity parameters are needed to fully describe the performance of ge ometrically similar machines, typically machine Reynolds numbe r and Mach number. However, these additional parameters have only secondary effects and are typi cally neglected for similarity analysis [50].

PAGE 54

36 With some general operating conditions fr om the simulation, a range of typical specific speed values can be computed and us ed to identify possibl e expander designs. For example, referring to Figure 3-3 the p eak efficiency value occurs near a boiling pressure of 0.7 MPa, this operating condition was used to compute a range of specific speed values and is presented in Table 4-2. By definition, specific sp eed is influenced by the capacity of the device through the exit volume flow rate. This effect is accounted for by considering two nominal out put values in the table. Table 4-2. Single stage specific speed calcula tions versus nominal work output and shaft speed. Ideal Output Shaft Speed 5000 rpm 20,000 rpm 60,000 rpm 5 kW 0.0084 0.033 0.10 30 kW 0.021 0.082 0.25 To provide some indication of suitable expanders, specific speed and diameter ranges taken from suitable references [49, 50] are presented in Table 4-3. What can be gathered is that partial admission axial turbines could match well with cycle conditions over a wide range. Full admission devices, both axial and radial inflow, would be suitable for only very high speeds or large power outputs. However, while partial admission operation may be better than full admission devices, by reducing clearance and secondary flow losses, other loss mechanisms appear and the net re sult is that partial admission operation will always be less effi cient than optimum full admission devices [51]. For cases of low speeds and/or sma ll outputs a positive-disp lacement reciprocating device may give good performance. Note that Table 4-3 is not inclus ive of all potential candidates, other types, for example rotary vane and scroll expa nders, could also be considered as will be shown. Also multiple stages could be used to divide the fluids energy content among separate stages and thus change the resultant specific speed values.

PAGE 55

37 Table 4-3. Approximate specific speed and sp ecific diameter ranges for efficient (>60%) single stage expander types [49, 50]. Expander Type Approximate nspec range Approximate dspec range Partial admission axial 0.008-0.1 10-50 Full admission axial 0.1-8 1-15 Radial inflow 0.1-1 2-10 Reciprocating piston 0.00002-0.008 14-70 Rotary piston 0.015-0.4 1.5-10 Technology Review With some background into the expansion situation present in the power-cooling cycle, a review is now presented which hi ghlights solutions of similar applications, proposed designs, and expected performance values. Dynamic Machines As indicated in the previous section, singl e stage, full admission turbines are more appropriate for larger outputs, which is confirmed by examples in the literature. Direct examples of turbines operati ng with ammonia or ammoniawater are found primarily in two research areas: Kalina cycle research and its derivatives and closed loop systems for ocean thermal energy conversion (OTEC). The Kalina-based research is more relevant and is discussed here. Closed loop OTEC systems typically employ ammonia as the working fluid, however, the limited temperatur e drop being exploited results in turbine pressure ratios of 1.4-1.5, for example [52], wh ich is lower than th e preferred conditions for the power-cooling cycle. Additionally, some discussion of steam turbine implementation will also be relevant. One of the many advantages emphasized by Ka lina cycle supporters is the fact that no new turbomachinery needs to be develope d, rather, conventional steam equipment can be employed successfully due to the fluid dynamic similarities of ammonia and water [53]. Not surprisingly most examples availa ble in the literature have employed steam

PAGE 56

38 equipment. Some information comes from th e few Kalina-based cycles that have been constructed. The first implem entation of this technology was a 3 MW heat recovery pilot plant at Canoga Park, CA [54]. A subsequent application was utiliz ation of a geothermal resource in Husavik, Iceland [55], where th e nominal output was 2 MW. Finally, a team at Waseda University in Japan is experiment ally investigating a derivative of the Kalina cycle technology which they term the Wase da Ammonia-Water Mixture Turbine System (W-MTS) [56]. Table 4-4 is a summary of the pertinent operating features for these turbines. Other installations have been reported, however, operational data is limited. Table 4-4. Reported turbine operating parameters and efficiencies for three systems using an ammonia-water working fluid [54-56]. Description x [kg/kg] Tinlet [C] Pinlet [MPa] Pexit [MPa] Exit Quality Size [%] Canoga Park 0.70 514 11.03 0.192 1 3.7 MW 90.1 Husavik, Iceland 0.95 121 2.72 0.534 0.946 2 MW 60 Waseda Univ. 0.62 162 1.5 0.385 0.990 60 kW 40 Steam Turbines It has been previously mentioned that conventional steam equipment has been proposed and used for operation with ammoniawater mixtures. This facilitates turbine design for large systems [53], and lowers the investment cost for smaller installations-where an off-the-shelf device would be used [57]. This section presents typical efficiencies for available steam turbines us ing the assumption that they will compare well to performance with ammonia-water. Over the range of 50 kW to approximately 10 MW the appropriate turbine choi ces range from small single stage machines to larger multistage units with a corresponding efficiency range of 50-80% [58, 59]. Larger output machines, 20-100 MW, report slightly better effi ciencies, to +80% [60]. However, with tailored designs, better efficiencies are expected, as evidenced by the designs for a

PAGE 57

39 Kalina-based, gas turbine bottomi ng cycle [53]. High efficiencies, +90%, were estimated but no verification is available. The comput ed operating parameters are presented in Table 4-5. Note that superheating is used to avoid condensed flow at the exhaust. Table 4-5. Estimated operating data for th e three turbine stages of a Kalina-based bottoming cycle [53]. Description x [kg/kg] Tinlet [C] Pinlet [MPa] Pexit [MPa] Exit Quality Size [%] High P 0.817 565.6 19.4 3.45 1 105 MW 94 Intermediate P 0.817 511.8 3.11 0.606 1 96 MW 96.5 Low P 0.817 141.8 0.572 0.281 0.998 23 MW 89.5 Displacement Machines In displacement machines the high speci fic energy of the nearly pure ammonia working fluid is not necessarily a disadvantag e for machine design since it is manifested as higher operating pressures. Within this class are two basic dis tinctions: reciprocating and rotary. Reciproca ting machines produce linear motion with parts such as pistons or diaphragms and require valves to alternat ely open and close for operation. Rotary devices create expanding chambers through the geometry of one or more rotating members. Valve operation with these device s is much simpler, usually the rotating components are used to cover and uncover th e inlet and exhaust ports. Reciprocating machines can be constructed to minimize leakage in the expansion chamber; however, the need for valve actuation hinders effi ciency. Because of their simpler valve construction, rotary machines are genera lly preferred even though leakage around the rotating members that comprise the expans ion chambers is a significant limit to efficiency. Displacement machines have generally b een relegated to small power outputs, being replaced in the larger sizes by ga s and steam turbines for power production

PAGE 58

40 applications. Badr et al. [61] evaluated avai lable expanders on the ba sis of efficiency and operation for small scale steam expansion. R eciprocating devices were judged to be less reliable and efficient than rotary devices for reasons similar to those already mentioned. The rotary machines evaluated were Wankel, screw, and vane expanders, which have published isentropic efficiencies (with st eam operation) of 13-25%, 25-40%, and 65-80% respectively [61]. However, a machine t ype not in common use during the survey by Badr et al. was the scroll expander. Since the scrolls intr oduction, primarily as compressors in air conditioning and refrigera tion systems, they have been proposed for use as expanders in small-scale Rankine sy stems [18, 20, 21]. There are some examples of their use in the literatu re; however, none appear to mention ammonia and very few with steam. Reported maximum efficienci es for R123 and R134a working fluids are approximately 67% [18], with compressed air values to 73% [20] have been seen, and 34% is a reported efficiency for non-lubr icated steam operation [62]. Table 4-6 summarizes the operational parameters fo r examples found in the literature. Table 4-6. Reported efficiencies of scroll expanders [18, 20, 62, 63]. Description Fluid Vol. Ratio Output Kane Topping HCFC 123 2.3 5 kW 67 % Kane Bottom HFC 134a 2.3 8 kW 67 % Oomori HCFC 123 2 400 W ~ 50 % Smith A/C Air ~ 2 120 W 74 % Smith Ref. Air ~ 4 < 500 W 72 % Kim Steam-no oil 4.6 15 kW 34 % On the other hand, displacement devices ha ve an added complication, lubrication. Lubrication, usually in the form of oil mixe d with the working fluid, serves two purposes: it reduces the friction between sliding surfaces within the device, and more importantly it enhances the sealing action of the expansion chambers [64] The high viscosity of oil compared to vapor working fluid prevents it from being quickly driven through small

PAGE 59

41 gaps created by close running parts. This e ssentially traps the working fluid for a short time creating a well-fitted seal. Therefore, to be effective in sealing, the oil must be applied from the high pressure side of the clea rance. In compressors this is not difficult because the oil can be throttled from the high pressure side to the compressor inlet. For expander duty, some work must be expended to inject the oil into the high pressure inlet side [20]. A second problem comes at the ex pander exhaust when the oil needs to be separated from the working flui d. It has been reported that oil lubricated steam expanders form an emulsion in the exhaust st ream that was difficult to separate [61]. Also, the condition of the oil itself, quantity and temperature, have their own impacts on expander performance. Temperature affects the oil viscosity and it must be balanced between the extremes of being too thin, and not able to seal the working fluid, versus being too thick, where viscous friction losses are excessive. Badr et al. [64] found that oil temperature alone affected rotary va ne expander isentropic efficiency by approximately 4%. Additional Considerations The previous sections discussed the relati ve merits and considerations for differing expander types, this section c onsiders some of the more pract ical design issues involved with expander implementation. When specifying or modifying an expander the corrosive nature of ammonia will need to be considered. The most critical issue is that copper and copper alloys are severely attacked by ammonia and will not su rvive service in an environment directly contacting ammonia. Another i ssue is that ammonia mildly attacks certain seal materials and even some lubricants. Most not ably these are Viton and silicone.

PAGE 60

42 Depending on the design, the expander sh aft may protrude from the housing and a seal will be needed. This is a concern fo r any closed thermodynamic system; however, it is particularly acute here due to the objecti onable properties of ammonia-water. In large installations, for example the Husavik Kalina cycle, conventional seals pressurized with nitrogen are used to prevent leakage [57]. Al so, the use of labyrint h seals with packings kept under slight vacuum have been proposed for large turbines [53]. Any ammonia leakage is diluted in water and is either recycled or used as fertilizer [53, 55]. In smaller systems these complicated sealing systems are not feasible and simple contacting shaft seals will likely be employed. For the experimental part of this work graphite-impregnated PTFE shaft seals have been used with some success. Given a proper design it should be possible to specify a seal with an acceptable lifetime. However, wear is inevitable so the best solu tion would be a hermetically sealed device, very similar in construction to the compresso r units of household refrigerators and small air conditioning units. One such proposal is the use of a high-speed (30-150 krpm) turbomachine with an integrally mounted gene rator and feed pump [65]. Similar ideas are proposed for small (10-200 kW) combus tion turbines to also reduce costs by eliminating the need for a reduction gearbox [66] The problem is of course corrosion of the copper windings that will be in an encl osed generator. Two solutions have been encountered; one was employed in a Gene ral Electric-designed ORC which was a magnetically coupled shaft coupling which transmitted torque through a housing wall [12]. Another, used in microturbine desi gns, is a magnetically coupled generator with shaft-mounted, rotating magnets inside the turbine housing an d a surrounding, but external, set of fi eld windings [66].

PAGE 61

43 A potential concern is the increase in expander exhaust pressure due to the implementation of the cooling heat excha nger. It has been shown that exhaust temperatures are sensitive to exhaust pressure with lower pressures preferred for cooling production. Off-the-shelf heat exchangers may introduce a more severe pressure drop since their economical designs typically trad e performance for compactness (cost). The same situation has been faced by ORC designe rs in the past, where regenerators were used to preheat the liquid boiler feed with the turbine exhaust. Low pressure drop designs were produced, for example the custom unit described by Batton and Barber [67], and similar devices would be needed for this application. Expansion process Compared to steam, ammonia wets more r eadily upon expansion. In fact, the work by Kremmer and Okurounmu [68], which is a study of the condensation process during rapid expansion, used ammonia because of its relatively high rate of nucleation and its rapid approach to saturation conditions, as compared to other candidate fluids. This implies that super-saturation of ammonia will occur to a lesser extent than with steam. The net effect will be more condensation with ammonia than that compared to equivalent steam exhaust qualities. Working fluid condensation could hinde r dynamic machine operation because of the momentum lost to the condensed droplets. These droplets are also responsible for erosion damage and can limit the useful lif e of the equipment. For example, the ammonia-water turbine used at the Husavi k geothermal station in Iceland required rebuilding due to condensate flow from an inadequate separator design [57]. Conditions for the expectation of maxi mum condensation are described in Chapter 7. As for contending with the issue, similar conditions are encountered in the

PAGE 62

44 utilization of geothermal steam that is pr oduced by flash boiling. The solution used for those conditions is to constr uct blades with erosion and co rrosion resistant materials and to incorporate condensate-draining channels into the blades [69] Condensation is typically not a design issue for displacement ma chines and may actually be beneficial for sealing purposes. Conclusion Based on the information in this chapter, some general divisions can be drawn. One is on the effect of scale and machine type on isentropic efficiency. For small outputs, positive displacement devices are pref erred, and with some development into the promising machine types efficiencies of +70% appear possible. The chief design parameter for these devices is the isentropic volume ratio of the working fluid. At the other end of the spectrum, large output devices, turbines are dominant and efficiencies to 95% have been demonstrated at the high end of this range an d 90% estimated at the lower end. The mid-range is a different matter, turbines are genera lly more economical; however, based on comparable steam performanc e, efficiencies range from 50 to 80%. With sufficient flow rates, successful turbin e designs can be formulated for nearly any pressure ratio large to small. The other separation comes from the thermo -physical properties of the nearly pure ammonia working fluid. For the purposes of cooling product ion saturated inlet conditions are preferred, which leads to conde nsation forming in the exhaust. Ammonia wets more readily than steam upon expansion, so the exit quality must be monitored to avoid damage to the turbine. At least for turbines, this requirement could essentially limit the minimum exhaust temperature to the corr esponding dew point temperature. To avoid condensed exhaust, similar measures as thos e used for steam equipment may be needed,

PAGE 63

45 for example superheating the inlet stream. Th e disadvantage is that this would move the inlet vapor conditions further from the preferre d ones outlined in Chapter 3. This would undoubtedly degrade any cooling production.

PAGE 64

46 CHAPTER 5 EXPERIMENTAL APPROACH An experimental study of system operati on was carried out to demonstrate subambient exhaust temperatures, confirm expected trends in exhaust temperature and vapor production, and identify inadequacies in the th eoretical modeling. Th is chapter contains a description of the experimental setup, genera l procedures used for experiments, and an outline of the experiments performed. Setup Description The experimental setup used for this work is detailed in this section. The system is based on the setup used by Tamm [70]. For this work Tamms boiling-absorption loop was modified and expanded. Primary additi ons to the experimental setup include a rectifier to condition the vapor and a turbine to extract work from the fluid. In Tamms experiments the high pressure and temperatur e vapor was throttled to the low-pressure side without performing work [70] because a turbine was not used. Instead, a heat exchanger was used to remove heat from the vapor, which was intended to thermodynamically simulate the effect of a turb ine [70]. In addition to adding a turbine, modifications to Tamms circulation loop we re also made to improve performance. The basic schematic of the current experi mental system is shown in Figure 5-1, which is intentionally similar to the mode led system of Figure 3-1. Key differences between the two are that ther e is currently no experimental cooling heat exchanger and the rectifier condensate drains directly to the absorber without mixing with the boiler weak solution. As shown in the next ch apter, experimental cooling production was

PAGE 65

47 minimal so a heat exchanger following the e xpander was not judged to be necessary for these tests. The change in plumbing for the rectifier condensate is a result of a repair modification to the experimental setup. Figure 5-1 also shows the location of instrumentation in the experimental syst em. Figure 5-2 is a photograph of the experimental system. Thermocouple Flow Measurement Fluid Sampling Port Pressure TransducerP SKeyT F Throttle Recovery Heat Exchanger TBoiler S F T Heat Source F T P T T Separator F Cooling Water S T T Rectifier T T S F T Absorber T P TP P T Cooling Water T ExpanderP T T F P Condensate Superheater Solution Pump Figure 5-1. Schematic of experimental setup. Heated water is used as the heat source for the experiment. It is heated in an electric water heater which uses phase change material for thermal storage. This water heater is controlled by an adjustable thermo stat. A second, conventional water heater is used solely as a hot water storage tank. A centrifugal pump is used to circulate water

PAGE 66

48 between the two tanks as well as to the boiling heat exchanger of the power-cooling system. This arrangement is used for c onvenience and control of the heat source. Figure 5-2. Photograph of experimental setup. The hot water pump circulates the heat s ource through one side of the boiling heat exchanger. This heat exchange r is composed of two individual flat plate heat exchangers. This construction of heat exchanger is also us ed for the recovery heat exchanger in which the cool strong solution recove rs heat from the weak liqui d exiting the separator. A rotary vane pump, driven by an electric motor, serves as the solution pump and is used to draw strong solution from the absorber and pump it through these heat exchangers into the separator, refer to Figure 5-1. Once the two-phase mixture enters the sepa rator it is separated by gravity into its liquid and vapor components. The separator it self is simply an empty tank, it contains no

PAGE 67

49 baffles or special equipment. The weak li quid drains from the bottom of the separator and is pressure-driven through the recovery he at exchanger through a throttling valve into the absorber. The vapor rise s out of the top of the sepa rator and depending on whether rectification is performed, either enters the bot tom of the rectifier or is directed through a bypass line to the top of the rec tifier. The rectifier is the combination of another flatplate heat exchanger and a packed-bed entrainm ent separator. Cooling for the rectifier is provided by the circulation of chilled water a nd ethylene-glycol mixture, the same fluid used to cool the absorber. Any condensate draining from the rect ifier passes through a sight glass and is throttled th rough a valve back to the absorb er. The sight glass is used to ensure that only liquid is be ing throttled back to prevent short-circuiting of the vapor. The vapor that leaves the top of the re ctifier is routed through a superheating section and on to the turbine. Superheati ng is achieved by a va riabletemperature heating tape that has been wrapped around a portion of the va por plumbing. There is also a throttling valve parallel to th e turbine so the vapor could be bypassed. From the turbine exhaust the vapor is routed to the absorb er where it is bubbled into a pool of basic solution liquid through a tube perforated with small holes. Absorber heat exchanger arrangement is presented in a subsequent section. As for the absorber cooling source, a va por-compression water chiller is used to cool a 50/50 mixture of ethyl ene-glycol and water. A centrifugal pump circulates fluid between a storage tank (approximately 110 gall on) and the chiller. The chiller has an internal, adjustable thermostat that mainta ins storage tank temperature. A separate centrifugal pump is used to circulate flui d from the storage tank through the heat exchangers in the absorber (yet another pump is used to circulate fluid to the rectifier).

PAGE 68

50 As with the heat source, a chiller is us ed to provide a range of convenient and controllable heat re jection conditions. Construction of the system itself is prim arily with stainless steel tubing connected with high quality compression fittings. In so me areas, mainly where larger diameters are needed, the construction is with threaded black iron pipe and fittings. Specific changes to the original system [70] are described in the following sections and complete experimental details are provided in Appendix B. Expander The expander for the experimental system has to conform to the parameters set out in Chapters 3 and 4, but it has the additional co nstraint of operating with low flow rates produced in the experimental system. Initia l testing was performed with a rotary vane compressor, which was modified to operate as an expander. Performance with this particular device was poor, apparently due to internal running friction since no oil lubrication was used. Testing of this device was abandoned in favor of the dynamic machine described next. From the specific speed considerations mentioned in Chapter 4, partial admission turbines can be well suited to the conditions generated by the power-cooling cycle. An off-the-shelf turbine that could be adapte d for partial admission operation was sought. The device that was found is a single-stage, radi al inflow turbine originally for use in an air-cycle cooling system. It was suitable for this application because of its small size and because it could be configured to operate in a partial admissi on configuration. The stator is made up of a ring of individual nozzles, a nd these could be indi vidually blocked off, with an epoxy in this case, so as to oper ate with a low flow rate in partial admission configuration. The nozzles have a convergent section only. Furthermore, all components

PAGE 69

51 were aluminum or steel so it would be comp atible with the ammonia working fluid. The original housing on the other hand, could not be used since it was vented to the atmosphere and the protruding sh afts were sealed only with a labyrinth seal. A new leakproof rear housing was construc ted to mount and enclose the turbine spindle, Figure 5-3. Also the bearings were replaced since the orig inal ones used a bronze ball retainer that was quickly consumed by the ammonia working fluid. Figure 5-3. Modified turbine us ed for experimental testing. Loading of the turbine is done in a simp le, non-contact manner because of the high shaft speeds encountered (20,000 60,000 rpm). The turbine drives an aluminum disc that is mounted inside the housing on the tu rbine shaft. A strong permanent magnet is used to induce a counter-current in the rotating disc. This creates a reaction torque that the turbine must overcome and is proportional to shaft speed. The magnitude of this counter-torque is controlled by changing the distance between the magnet and the disk. Of course this device produces no mechanical work, all of the energy is turned to heat

PAGE 70

52 and dissipated (< 300 W). While this met hod has the disadvantages of not producing mechanical work and potentially adding heat to the spindle of the turbine, it eliminates the need to incorporate a shaft seal since th e field lines of the magnet can penetrate the walls of the aluminum turbine housing. This makes the experimental apparatus easier to construct and maintain. Bearing life may be shortened, however, that is not a concern for this work. On the other hand, h eat transfer from the hot disk to the (ideally cold) turbine exhaust may prove to negatively affect result s. Discussion relati ng to this point is provided in the subsequent chapters. According to Figure 5.1, the inlet temperatures and pre ssures of the turbine are measured with appropriate transducers. The turbine was also outfitted for shaft speed measurements. The sensor is based around an infrared emitter-receiver pair, that registers a pulse when a hole in the turbine shaft ali gns between them. Two pulses occur for each revolution of the shaft, therefore, the freque ncy of this signal is proportional to the shaft speed. Rectifier The significant impact of rectification on cooling production was identified in the theoretical modeling of Chapter 3. This prompted the implemen tation of a rectifier in the experimental setup. Construction is very si mple, cooling and condensation take place in a flat-plate brazed heat exchanger and condens ate separation occurs in a vertical section of 2-inch pipe filled with inch Berl saddle packing. The packing captures any entrained liquid droplets. The column was originally intended to function as a direct contact heat exchanger; however, performa nce in that mode of operation was not satisfactory so the extern al condenser was added.

PAGE 71

53 Absorber Modifications to the absorber were directed at correcting two issues: increasing the heat rejection capacity and di rectly cooling the exiting st rong solution. The initial configuration of the absorber is shown in Fi gure 5-4. As can be seen, heat rejection was taking place only as the weak solution was dr ipping over the heat exchanger stack. No cooling was taking place in the liquid pool wh ere a significant amount of heat addition was occurring due to the bubbling in of the va por. This resulted in performance problems such as pressure and temperature build-up in the absorber and served to hinder pump performance. To alleviate these issues a heat exchange coil was designed and placed below the existing heat exchanger stack in the liquid pool. A new bubbling tube was fabricated to accommodate the heat exchanger coil, as shown in Figure 5-4. Original absorber configuration Weak In Vapor In Strong Out Liquid Level Shelf Bubbling Tube Heat Exchanger Stack Strong Out Modified absorber configuration Vapor In Weak In Bubbling Tube Coil Assembly Figure 5-4. Original and modi fied absorber configurations.

PAGE 72

54 The coils are made of 9.5 mm OD alumin um tubing, and the entire assembly is made of 3 separate coils connected in para llel. Each of the 3 coils is made of approximately 3.4 m of tubing, so the entire coil assembly contains approximately 10.2 m of tubing. Furthermore, this coil assembly is in series with the upper heat exchanger stack, with the coolant passing through the coil assembly before the stack. Pump In previous work, cavitation problems cau sed by pumping strong solution very near saturation, were a hindrance to system operati on [70]. Several measures were taken to alleviate those problems for this work. Fo r instance the piping connecting the outlet of the absorber to the inlet of the pump was simplified, which was comprised of reduced length and number of bends, and it was increased in diameter to match the exit fitting of the absorber. In addition, the diaphragm pump was replaced. From an applications standpoint, a diaphragm pump should have b een an appropriate choice for the powercooling cycle. However, the old pump, which was removed from other equipment, appeared to have a minimum inlet pressure re quirement and did not pump well, or at all, at low absorber pressures. This was partic ularly troubling at star tup when the absorber was cool. A rotary vane pump was chosen as a replacement partially because it could handle a liquid-vapor mix but also because it wa s compatible with a drive already in the lab. Service was adequate, but far from ideal. Perhaps this is not surprising since Barber and Prigmore advised of poor pump performa nce when giving design guidelines for ORC engines [49]. For estimation purposes Barber and Prigmore used a pump efficiency of 40%.

PAGE 73

55 Data Collection This section describes the general equipment used fo r data collection, further details can be found in Appendix B. Most of the collected data is recorded with a computer-interfaced collection system. All temperatures are measured with T-type thermocouples and pressures are detected with pressure transducers. Measurements are saved to a PC through the appropriate interface cards and data acquisition software. Two instrument types are used for measuring flow rates: the vapor flow into the turbine is measured with a turbine-type flow meter th at provides a signal whose frequency is proportional to flow rate, and the strong, wea k, heat source fluid, and coolant flows are measured with float-type rotame ters and are recorded manually. Provisions are in place to sample the working fl uid at points of interest in the cycle. Syringe sampling ports were placed on the st rong, weak, rectifier vapor inlet, and the rectifier vapor outlet lines, see Figure 5-1. Th ese correspond to four of the five different concentrations that are present during steady state operation of the cy cle, Figure 3-1. The procedure for liquid samples is to sample the working fluid with a syringe and then determine concentration with a gas chromatogr aph (GC) analysis. This procedure works well with liquid samples, but not as well with the saturated vapor samples. The vapor condenses easily in the syringe thus causing errors in the GC analysis. Therefore, vapor concentrations are determin ed from property relations using the locally measured temperature and pressure and using the assump tion of saturated vapor at the separator and rectifier exits. Experimental Method For each set of conditions to be tested, a st andard test routine was established and is described here. The first step in establishi ng stable system operat ion was to arrange the

PAGE 74

56 parameters that cannot be adjusted duri ng operation, for example the number of open turbine nozzles. Next, the heat source and heat rejection subsystems were started and allowed to circulate until temperatures stabil ized; then, circulation of the basic solution was started. The heat source flow rate wa s controlled to maintain the desired boiling temperature of the two phase mixture leavi ng the boiler. With basic solution flow established the weak solution flow from the se parator to the absorber was controlled so as to maintain the desired level of solution in the absorber. Vapor flow was regulated only by the nozzle restriction of the turbine. As the solution in the absorber was heated by the returning liquid and vapor streams, the absorb er coolant flow was adjusted to maintain the desired absorber pool temperature. When the rectifier was active, vapor flow was diverted through the vapor heat exchanger a nd the packed bed entr ainment separator. Vapor coolant flow was adjusted to maintain the desired rectifier exit temperature. The feedback and adjustments mentioned were performed manually. Fortunately, during testing pseudo-steady-state operating conditi ons were encountered and the adjustments were quite manageable. With the system operating at a specified set of conditions data acquisition could begin. For each set of cond itions several individual meas urements were made, usually eight. The only parameter that was changed during these measurements was the loading of the turbine. This was done to find the sh aft speed where optimum efficiency occurred. A period of five minutes was judged to be ad equate between adjustments to the turbine loading. Therefore, for each set of operati ng conditions, testing lasted approximately 4045 minutes.

PAGE 75

57 Experiments Performed Experiments were designed to demo nstrate sub-ambient expander exhaust temperatures and to isolate the important trends identified during the theoretical analysis. The effective techniques that were available and the variations performed with them are discussed in this section. Boiler Exit Temperature As mentioned in the previ ous section, the mixture te mperature exiting the boiling heat exchanger was controlled by the flow and temperature of the heat source fluid. Nominal exit temperatures of 60, 80, and 95 C were considered. Basic Solution Concentration The basic solution concentration was va ried over a fairly narrow range by monitoring the solution level in the absorber. General effects were to increase vapor production and concentration, however, at the expense of absorption pressure. Three absorber levels were tested which resu lted in basic solution concentrations of approximately 0.381, 0.396, and 0.414. Superheating Heating of the vapor before it entered the turbine was performed with an electrical resistance heating tape wh ich was wrapped around a section of the vapor piping. The electrical tape heating was cont rolled with a variable transf ormer. Given the arrangement only a small amount of superheating was possible, approximately 5 10 C. Absorption Temperature Similar to the boiler exit temperature, the absorber liquid pool temperature was controlled by varying the ch illed fluid temperature and flow rate. Considered temperatures were nominally 25 and 35 C.

PAGE 76

58 Nozzle Flow Area The effect of nozzle flow area was to change the flow restriction imposed by the turbine. This directly cont rolled the coupled effects of boiling pressure and vapor concentration. For a discernable change the number of open nozzles tested were one and four. Rectification As with the theoretical m odeling, the amount of rectifi cation in the experimental setup was determined by the rectifier exit temper ature. This in turn was controlled by the coolant flow rate through the condensing h eat exchanger. Only two conditions were tested, either no rectification was performed or the exit temperature was nominally set for 35 C. Conclusion This chapter has presented an overview of the experimental system that was used for this work. As shown in the concluding chapters, the data from these experimental tests can be used to provide ge neral verification of the trends identified in Chapter 3 and a moderate demonstration of the power-cooling cycles key concept. Also, unexpected behavior of this real system adds to the collective design experience.

PAGE 77

59 CHAPTER 6 EXPERIMENTAL RESULTS In the previous chapter the techniques a nd types of experiments were described. This chapter presents the results of those te sts as they relate to the operating trends identified in the theoretical analysis. Overa ll, the anticipated trends have been confirmed by these results. However, there were unexpected deviations, and these included problematic turbine performance and other equipment limitations that obscured some effects. Confirmation of Trends The results based on system modeling ha ve been informative regarding the characteristics and potential of the power-cooling cycle. C onfirmation of this expected performance is needed in order to place c onfidence in the derived conclusions. This section describes the experimental results as they relate to the operating mechanisms under discussion. Pressure Variation The boiling pressure is the first parameter to be considered. Its effect is to control the vaporization of the basic solu tion in the boiler. As pre ssure decreases more vapor is formed and its concentration drops--ultimately until the saturation pressure is reached and all of the basic solution has va porized. These effects have been isolated experimentally and are presented in Figure 6-1.

PAGE 78

60 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 0.20.30.40.50.60.70.8Boiling Pressure [MPa]Vapor Ammonia Concentration0 0.05 0.1 0.15 0.2 0.25 0.3 0.35Vapor Mass Flow Fraction 4 nozzle xv 1 nozzle xv 4 nozzle mass flow 1 nozzle mass flow Measured Data Simulated Vapor Mass Flow Simulated Vapor Concentration (+/-0.00067 kg/kg) Figure 6-1. Measured effect of pressure va riation on vapor quantity and concentration. Simulated data superimposed to extend observed trend. To obtain Figure 6-1 the number of open no zzles in the turbine was changed from 1 to 4. The effect of increasing the vapor mass flow rate and decreasing the vapor concentration is clearly show n. Simulated results are supe rimposed on the figure and are used to extend the trend to th e limits of par tial vaporization. Concentration Variation Basic solution concentration variations have similar ultimate effects as those of the boiling pressure. That is the vapor mass flow fraction will increase with the basic solution concentration. However, with fixe d boiling temperature and pressure, the vapor concentration remains constant. These trends are seen in Figure 6-2, which also has simulated results superimposed. Since the liqui d level in the absorber was used to vary the basic solution concentration, the span of concentrations was limited to a somewhat

PAGE 79

61 narrow range. As seen in Figure 6-2 the meas ured vapor flow rate is consistently lower than that expected from the equilibrium mode l indicating some inefficiencies in the vapor production process. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.20.250.30.350.40.450.5Basic Solution ConcentrationVapor Ammonia Concentration0 0.05 0.1 0.15 0.2 0.25 0.3 0.35Vapor Mass Flow Fraction Measured Concentration, +/-0.00067 kg/kg Measured Mass Flow Simulated Vapor Mass Flow Simulated Vapor Concentration Figure 6-2. Measured effect of basic solution concentration on vapor production. The other effect of basic solution concentr ation is to reduce the expander pressure ratio by the increase of absorption pressure wi th increasing concentration. The relevant data are presented in Table 6-1. Table 6-1. Measured decrease of absorption pressure with basic solution concentration. Basic Concentration Ave. Absorption Temp. Ave. Absorption Press 0.414 34.8 C 0.250 MPa 0.396 34.7 C 0.237 MPa 0.381 34.4 C 0.225 MPa Temperature Variation Temperature is the last boiling condition to be considered. It also changes the degree of vaporization and thus the amount and concentration of vapor that is formed.

PAGE 80

62 The simulated effect of temperature change ha s been shown in Chapter 3, however, in the experimental system there is some coupling between temperature and pressure because of the fixed turbine nozzle restricti on, that is the pressure is not held constant. Figure 6-3 is a plot of the change to vapor mass flow fraction (mass flow vapor/mass flow basic solution) with boiling temperature. The expected trend of an increase in vapor flow with temperature is shown, but it is somewhat curt ailed due to the associated rise in boiling pressure, also shown. Agreement with equi librium modeling is reasonable except for the highest temperature case where an unusually high concentration reading is the suspect. 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 50556065707580859095100Boiling Temperature [C]Boiling Pressure [MPa]0 0.05 0.1 0.15 0.2 0.25 0.3Vapor Mass Flow Fraction Pressure (+/-2.24 kPa) Vapor Flow Fraction A verage pressure values used in simulation Simulated vapor mass flow fractions Figure 6-3. Measured change in vapor flow rate (relative to basic solution flow) due primarily to changes in boiling temperature. The effect is counteracted by the indicated rise in boiling pressure. Absorption Pressure The effect of absorption pressure on c ooling production was shown in Chapter 2 with the aid of an ammonia-water binary phase diagram. As it pertains to that diagram,

PAGE 81

63 increasing or decreasing absorption pressures shift the two-phase envelope either up or down respectively. Increases to pressure woul d then increase the dew point of the turbine exhaust, which is an initial indicator of the minimum temperature possible. Experimentally measured turbine inlet a nd outlet pressures and temperatures are presented in Table 6-2. As shown with th e sample data, the lower exhaust pressure corresponding to the lower abso rption temperature allows the vapor to be expanded to lower temperatures. Table 6-2. Measured data indicating effects of absorption temperature. Turbine Parameter 25 C Absorption T 35 C Absorption T Inlet Pressure [MPa] 0.329 0.329 Exhaust Pressure [MPa] 0.157 0.217 Inlet Temperature [C] 56.0 55.5 Exhaust Temperature [C] 42.0 46.1 Rectifier Penalty Rectifying the vapor before it enters the expander was shown to have a positive effect on exhaust temperatures and cooling prod uction to a certain extent. The benefits to vapor temperature and concentration from rec tifier operation is presented in Table 6-3. Also indicated in the table is the concom itant loss of vapor due to condensation. Naturally, the cases with the lowest inlet con centration have more of the vapor condensed to reach essentially the same ending concentration. Table 6-3. Averaged valu es for rectifier operation. Rectifier Parameter Nominal Boiling Temp. = 60 C 80 C 95 C Inlet Temp. [C] 58.7 78.1 94.4 Exit Temp. [C] 34.0 36.2 33.9 Inlet Concentration 0.968 0.927 0.854 Exit Concentration 0.993 0.994 0.996 Normalized Inlet Mass Flow 1 1 1 Normalized Exit Mass Flow 0.953 0.866 0.695

PAGE 82

64 Another way to view the losses associated with rectificati on is to consider the work that was sacrificed for lower exhaust temperat ures. Table 6-4 presents such information for a couple of the cases in Table 6-3. Th e measured work out is the work output, per kilogram of basic solution flow, based on e xperimentally measured conditions at the point of maximum turbine efficiency. The computed work out is the estimated work that could have been produced with the vapor not going through the rectifier but straight to the turbine. For this calculat ion, the turbines efficiency was assumed to be the same as the maximum measured value. The data in Ta ble 6-4 clearly show the detrimental effect of excessive rectification on work production. Table 6-4. Values for rectifier operati on highlighting penalty to work production. Parameter Nominal Boiling T = 60 C 95 C Measured Exhaust Temp. 25.9 C 22.3 C Max Measured Work Out 563 J/kg 1660 J/kg Computed Work w/no Rect. 612 J/kg 2600 J/kg Computed Drop in Work 7.9 % 36 % Concept Demonstration The realities of the experimental setup re quired a compromise in the testing plan. Based on experimental measurements, the turbin e, even with a single nozzle, is slightly oversized for the experimental setup. The re sult is that the boiling pressure falls to a value that is below the optimum identified in the theoretical analysis The consequences come in the form of reduced pressure rati os and vapor concentrations, both effects degrade cooling production. To compensate, boiling temperatures were increased to compromise between sufficient vapor flow rate for turbine operation and sufficient pressure to allow for a high degree of r ectification without sub-ambient condensing temperatures. As can be concluded from the previous theoretical analysis, this incurred

PAGE 83

65 much loss due to the rectifier, but the resul ting vapor mass flow and turbine inlet pressure was higher than could be achieved at lower temperatures. Efficiency was essentially traded for more suitable vapor conditions. Figure 6-4 shows the measured temperatures across the turbine in relation to the measured absorption temperature and Table 6-5 presents the parameters for this testing. Successive stages of rectifying and superh eating enabled the production of vapor with 0.993 concentration and temperatures ranging from the vapor saturation temperature, approximately 39 C according to Figure 6-4, up to the useful limit of the superheater. 15 18 21 24 27 30 33 35384144475053Expander Inlet Temp. [C]Temperature [C] Expander Exhaust Temp. Absorption Temp. 20 % 25 expander = Figure 6-4. Experimental measurement of the expansion of vapor to temperatures below those at which absorption-condensation is taking place. Obviously the minimum exhaust temperatur es of Figure 6-4 are not suitable for a cooling load, however, it is a clear measurem ent of the power-cooling cycle concept as it was explained in Chapter 2. In Chapter 2 the power-cooling cycle was contrasted with

PAGE 84

66 pure working fluid Rankine cycle operation, wher e it is impossible to expand the vapor to a temperature below that at which condensati on is taking place. While not a dramatic demonstration, Figure 6-4 clearl y shows expansion of the va por to temperatures well below the absorption-condensation temperature. Table 6-5. Averaged conditions for the testing of Figure 6-4. Expander Inlet Pressure: 0.516 MPa Expander Exit Pressure: 0.208 MPa Vapor Flow Rate: 0.00299 kg/s Vapor Concentration: 0.993 kg/kg Rectifier Inlet Temp.: 83.4 C Absorber Temp.: 31.4 C Superimposed with the data points of Fi gure 6-4 are lines of simulated performance for several isentropic efficiencies. At the lo wer inlet temperatures the experimental data is approximated by an expansion process with 20% isentropic efficien cy. At higher inlet temperatures the experimental data drifts aw ay from the 20% line and appears to improve in efficiency. This is an unexpected devi ation in the measurements and its possible source is discussed furthe r in the next section. Expander Performance Some difficulties with the thermodynami c performance of the turbine were encountered. First, the efficiency with th e ammonia-water working fluid was lower than the anticipated efficiency obt ained from air testing the turb ine. Second, some of the results based on thermodynamic measurements seem to indicate that the turbine efficiency is sensitive to inle t conditions. This section is a summary of the analysis into these phenomena and the conclusions regard ing the turbine performance measurements. Initial testing with the turbine was perf ormed with compressed air as the working fluid since it was simple to control and leaks we re not a problem. Deta ils of these tests as they relate to this work are provided in Appendix D. Based on this air testing the

PAGE 85

67 expected turbine efficiencies were in the ra nge 20-30%. The optimum ratio of ideal jet velocity to rotor tip speed was approximatel y 0.3. In order to maintain a similar ratio when testing with the ammonia-water mixture it would be necessary for the rotor speed to increase because of the higher ideal jet veloc ity for ammonia-water. However, this was not observed, possibly due to the reduced mass flow rate of ammonia-water as compared to the tests with air. This likely caused additional incide nce losses and resulted in the lower efficiency. Using thermodynamic measurements the turb ine efficiency appeared to vary and seemingly worsened as cooler exhaust condi tions were approached. A few thoughts on the measured performance are given here. In general, the observations from the experimental testing followed these trends: the expander exhaust c onsistently expanded to a point at or near the de w point for the measured exhaust pressure and estimated vapor concentration. This resulted in good indicated performance wh en the inlet temperatures were significantly higher than the exhaust dew point and poor indicated performance when the inlet temperatures were not significantly higher, for example the cases with rectification. A few possible ex planations are discussed below. The expander is a partial admission dyna mic turbine which was not designed to expand a two-phase working fluid. If enough fl ow were to condense it would alter the momentum transfer in the turb ine and there would be a corr esponding drop in efficiency. However, when the amount of condensation is examined for the experimental conditions, the concomitant effect on efficiency should be small. For example, Figure 6-5 presents the expected expander exit quality values for simulated conditions similar to those of the experimental testing. As can be seen, ev en for an isentropic expansion, the minimum

PAGE 86

68 expected exit quality does not go below 0.97 for these conditions. This amount of condensation would result in a minor expected penalty, 0.95 or higher [50]. Unless there were substantially more condensed flow th an expected, it does not appear to explain expander performance. 0.97 0.975 0.98 0.985 0.99 0.995 1 3537394143454749515355Expander Inlet Temp. [C]Equilibrium Exhaust Quality 100 % 70 % 30 % Pinlet/Pexit = 0.516 MPa/0.208 MPa xvr= 0.993expander= Figure 6-5. Expected equilibrium exhaust qualities for conditions similar to those of the experimental study. The exit quality is e xpected to be above 97% even for an isentropic device. Another possible explanation could come from errors in the thermodynamic power measurements. For pure component fluids ne ar saturation conditions the sensitivity of temperature to enthalpy changes is poor due to their isothermal phase change. This could introduce significant error in temperaturebased enthalpy measurements. For the ammonia-water binary mixture and the range of condensation which is being considered, however, condensation does not take place is othermally, even for a very high ammonia concentration vapor. Figure 6-6 is a temp erature-enthalpy diagram for two fluids, pure

PAGE 87

69 ammonia and a high concentrati on ammonia-water mixture, at a pressure typical of the exhaust conditions for the expander. The cu rves are similar except near the saturated vapor-two phase mixture boundary, where conde nsation begins for the mixture in the equivalent sensible heating range for the pure fluid. Figure 6-6 indicat es that expansions resulting in qualities of 0.97 or greater fall within a range where the sensitivity of temperature to enthalpy changes is good. Furt hermore, as Figure 6-6 also shows, if the near-isothermal phase change region were being encountered, the temperature would be drastically lower than what has been measured. -25 -15 -5 5 15 25 35 -2000200400600800100012001400Enthalpy [kJ/kg]Temperature [C]Mixture Vapor Quality = 1 0.98 0.97 0.99 0.993 Ammonia Concentration Pure AmmoniaPressure = 0.208 MPa Figure 6-6. Temperature-enthalpy diagram covering the phase change of pure ammonia and a high concentration ammonia-wate r mixture. For mixture qualities above 97% the sensitivity of enth alpy to temperature appears good. Direct measurements of power output were attempted and are described in Appendix D. The mechanism worked while testing with air, however, it gave inconclusive results with amm onia-water testing. Due to th is, air testing results were

PAGE 88

70 used to estimate the no-load power needed to drive the expander. These values can then be compared to the no-load results from th e ammonia-water testing, which is shown in Figure 6-7. 0 20 40 60 80 100 120 2900031000330003500037000390004100043000Shaft Speed [rpm]Power [W] Ammonia-water no-load test points labeled with the nominal expander inlet temperature 95 C 60 C 35 C 35 C 35 C No-load power consumtion based on air testing 80 C Figure 6-7. Comparison between the measur ed no-load power consumption of operation with compressed air (solid line) a nd ammonia-water (individual points). As can be seen in Figure 6-7, the air testi ng results indicate an approximate no-load power consumption of 10 to 25 W over the sh aft speeds considered. On the other hand, the ammonia-water test points va ry substantially. For inst ances where the expander inlet temperatures are nominally 60 C or lower (the multiple 35 C readings are all cases where rectification was used) the no-load power consumption is in the range of 10 to 42 W, and at inlet temperatures of 80 C and a bove the consumption is greater than 110 W. While these results are not conclusive in themse lves (because not all of the test conditions were equivalent) they do imply that the pow er output is greatly over-estimated for cases

PAGE 89

71 with high expander inlet temperatures. This suggest that an unaccounted heat transfer from the hot inlet fluid may be skewing the energy balance of the expander. Because of this, those particular experimental results have not been incorporated into this work. Admittedly, this expander was not the best choice for this size of system, as evidenced by the recommendations for expanders given in Chapter 4. It was, however, a choice between relative performance among tested devices and ease of adaptability to the ammonia-water working fluid. Had the anom alies just discussed not occurred and the efficiency was within the anticipated range of 20-30%, exhaust temperatures 5-10 C cooler could have been expected. Conclusion A demonstration of the key concept of the power-cooling cycle has been provided and the trends important to cooling production have been verified. However, the small scale of the experiment complicated the te sting conditions so full agreement was not achieved, neither was a truly convincing ex ample of combined power and cooling outputs. These complications were more evident in the turbine than in the other components where performance was poor and some readings were erroneous. Another issue is the fact that even with the minimum flow of the turbine, a single open nozzle, it regulated pressure to a lower level than that considered optimum by the theoretical analysis. On the other hand, the experimental setup was not designe d to be an economic success. Rather, it is a test-bed for explor ing operating issues with the power-cooling cycle and some observations from it are in cluded in the conclusions next chapter.

PAGE 90

72 CHAPTER 7 DISCUSSION AND CONCLUSIONS The conclusions of this work can be broadl y divided into two ca tegories: those that are derived from a theoretical analysis of the power-cooling cycle thermodynamics and those that result from the deviations enc ountered during the expe rimental study. Based on this information, the following conclusions regarding cooling production have been formed. In general, cooling production with this cycle is counter-productive to work output. This is a direct consequence of the need to reduce the entropy of the vapor in order to expand it to low temperatures. The effectiv e COP parameter, introduced previously, is used to quantify the trade-off of work a nd cooling and to select favorable cooling conditions. Characteristics of these optimum conditions are explored with regards to system implementation and operation. As for th e experimental results, they have been presented in Chapter 6 to verify many of the operating mechanisms of the power-cooling cycle. However, while these mechanisms are an aid for system design and evaluation, experimental testing indicates that they w ill have only secondary effects on the operation of a real system. The primary effects, whic h were largely unaccounted for, are described and their impact on system performance evaluated. Cooling Conditions During the discussion of the operating m echanisms of the power-cooling cycle it was noted that cooling production has a ma ximum value for a given heat source temperature. This maximum is a result of th e balance between vapor mass flow rate and

PAGE 91

73 minimum temperature from the exhaust. In this section, the balance of conditions are quantified by using the effective COP paramete r introduced in Chapter 3. Initially, the optimum amount of rectification is determin ed, then the overall en ergy advantage of the power-cooling cycle is evaluated by comp arison of cooling-optimized and workoptimized systems. Expander choice and expected performance has been discussed in Chapter 4. The general conclusion was that the expected e fficiency increases w ith power output, ranging from 60-70% for multi-kW displacement machines to +90% for multi-MW dynamic turbines. Turbines also cover the mid-output range with widely varying efficiencies, 6090%, where much of the variation depends on whether stock steam turbines are used or custom design takes place. For the remaini ng simulation results, efficiencies have been chosen to place bounds on the anticipated performance. Optimum Rectification For the heat rejection temperatures consider ed in this work, some rectification is necessary to achieve any pract ical cooling, so the determination of optimum cooling conditions is largely an issue of determini ng the optimum amount of rectification. The effective COP value presented in Chapter 3 is used to determine this balance, it is repeated here as Equation 7-1. rect with rect no cool effW W Q COP (7-1) The optimization strategy given a boiling te mperature, absorption temperature, and expander efficiency, is to chose the boiling pr essure, rectification temperature, and basic solution concentration that maximize the effec tive COP value. This results in the most

PAGE 92

74 energy effective trade-off of work and coo ling production given the specified inputs. Figure 7-1 is a plot of the maximum effectiv e COP as a function of boiling temperature. 2.5 3 3.5 4 4.5 5 6065707580859095100Boiling Temp. [C]Eff. COP Expander Efficiency = 60 % 90 % Figure 7-1. Maximum effective COP values where the work compone nt is the amount of work lost due to operation with rectifi cation vs. equivalent conditions with no rectification. The results of the maximum effective COP appear quite good. In the best cases, the work that is lost due to rectification pe nalties is effectively traded for 4-5 times the quantity of cooling. This is better perform ance than most conventional cooling systems, however, it is not quite the complete picture of energy efficiency since work production can also be optimized. Overall Optimum Cooling For a given heat source temperature, wo rk optimization results in the system configuration evolving toward a pure component working fluid Rankine cycle, that is a high ammonia concentration in the basic solution. General optimum characteristics for

PAGE 93

75 the power-cooling cycle have already been disc ussed and a brief comparison of the two is given in Table 7-1. Table 7-1. Typical operating characteristics for cooling and work optimized cycles. Optimization Vapor Mass Flow Fract ion Basic Solution Concentration Cooling 5-10 % approx. 0.50 Work +90 % approx. 0.95 For a more stringent evaluation, cooling production can be assessed by the amount of lost work that could be obtained from a work-optimized system. When considering this scenario the overa ll effective COP has the following formulation. / / /1coolcw overall workopt workoptwcool wcoolQR COP WW (7-2) Where the subscripts work opt and w/cool refer to parameters for the work optimized case and those with cooling production, respectively. The rightmost formulation represents the actual implementa tion with dimensionless parameters. The term Rc/w is the ratio of cooling to work. Figure 7-2 presents the maximum values of the overall, effective COP as a function of boiling temperature. Based solely on energy considerations and the assumptions inherent to Figure 7-2, at the best conditi ons nearly equal amounts of work must be forfeited for the amount of cooling gained. The ratio of cooling to net work output is also presented in Figure 7-2 to provide the relative magn itudes of each. Exhaust Temperature The exhaust temperature will determine su itable cooling applications, e.g. space conditioning vs. refrigeration. The correspond ing temperatures for the optimized cooling cases of Figure 7-2 are presented in Figure 7-3.

PAGE 94

76 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 6065707580859095100Boiling Temp. [C]Eff. COP0 0.2 0.4 0.6 0.8 1 1.2Ratio of Cooling/Net Work Output Eff. COP Cooling/Work Ratio 60% = Expander Efficiency 90% 90% 60% Figure 7-2. Maximum overall effective COP values as defined by Equation 7-2. -10 -5 0 5 10 15 6065707580859095100Boiling Temp. [C]Exhaust Temp. [C] Expander Efficiency = 90 % 60 % Figure 7-3. Corresponding exhaust temperatur es for the optimum conditions presented in Figure 7-2.

PAGE 95

77 However, these values are complicated by numerically maximizing the effective COP values and do not provide a complete picture of the exhaust temperatures attainable. Consideration of the relati onship between overall effective COP and other system parameters provides a better indication of ach ievable exhaust temperatures. Figure 7-4 attempts to show this relationship by plotti ng results of constant temperature operation in a boiling pressure-basic soluti on concentration plane. Thr ee parameters are plotted in Figure 7-4: the overall effective COP value, the expander exhaust temperature, and the vapor mass flow fraction (mvr/ms). The maximum effective COP value is centered within the 1.08 isoline at approximately a pre ssure of 1.03 MPa and a basic solution concentration of roughly 0.46. The correspond ing exhaust temperat ure and vapor flow fraction is -4.5 C and 7.8%, respectively. Fi gure 7-4 shows that the effective COP is much more sensitive to the vapor flow fracti on than the exhaust temperature. Therefore, with a mild penalty to effective COP values, a large range of exhaus t temperatures could be accessed. For example, while operating within the 1.08 effective COP isoline, the exhaust temperatures could vary between appr oximately -7 C and -2.3 C. Furthermore, if the effective COP were diminished by 7.5% (the 1.0 isoline), th e range of possible exhaust temperatures would be be tween roughly -12 C to 5.6 C. Implementation The preceding results identify preferred condi tions for cooling production. In this section some of the implications of these c onditions are considered as they relate to physical implementation of the power-cooling cycle.

PAGE 96

78 0.5 0.8 1.1 1.4 1.7 0.40.440.480.520.560.6Basic Solution ConcentrationBoiling Pressure [MPa] Vapor Mass Flow Fraction = 0 % 5 % 10 % 15 % 1.08 1.0 Overall Effective COP = 0.9 0.7 9o = Exhaust Temp. [C] 4.5o 0o -4.5o -7o -10o Figure 7-4. Design point map showing the rela tive sensitivity of overall effective COP to vapor mass flow fraction and exhaust temp erature. Sensitivity to mass flow is high, while with a mild penalty to effec tive COP a wide range of temperatures could be expected. Boiling temperature of 80 C. Vapor Quality Considering the optimum cooling conditions certain regions may cause exit quality concerns. Combinations of high boiling te mperatures, high expander efficiencies, and effective cooling production result in relatively low equilibrium exhaust qualities, 9095%. This is due to the fact that at these conditions th e entropy of the vapor is low enough that it can be expanded to conditions that have significant condensation. It was concluded in Chapter 4 that dyna mic machines, turbines, were the better expander choice for medium to large work outputs. However, these machines are sensitive to vapor quality, in terms of both operating efficiency and erosion damage.

PAGE 97

79 From examination of steam turbine performance, the effect of vapor quality on machine efficiency is mild, for example an exit quality of 90% would result in an expected efficiency penalty of 5-10% [50]. On the other hand, erosion of flow path components is a significant concern sin ce it could require ma chine rebuilding or replacement. For these cases cooling may have to be curtailed, either by limiting rectification or superheating, to avoid unacceptable conditions. Rectifier Implementation It was pointed out in Chapter 3 that the physical setup of the rectification process can have an effect on the penalties to work production because it can affect the quantity of vapor that continues on to the expander. Here, these effects are reconsidered in light of the conclusions regarding optimum rectification. In Chapter 3 upper and lower bounds were placed on the efficiency of the rectification process. It wa s pointed out that with high am ounts of rectification, lower rectifier exit temperatures, the difference be tween the two becomes excessive while there is not a significant difference with minimal rectification. The effects of minimum and maximum rec tifier efficiencies on the optimum overall effective COP is shown in Figure 7-5. From the figure it can be seen that compared to the minimum performance case, an approxima te 20% improvement could be obtained for most conditions. This benefit will be weighe d against the added complexity of a direct contact heat exchanger. Experimental Observations Correlation of the experimental results to simulated data revealed unexpected and unaccounted trends in system operation. Ma ny of these deviations overshadowed the

PAGE 98

80 basic operating mechanisms that were studied analytically. This se ction explains these deviations and any impact on system design. 0.4 0.6 0.8 1 1.2 1.4 6065707580859095100Boiling Temp. [C]Overall Eff. COP0 10 20 30 40 50Eff. COP Increase [%]COP with max. possible rectified vapor production COP with minimum rectified vapor production % Increase Figure 7-5. Effect of the minimum a nd maximum bounds of rectifier operation on effective COP values. Absorption Pressure The experimentally measured absorption pr essure was higher than that predicted by using saturation properties at the basic solution concentration and the measured temperature. This is not surprising since previous expe rimental investigation [70] showed similar results. However, part of the explanation is an unexpected coupling between the boiling temperature and basic solution concentration. Basic solution concentration increased with temperature, t hus increasing absorption pressure, due to the storage of weak soluti on in the separator.

PAGE 99

81 At steady state operation all of the amm onia and water leaving the absorber are returned to it so that the concentration of the solution in the absorber remains fairly constant. It would be expect ed then, even at different operating conditions, that the concentration remains at its original value. However, in the experi mental setup there is some stored liquid in the separa tor that causes this scenario to be out of balance. The liquid is retained to form a vapor seal in the separator and amounts to approximately 25% of the total working fluid mass. This is weak solution liquid, which is a larger percentage water than the basic solution, so by basica lly removing it from the working fluid, the basic solution concentration inevitably incr eases. Furthermore, the weak solution concentration is dependent on the boiling te mperature and this results in the coupling between boiling temperature and basic soluti on concentration. A formulation of this effect is given as Equation 7-3. _1basicorigweak basicnew x yx x y (7-3) Where the new basic solution concentration, xbasic new, is dependent on the original concentration, xbasic orig, the current weak solution concentration, xweak, and the fraction of the total mass of working fluid that is stored, y. The effect is plot ted in Figure 7-6 as a function of both boiling temperat ure and stored fraction. Also included in the figure are experimental measurements of the basic solution concentration at the approximate amount of storage. As shown in Figure 7-6, the change in basic solution concentration is more pronounced with higher boiling temperatures a nd stored fractions. Accounting for this effect when determining absorption pressure, Figure 7-7, explains much of the difference

PAGE 100

82 when simply assuming a constant concentration, 0.35 for this case. The remainder of the difference is attributed to non-equilib rium performance of the absorber. 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0102030405060Stored Fraction [%]Strong Concentration 60C Measurement 80C Measurement 100C Measurement Simulated data boiling temp. = 100 C 80 C 60 C Figure 7-6. Computed effect of weak solu tion storage on basic solution concentration. Experimentally measured basi c concentrations also shown. The crucial effect of absorption pr essure on cooling production has been mentioned, lower pressures are ce rtainly preferred. Therefore, if not taken into account, an undue pressure rise caused by this effect could limit c ooling production. The obvious solution, according to Figure 7-6, is to minimize the stored fraction.

PAGE 101

83 0.1 0.12 0.14 0.16 0.18 0.2 0.22 0.340.350.360.370.380.390.40.41Strong ConcentrationAbsorber Pressure [MPa] Computed Equlibrium Pressures Measured Absorber Pressure (+/-0.538 kPa) Figure 7-7. Computed absorption pressures taking into account th e changes of basic solution concentration compared with measured absorber pressures. Rectifier Pressure Effect An unanticipated decrease in boiling pr essure occurred between operations with and without the rectifie r, Figure 7-8. The drop in pressu re was due to condensation of vapor in the rectifier, which essentially behaved as a sec ond outlet (the first being the expander) for the high pressure vapor. Consis tent with this explanation is the observation that the pressure drop is mo st severe for the cases that have the highest amount of condensed vapor. Figure 7-9 is a plot of the corresponding amounts of condensed vapor for the same tests presented in Figure 7-8. For the points without re ctification the vapor lost between the separator exit and expander inlet is near zer o or is zero as expected. However, for the cases with re ctification the amount of vapor condensed in the rectifier increases with boiling temperature. Consid ering both Figures 7-8 and 7-9, the maximum

PAGE 102

84 pressure drop was approximately 0.058 MPa where roughly 30% of the vapor produced, was condensed. 0.3 0.35 0.4 0.45 0.5 0.55 50556065707580859095100Boiling Temperature [C]Boiling Pressure [MPa] Boiling P with rectif ication (+/-2.24 kPa) Boiling P no rectification (+/-2.24 kPa) Figure 7-8. Measured drop in boiling pressure due to rectifier operation. Using average values for the worst case, nom inally 95 C, the effect on the rectified vapor concentration amount to a decrease of 0.05% and a reduction in isentropic enthalpy drop of 1.7%. These are somewhat minor, th e largest impact, however, was a reduction of 4% in vapor mass flow rate. This ex tra flow rate loss stems from the lower concentration vapor that is produced as a re sult of the lower pressu re boiling condition. Because of the lower concentration vapor, 4% mo re of it is condensed in the rectifier as compared to the vapor that would have been generated if the pressure remained high. Again, this was a slightly unexpected effect and emphasizes the point that the vapor escaping through condensation in the rectif ier should accounted for in a system design.

PAGE 103

85 0 5 10 15 20 25 30 35 50556065707580859095100Boiling Temp. [C]Condensed Vapor in Rectifier [%] Vapor condensed, with rectification Vapor condensed, no rectification Figure 7-9. Amount of the produced vapor that was condensed in the re ctifier. This data corresponds to the results of Figure 7-8. Conclusions The parameters affecting cooling with this cycle have b een identified and thoroughly examined both analytically and expe rimentally. Operating trends have been verified and unforeseen effects identified fr om experiments. A demonstration of subambient expander exhaust has been provided wh ich to a limited extent, demonstrates the key concept of the power-cooling cycle. Base d on these cumulative investigations the practical operating region for cooling production has been outlined and preferred operating procedures identified. The fina l conclusions are summarized as follows. In determining the optimum degree of r ectification the trade-off between work sacrificed and cooling gained can be used. This effective coefficient of performance can be as high as 5 for the cooling gained relative to the work lost due to the rectifier.

PAGE 104

86 When considering the impact of cooli ng production on the optimum amount of work that could be produced, however, the equivalent effective COP is lower with maximums near 1.1. Characteristics of this overall optimum in clude basic solution concentrations of 0.40-0.55 and vaporization fractions of 5-10% Minimum expected temperatures are in the range of -4.5 C. However, with an approximate 10% penalty to the overall effective COP an 18 C range of exhaust temperatures can be accessed. As for expander implementation, in the smaller size range, displacement machines have advantages over turbines and ha ve reached +70% efficiencies with comparable operation with steam. In la rge applications there are examples of dynamic turbines operating with ammoniawater and achieving +90% efficiency. Applications in the mid-range will likely be relegated to lower efficiencies, 6080%, without resorting to custom designs in which case estimates of 90% efficiencies have been put forward. At combinations of high heat source te mperatures and optimum cooling production with high efficiency expander operation (t ypical of dynamic machines at larger sizes), condensate formation in the expander is predicted to be greater. For these cases, cooling production may n eed to be curtailed to avoid efficiency penalties in the expander and possible erosion damage. An approximate 20% increase in the overall effective COP value could be expected by incorporating a boiler-rectifier system that more closely approximates the maximum theoretical amount of vapor production. When off-performance effects are taken in to account, most of the operating trends have been confirmed by the experimental results. Additionally, aside from the noted exceptions, the simulation modeli ng does a reasonable job of estimating some experimentally measured parameters such as vapor flow rate and concentration. Simple equilibrium modeling of the abso rber is not sufficient to predict the absorption pressure. A portion of this diffe rence is caused by the change in basic solution concentration due to storage of weak solution. This effect is more pronounced at higher boiling temperatur es and larger stored fractions. The condensation of vapor in the rectifier es sentially acts as a secondary exit for the vapor thus lowering the boiling pressure when compared to equivalent cases without rectification. For the worst case encountered the pressure decrease was 0.058 MPa and this caused a corresponding 4% decrease in vapor mass flow rate. The apparent efficiency of the turbine used in the experimental work seemed to get worse as optimum conditions for cooling we re approached. No conclusive reason has been identified except for unaccounted heat transfer either fr om the inlet vapor or to the exhaust stream.

PAGE 105

87 It is difficult to quantify all of the ad vantages and disadvant ages that the powercooling concept has in comparison to othe r options. However, the characteristics described in this work can be used to ev aluate the energy effec tiveness of the powercooling concept. Furthermore, several desi gn and modeling issues ha ve been raised so that they will be available for the next generation of studies into this concept.

PAGE 106

88 CHAPTER 8 RECOMMENDATIONS The previous chapter presented the cumula tive results of this work. Considering those as a preface, the current chapter explores future suggestions for the power and cooling concept. Experimental Testing In retrospect the most severe underestim ation when designing the experiments for this work was the impact of the small va por flow rate on expander operation. Not only did this complicate the selection of a suitabl e device, but it resulte d in a correspondingly small power output that was difficult to measure. It can be inferred from Chapter 4, which concerns expander design considerations, that the expander used for this work was not th e best choice in terms of performance. It was, however, a choice between relative performance among tested expanders and adaptability to the ammonia-water working fluid. Had the anomalies previously discussed not occurred, the conditions were se t to have a more dramatic demonstration. For the small size of the current se tup and likely any other prototype, the recommendation would be to use a displa cement expander rather than the dynamic machine used in this work. At the outset of this work, the dynamic machine which was eventually used had certain practical adva ntages over the displacement devices being considered. Primarily these were simpler ope ration with adequate efficiency, no need for oil, no shaft seal, and a very simple load ing mechanism. However, considering the

PAGE 107

89 difficulties encountered with making an accurate power output measurement, the complications of displacement machines seem quite manageable. Power measurements were complicated by two factors, the first is that small amounts of inadvertent heat transfer (losses to the environment) could cause significant errors in the computation of power produ ced based on thermodynamic measurements. Secondly, the small quantity of power produced quite a small reaction torque (this was complicated further by the high shaft speeds) that was difficult to measure accurately. Increased flow (larger potent ial output) would improve thin gs by making the expander inlet/exhaust states less sensitive to heat transf er losses. Also with a larger power output, direct measurements of the reaction torque could be simplified. This could also be improved if the shaft speeds were lower, as might be the case with a displacement expander. Aside from a more efficient expander, the current experimental system would benefit most from improved solution pump pe rformance. Under the best circumstances the measured flow rate (approximately 0.5 gpm ) was less than half of the rated capacity of the pump (1.3 gpm). With an increase in vapor production the boiler pressure would rise for a given boiler exit temperature a nd nozzle configuration. This would be an improvement for cooling production and likely even for the performance of the current turbine. There appear to be two problems with the current pump configuration, the first is the fact that the working fluid is at or ve ry near saturation, so a ny significant pressure drop could induce cavitation, secondly, the pum p inlet piping from the absorber still serves as a restriction.

PAGE 108

90 Practical Application In considering the possible applications of the power-cooling cycle, several facts about cooling production need to be cons idered. Not the least of which is the consideration of the amount of energy need ed to produce cooling in this manner. The overall effective COP values presented in Figure 7-2 indicate that nearly equal amounts of work (that could have been produced with an optimized system) are sacrificed for cooling produced. As suggested by thos e results and reemphasized here, cooling production in this manner may not be as en ergy efficient as conventional work-driven cycles. Therefore, it appears that for this concept to be useful the convenience of combined power and cooling will need to outweigh this disadvantage. Additional thoughts concerning the future evaluation and possible application of the power-cooling cycle follow. ORC Comparison From the conclusions of this work, it app ears that any use of this concept will be primarily as a power cycle w ith the secondary feature of cooling production. A solid performance comparison of this cycle w ith the more-proven ORC technology would likely give an indication for the potential use of this cycle. The ammonia-water based power cycle was reintroduced by Kalina [23] to take advant age of the mixtures boiling temperature glide and reduce heat transfer irreversibilities when compared to steam Rankine cycle operation. This advant age may be extended to lower operating temperatures where organic Rankine cycles are typically employed, however, these fluids generally have smaller latent he at values than water, so the advantage may be reduced or even eliminated. A comparison of this power-cooling cycle and its chief competition, organic Rankine cycles, would give an indication of its me rit as a power cycle alone.

PAGE 109

91 Another basis of comparison would be th at ORC engines have proven, efficient expander performance in small sizes, and the power-cooling concept may be at a disadvantage in this area because of the lack of efficient designs, or evidence of them, for ammonia-water. The underlying reasons for this were discussed in Chapter 4. The opposite may be true for larger output systems where turbines for organic fluids would be inordinately large, however, considering that most low te mperature applications, solar thermal, geothermal, etc., and the whole distributed generation concept have been focused on small size ranges, this may be of little consolation. Cooling Production As for the cooling process itself, the cooli ng produced with this cycle is sensible rather than latent. Therefore, analogous to the heat addition pro cess, this may provide some advantage in exergy efficiency. The s uggestion here is that to effectively use the cooling from this cycle, the temperature profile of the exhaust stream will have to be taken into account. This will likely requi re counter-flow heat exchangers, but more importantly it may reduce the heat transfer irre versibilities as compared to latent-based cooling, which has an isothermal region. If this could be shown, then the argument could be made that this sensible cooling could sati sfy a load with fewer irreversible losses. This could be a particular advantage when compared to absorption cooling cycles where water (lithium-bromide system) and a mmonia (aqua-ammonia system) are commonly used as refrigerants. Both have high late nt heats so the irreversibilities may be significant. As for applications for the cooling out put, the obvious and immediate thoughts, for example serving both the power and cooling n eeds of a residence, are likely not a good match for the characteristics of the cycle. Wh en the ratio of cooli ng to power and overall

PAGE 110

92 effective COP values are considered, one realiz es that the amount and cost of cooling is not suitable for these traditional applicati ons. For example, the quantity of cooling produced is at most equal to the power output, but would typically be less to attain higher operating efficiencies. This is almost th e opposite of what the expected demand for a residence might be. Furthermore, the cost of cooling, in terms of work that could have been produced (overall effective COP), indicate that serving large co oling loads in this manner may be inefficient. Rather than these uses, it may be more fru itful to identify nich e applications that this cycle could be applied to. Applications should be considered where the cooling is used internally in the power cycle itself to either improve overall performance or reduce equipment count. For example, considering th at the amount of cooling output can be just a fraction of the work output, referring to the simulation results of Figure 3-3, a configuration of the power-coo ling cycle could be implemented where this fraction of cooling could be used to keep the generator or other power conditioning device cool. This could provide some performance improveme nt or simply eliminate the need for an additional cooling system. This kind of practi cal advantage will need to be identified and exploited. Conclusion Summing these considerations results in the power-cooling c oncept being a power cycle that may be competitive with other heat engines and could produce small amounts of cooling, or even intermittent cooling, w ith some cost to performance. Where efficiency is of concern, a balance will need to be struck between the benefit associated with cooling and the penalty incurred due to its production. The effective COP term can be used to evaluate the suitability of modi fying the expansion conditions to accommodate

PAGE 111

93 cooling in this manner. Furthermore, possi ble applications for th is cycle may include those where cooling is not a primary output, but it is used to improve overall efficiency or simplify implementation.

PAGE 112

94 APPENDIX A PROPERTY EVALUATION For the evaluation of the working fluid pr operties the method described by Xu and Goswami [26] is employed. The Gibbs free energy method is used to determine mixture properties while empirical equations for bubble and dew point temperatures are used to determine phase equilibrium. Good agreemen t between this method and experimental data have been shown [26]. Pure Component Properties Beginning with the components needed to determine mixture properties, the Gibbs free energy for a pure component is given as Equation A-1. 00000 TPT P P TPTc GhTscdTdPTdT T (A-1) The subscript 0 indicates reference state propertie s. The empirical relations used by Ziegler and Trepp [71] for constant pre ssure specific heat and specific volume are used here, Equations A-2 through A-5. Liquid phase relations: 2 1234 l A APATAT (A-2) 2 123 l PcBBTBT (A-3) Gas phase relations: 2 3 24 1 31111 gC CCP RT C PTTT (A-4) 2 123 g PcDDTDT (A-5)

PAGE 113

95 Substitution and integration of Equations A2 and A-3 into Equation A-1 results in the following expression for the reduced liquid phase Gibbs free energy. 2233 3 2 ,0,01,0,0,0 22 3 12,0,0 ,0 222 2 134,0,023 ln 2 2lll rrrrrrrrrr r r rrrrrr r rrrrrrB B GhTsBTTTTTT BT T BTBTTTTT T A AATATPPPP (A-6) Similarly, substitution and integration of Equations A-4 and A-5 into Equation A-1 results in the following expression for th e reduced gas phase Gibbs free energy. 2233 3 2 ,0,01,0,0,0 22 3 12,0,0 ,0 ,0 ,0,0 1,023 33411 ,0,023 lnln 2 4312ggg rrrrrrrrrr r r rrrrrrr r r rrrr rr rr rrrrD D GhTsDTTTTTT D TP DTDTTTTTT TP PPTP PP CPPCC TTTT ,0,0 1112 ,0,0 33 3 ,0,0 4 111112 ,0,011 1211 3rr rr rrr r rrrPT TT PPT CP TTT (A-7) Table A-1 presents the coefficient values for Equations A-6 and A-7. The reduced properties are defined in Equa tions A-8 through A-13. The associated reference values are presented in Table A-2. r BT T T (A-8) r BP P P (A-9) r BG G R T (A-10)

PAGE 114

96 r Bh h R T (A-11) rs s R (A-12) B r BP R T (A-13) Table A-1. Coefficient and reference state values for ammonia and water. CoefficientAmmoniaWater A10.039714230.02748796 A2-1.790557E-05-1.016665E-05 A3-0.01308905-0.004452025 A40.0037528360.000838925 B116.3451912.14557 B2-6.508119-1.898065 B31.4489370.2911966 C1-0.010493770.02136131 C2-8.288224-31.69291 C3-664.7257-46346.11 C4-3045.3520.0 D13.6736474.01917 D20.09989629-0.0517555 D30.036176220.01951939 hl r,04.87857321.821141 h g r,026.46887360.965058 sl r,01.6447735.733498 s g r,08.33902613.45343 Tr,03.22525.0705 Pr,02.0003.000 Table A-2. Reference values for reduced property computation. TB 100 K PB 10 bar R 8.314 kJ/(kmol K)

PAGE 115

97 For a pure component, the molar specific properties of enthalpy, entropy, and volume are related to the reduced Gibbs free energy through Equations A-14 through A16. 2rr Br rr P G hRTT TT (A-14) rr r P G sR T (A-15) rBr Br TRTG PP (A-16) Liquid Mixture Properties The Gibbs excess energy function for liquid mixtures allows for deviation from ideal solution behavior. Th e relation used here is the same one used by Zeigler and Trepp [71] and is given in Equation A-17. 2 12321211E rGFFxFxx (A-17) The parameter x in Equation A-17 is the ammoni a mole fraction of the ammoniawater mixture. The coefficients are defined in Equations A-18 through A-20. 56 11234 2rrr rrEE FEEPEEPT TT (A-18) 1112 278910 2rrr rrEE FEEPEEPT TT (A-19) 1516 31314 2r rrEE FEEP TT (A-20) In turn, the coefficients used in E quations A-18 through A-20 are those proposed by Ibrahim and Klein [72] a nd are given in Table A-3.

PAGE 116

98 Table A-3. Coefficient values used to compute excess properties. E1-41.733398 E90.387983 E20.02414 E100.004772 E36.702285 E11-4.648107 E4-0.011475 E120.836376 E563.608967 E13-3.553627 E6-62.490768 E140.000904 E71.761064 E1524.361723 E80.008626 E16-20.736547 The excess enthalpy, entropy and volume fo r liquid mixtures is presented in Equations A-21 through A-23. 2 ,rE E r Br rr P xG hRTT TT (A-21) ,rE E r r P xG sR T (A-22) ,rE E Br Br TxRTG PP (A-23) Finally, the liquid mixture properties can be computed with Equations A-24 through A-27. 1lllE mawhxhxhh (A-24) 1lllEmix mawsxsxsss (A-25) ln1ln1mixsRxxxx (A-26) 1lllE mawxx (A-27) Where the subscripts m a and w indicate mixture, ammonia, and water properties respectively.

PAGE 117

99 Vapor Mixture Properties Vapor mixtures of ammonia and water are treated as ideal solutions. The following formulations are used to compute th e mixture thermodynamic properties. 1 g gg mawhxhxh (A-27) 1 g ggmix mawsxsxss (A-27) 1 g gg mawxx (A-27) Equilibrium Conditions To determine the phase equilibrium of ammonia-water mixtures, the bubble and dew point temperatures are computed from th e explicit relationships of reference [73]. The relations themselves and associated de finitions are presented as Equations A-28 through A-31. 710 11lni j C bubbleCiij ijP TTCCx P (A-28) 64 11ln1.0001lni j C dewCiij ijP TTaAx P (A-29) 4 1i Ccritwatercriti iTTax (A-30) 8 1expj Ccritwatercriti iPPbx (A-31) Note that in Equations A-28 through A-31 th e temperatures are in F and pressures have units of psia. The coefficient values are presented in Table A-4.

PAGE 118

100 Table A-4. Coefficient values for the de termination of mixture bubble and dew point temperatures. Ciai153.634521459000000153.170553460000000 -13.030554389200000-11.770568746100000 -1.148452829910000-1.781263559570000 0.5503580944470000.647385455059000 -0.075345014842700-0.071995075189800 0.0048111666267000.002854239507860 -0.000120433757177 Cij-462.460321366000-9668295.895040000-3583589.8687500004807.07241098 23739.9986309000005922081.87086000012243265.38150000013565.10033090 -194504.352920000000-1432405.521250000-22307970.015600000-466407.78083200 639383.528867000000421.44312220822896656.8499000002827083.44764000 -523748.057636000000-14560.354925000-12483324.809100000-8469715.15799000 -2328271.47551000000053051.4495633002813311.71633000014459588.89620000 7562418.534990000000382763.793582000-248.783804168-14281087.53310000 7596403.5967800000002132412.4695900000-3064.8207065800-54497.09733360000 -1684002.644820000000-3699199.659140000071.79547520523.97454953787 126.9655807280003688365.225460000051780.6666590000-77.02684646900 -2090.452705740000-1975122.3929600000-209714.8998560000541.19105807000 1993.171011660000440201.4460680000405011.9853550000-1696.60270972000 100706.510396000000-33.5343446156-428310.46156600001713.45942707000 -687388.808612000000601.8785866890238153.69832600004019.01019872000 -14844.792800400000113.7620645460 19481.009455100000-258.7504969220 -12107.079450100000311.0025852180 2966.928043860000-123.9179934540 -0.170806170177-123.4806274920 3.481828592990154.3750421140 -27.795758774300-48.5083828701 Aij194.793913463000-4.7886691858100-0.90857587517000-0.019166461330400 74.236124188000-0.2254167334760-0.35675269114700-0.001701425386700 9.84103819552013.01754473670000.023806727550200.001954417029830 0.4368438527456.15865641170000.004955939339520.002805333489370 -74.3508283362000.7897403371410-0.000718635741530.001389943656300 -33.2941879809000.0321510834958-0.025102638353300.000116422611616 acrti,ibcrti,i205.88890000.368105523897 280.9305560-3.667954887500 -317.013888946.600047080900 263.1944440-262.921061996000 732.995369360000 -1076.061348900000 797.948078048000 -235.903904222000

PAGE 119

101 Computer Implementation The relations presented for determining mixture properties and phase equilibrium have been coded in C++. This section presen ts the subroutine code used in this work to compute the saturation temperatures, enthalpy, entropy, and specific volume. Variable names should be self-explanatory because th e labeling system is consistent with the relations already presented. Saturation Temperatures This subroutine computes the saturation temperatures, bubble and dew points, for a specified pressure and mixture concentration. //////////////// //////////////// //////////////// //////////////// //////////////// double SatTemp(double P, double xa) { double Ppsi=P/0.006894757; // Convert P to psi // Determine Tc int i=0; double sum=0; for (i=0; i<=3; ++i){ sum=sum+(a[i]*pow(xa,(i+1))); } Tc=Tcw-sum; // Determine Pc i=0; sum=0; for (i=0; i<=7; ++i){ sum=sum+(b[i]*pow(xa,(i+1))); } Pc=Pcw*exp(sum); // Determine bubble temp, Tbub i=0; sum=0; int j=0; double sum2=0; for (i=0; i<=6; ++i){ j=0;

PAGE 120

102 sum=0; for (j=0; j<=9; ++j){ sum=sum+(Cij[i][j]*pow(xa,(j+1))); } sum2=sum2+((Ci[i]+sum) *pow(log(Pc/Ppsi),(i+1))); } Tbub=(((Tc-sum2)-32)/1.8)+273.15; // Determine dew temp, Tdew i=0; sum2=0; for (i=0; i<=5; ++i){ j=0; sum=0; for (j=0; j<=3; ++j){ sum=sum+(Aij[i][j]*(pow(log(1.0001-xa),(j+1)))); } sum2=sum2+((ai[i]+sum) *pow(log(Pc/Ppsi),(i+1))); } Tdew=(((Tc-sum2)-32)/1.8)+273.15; return (Tbub, Tdew); } //////////////// //////////////// //////////////// //////////////// //////////////// Enthalpy This subroutine accepts values for pressure, temperature, and mixture concentration, with which it computes the va por phase and liquid phase enthalpies. The determination of phase is not performed here; that is handled in the calling program which also determines the appropriate enthalpy, vapor or liquid, to use. //////////////// //////////////// //////////////// //////////////// //////////////// double PTXh(double P, double T, double xa) { // Reduced Temps and Pressures Tr=T/Tb; Pr=(P*1000)/Pb; //Convert MPa to kPa before dividing // Ammonia mole fraction ya=xa/((xa/Ma+(1-xa)/Mw)*Ma);

PAGE 121

103 // Mixture molecular mass Mm=ya*Ma+(1-ya)*Mw; // Gas phase properties hga=-R*Tb*(-hgroa+D1a*(Troa-T r)+0.5*D2a*(Troa*Troa-Tr*Tr) +(D3a/3)*(pow(Troa,3)-pow(Tr,3))-C1a*(Pr-Proa) +4*C2a*(Proa/pow(T roa,3)-Pr/pow(Tr,3)) +12*C3a*(Proa/pow(Troa,11)-Pr/pow(Tr,11)) +4*C4a*(pow(Proa,3)/pow(Tro a,11)-pow(Pr,3)/pow(Tr,11))); hgw=-R*Tb*(-hgrow+D1w*(Trow -Tr)+0.5*D2w*(Trow*Trow-Tr*Tr) +(D3w/3)*(pow(Trow,3)-pow(Tr,3))-C1w*(Pr-Prow) +4*C2w*(Prow/pow(Trow,3)-Pr/pow(Tr,3)) +12*C3w*(Prow/pow(Trow,11)-Pr/pow(Tr,11)) +4*C4w*(pow(Prow,3)/pow(Tro w,11)-pow(Pr,3)/pow(Tr,11))); hgm=(ya*hga+(1-ya)*hgw)/Mm; // Liquid phase properties hLa=-R*Tb*(-hLroa+B1a*(Troa-T r)+0.5*B2a*(Troa*Troa-Tr*Tr) +(B3a/3)*(pow(Troa,3)-pow(Tr, 3))+(A4a*Tr*Tr-A1a)*(Pr-Proa) -0.5*A2a*(Pr*Pr-Proa*Proa)); hLw=-R*Tb*(-hLrow+B1w*(Trow -Tr)+0.5*B2w*(Trow*Trow-Tr*Tr) +(B3w/3)*(pow(Trow,3)-pow(Tr, 3))+(A4w*Tr*Tr-A1w)*(Pr-Prow) -0.5*A2w*(Pr*Pr-Prow*Prow)); hE=-R*Tb*((1-ya)*ya*(-E1-E 2*Pr-2*E5/Tr-3*E6/pow(Tr,2) +(-E7-E8*Pr-2*E11/Tr3*E12/pow(Tr,2))*(2*ya-1) +(-E13-E14*Pr-2*E15/Tr-3*E16/pow(Tr,2))*pow((2*ya-1),2))); hLm=(ya*hLa+(1-ya)*hLw+hE)/Mm; return hgm, hLm; } //////////////// //////////////// //////////////// //////////////// //////////////// Entropy This subroutine computes vapor and liquid phase entropy values given the inputs of pressure, temperature, and mixt ure concentration. As with the enthalpy subroutine, phase determination is not considered here but is dealt with in the calling program. //////////////// //////////////// //////////////// //////////////// ////////////////

PAGE 122

104 double PTXs(double P, double T, double xa) { // Reduced Temps and Pressures Tr=T/Tb; Pr=(P*1000)/Pb; //Convert MPa to kPa before dividing // Ammonia mole fraction ya=xa/((xa/Ma+(1-xa)/Mw)*Ma); // Mixture molecular mass Mm=ya*Ma+(1-ya)*Mw; // Entropy of mixing smix=-R*(ya*log(ya+.000001)+(1-ya)*log(1.000001-ya)); // Gas phase properties sga=-R*(-sgroa-D1a*log(Tr/Troa)+D2a*(Troa-Tr) +0.5*D3a*(Troa*Troa-Tr*Tr)+log(Pr/Proa) +3*C2a*(Proa/pow(T roa,4)-Pr/pow(Tr,4)) +11*C3a*(Proa/pow(Troa,12)-Pr/pow(Tr,12)) +(11/3)*C4a*(pow(Proa,3)/pow(T roa,12)-pow(Pr,3)/pow(Tr,12))); sgw=-R*(-sgrow-D1w*log( Tr/Trow)+D2w*(Trow-Tr) +0.5*D3w*(Trow*Trow-Tr*Tr)+log(Pr/Prow) +3*C2w*(Prow/pow(Trow,4)-Pr/pow(Tr,4)) +11*C3w*(Prow/pow(Trow,12)-Pr/pow(Tr,12)) +(11/3)*C4w*(pow(Prow,3)/pow(T row,12)-pow(Pr,3)/pow(Tr,12))); sgm=(ya*sga+(1-ya)*sgw+smix)/Mm; // Liquid phase properties sLa=-R*(-sLroa-B1a*log (Tr/Troa)+B2a*(Troa-Tr) +0.5*B3a*(Troa*Troa-Tr*Tr)+(A3a+2*A4a*Tr)*(Pr-Proa)); sLw=-R*(-sLrow-B1w*log( Tr/Trow)+B2w*(Trow-Tr) +0.5*B3w*(Trow*Trow-Tr*T r)+(A3w+2*A4w*Tr)*(Pr-Prow)); sE=-R*(1-ya)*ya*(E3+E4*Pr-E 5/(Tr*Tr)-2*E6/pow(Tr,3) +(E9+E10*Pr-E11/(Tr*Tr)2*E12/pow(Tr,3))*(2*ya-1) +(-E15/(Tr*Tr)-2*E16/pow(Tr,3))*pow((2*ya-1),2)); sLm=(ya*sLa+(1-ya)*sLw+sE+smix)/Mm; return sgm, sLm; }

PAGE 123

105 //////////////// //////////////// //////////////// //////////////// //////////////// Specific Volume This subroutine returns values for the liquid and vapor phase specific volume. Phase determination is handled in the calling program. //////////////// //////////////// //////////////// //////////////// //////////////// double PTXv(double P, double T, double xa) { // Reduced Temps and Pressures Tr=T/Tb; Pr=(P*1000)/Pb; //Convert MPa to kPa before dividing // Ammonia mole fraction ya=xa/((xa/Ma+(1-xa)/Mw)*Ma); // Mixture molecular mass Mm=ya*Ma+(1-ya)*Mw; // Liquid specific volume components vLa=((R*Tb)/(Pb))*(A1a+A2a*Pr+A3a*Tr+A4a*Tr*Tr); vLw=((R*Tb)/(Pb))*(A1w+A2w*Pr+A3w*Tr+A4w*Tr*Tr); vLE=((R*Tb)/(Pb))*(ya*(1-ya)* (E2+E4*Tr+(E8+E10*Tr)*(2*ya-1) +E14*pow((2*ya-1),2))); vLm=(ya*vLa+(1-ya)*vLw+vLE)/Mm; // Vapor specific volume components vga=((R*Tb)/(Pb))*(Tr/Pr+C1a+C2 a/(pow(Tr,3))+C3a/(pow(Tr,11)) +C4a*Pr*Pr/(pow(Tr,11))); vgw=((R*Tb)/(Pb))*(Tr/Pr+C1w+C2 w/(pow(Tr,3))+C3w/(pow(Tr,11)) +C4w*Pr*Pr/(pow(Tr,11))); vgm=(ya*vga+(1-ya)*vgw)/Mm; return vLm,vgm; } //////////////// //////////////// //////////////// //////////////// ////////////////

PAGE 124

106 APPENDIX B MODEL FORMULATION The thermodynamic model used to generate the simulated results of this work is described here. First, the i ndividual component formulations are presented, and then the implementation of these into pr ogramming code is presented. Thermodynamic Formulations The flow diagram used for the modeling is pr esented in Figure B-1. It is identical to Figure 3-1 except now the individual stat e points have been numbered. Also reproduced here is Table 3-1 (as Table B-1) wh ich describes the stream labels in Figure B-1. The assumptions and procedures outline d in Chapter 3, however, are not repeated here, but it may be benefici al to review them in conjunction with the following descriptions. Formulation begins in the absorber where two inputs are given, the absorber exit temperature and basic solution c oncentration. Assuming the basic solution is at or very near saturation, property data is used to de termine the system low pressure. System high pressure is usually specified. With the pr essure ratio known, an energy balance across the pump yields Equation B-1. 1 1 2 2h h h hs pump (B-1) Using property data, state point 2 can be determined (using Pboiler, xs, and h2). Moving to the boiler, where th e pressure and temperature ar e known, the vapor and weak solution ammonia mass concentrations and properties can be determined assuming saturated conditions at boiler exit, states 5 a nd 6. Furthermore, the mass flow rates of

PAGE 125

107 each are found from a mass balance of the bo iler and saturated equilibrium conditions, Equation B-2. 5 6 5 1 3 5 3 61x x x x m m m m (B-2) Heat Source Recovery Heat Exchanger Throttle Boiler Absorber Coolant Separator Coolant Rectifier Expander vr Solution Pump v wb w wr s Superheater Heat Source Cooling Heat Exchanger Cooled Fluid 1 2 3 4 5 7 8 6 11 12 13 14 9 10 Figure B-1. Schematic used for the theoretical modeling. Table B-1. Flow identification for the configuration of Figure B-1. Identifier/ Subscript Description s Basic (strong) solution flow from absorber through boiler v Vapor flow produced from pa rtial vaporization in boiler vr Rectified vapor passing through tu rbine and cooling heat exchanger w Weak (in ammonia) solution liquid returning to absorber wr Weak condensate formed in rectifier wb Weak liquid produced from pa rtial vaporization in boiler

PAGE 126

108 With this information the boiler heat transfer is determined. From a boiler energy balance Equation B-3 can be determined. 665533boilerQmhmhmh (B-3) Note that at the moment the heat source medium is not taken into account, i.e. qboiler represents only the amount of heat accepte d by the system and does not account for the efficiency or effectiveness of that heat transf er. Also note that stat e point 4 is not used, it is assumed that it is a two-phase mixtur e which is divided into states 5 and 6. Now computation proceeds to the rectifie r where the vapor inlet state is known, state 6, and the rectifier ex it temperature is specified, T11. Also, points 7 and 11 are interrelated through saturated equilibrium conditions, Equation B-6. 7 11 7 6 6 7 6 111x x x x m m m m (B-6) There is now enough information to comp lete the rectifier energy balance and determine the rectifier heat transfer using the following expression. rectifier66111177Qmhmhmh (B-7) There is now enough information to return to the weak solution exit from the boiler. At this junction th e weak solution from the boile r, rectifier condensate, and cooling strong solution are mixed. State 8 is determined from the junction mass and energy balances, Equations B-8 and B-9. 837mmm (B-8) 3377 8 8mhmh h m (B-9)

PAGE 127

109 For the recovery heat excha nger exit conditions, an eff ectiveness is assumed, and based on the maximum heat transf er that can occur, the exit states can be determined. Assuming counter-flow operation, the limiting exit conditions are either that T3 is raised to T8 or T9 is cooled to T2 so the maximum heat exchange is the minimum of Equation B10. max recovery33max2889minminofor Qmhhmhh (B-10) Subsequently, the recovery unit exit stat es are determined by Equations B-11 and B-12. recoverymax recovery 32 3Q hh m (B-11) recoverymax recovery 98 8Q hh m (B-12) Of course each end state above has the possi bility of being a two phase mixture so the enthalpies at those points, 3 and 9, s hould reflect mixture properties. The weak solution throttle is assumed to be isenthalpic, Equation B-13. 10 9h h (B-13) This completes the strong and weak portions of the cycle, now the vapor circuit can be completed. The vapor entering the superhea ter is assumed to be saturated, state 11. As with the rectifier operation, the superhea ter operation is determined by specifying the exit temperature. Since pressure and c oncentration are also known, the other thermodynamic properties can be determined. Th is brings computation to the inlet of the expander. An isentropic expander efficiency is assumed and the ex it conditions are found through Equation B-14.

PAGE 128

110 s turbineh h h h13 12 12 13 (B-14) Having determined the expander exit enthal py, there is sufficient information to determine its exhaust temperature (using Pabsorber, xvr, and h13). Based on this exhaust temperature, the cooling heat exchanger may not be active if the temperature is not below the limit specified for cooling production. For the analysis of c ooling production this threshold temperature was 15 C. If T13 is higher than this, then the cooling heat exchanger has no effect. If the exhaust temp erature is below the threshold the vapor is heated to another specified temperature, typi cally the threshold temperature. This of course fixes the properties at point 14 and the cooling obtai ned can be determined from an energy balance of the cooling heat exchanger, Equation B-15. 1314coolingvrQmhh (B-15) Finally, all of the conditions needed to com pute the heat rejected in the absorber are know. From an energy balance of the absorber the rejected heat can be determined from Equation B-16. 1010141411absorberQmhmhmh (B-16) This concludes the calculations needed to solve for the conditions in this configuration of the pow er-cooling cycle, attention is now turned to the coding of these formulations for computer execution. Computer Implementation This section presents much of the comput er code used to compute cycle operating parameters. The subroutines that follow, which are built upon the fundamental property routines presented in Appendix A, have been used in various combinations to compute cycle performance. Along with the br ief description preceding each routine,

PAGE 129

111 determination of its function should be self-e vident. Many of the ro utines are used to determine properties using inputs other th an pressure, temperature, and mixture concentration (the inputs for the fundamental routines presented in Appendix A). To do this an iterative search routine is us ed, which employs the bisection method. Saturated Liquid Pressure This subroutine computes the liquid satu ration pressure given a temperature and liquid concentration as inputs. It is typically called to dete rmine absorber pressure since basic solution concentration and ab sorption temperature are specified. //////////////// //////////////// //////////////// //////////////// //////////////// double TXSatLiq(double T, double xa) { double Plmin=0.005; // Search interval, may need adjustment double Plmax=20; // P's in MPa double Plmid; double fPlmid=1; SatTemp(Plmax,xa); double fPlmax=Tbub-T; SatTemp(Plmin,xa); double fPlmin=Tbub-T; int i=0; while (sqrt(fPlmid*fPlmid) > .00001){ Plmid=(Plmin+Plmax)/2; SatTemp(Plmid,xa); fPlmid=Tbub-T; ++i; if (i > 100){ PsatL=0; return PsatL; } if ((sqrt(pow((Plmax-Plmin),2)))<0.0001){ PsatL=Plmax; return PsatL; } if ((fPlmin*fPlmid) <= 0){

PAGE 130

112 // Root is in the left half interval Plmax=Plmid; fPlmax=fPlmid; } else { // Root is in the right half interval Plmin=Plmid; fPlmin=fPlmid; } } PsatL=Plmid; return PsatL; } //////////////// //////////////// //////////////// //////////////// //////////////// Two-Phase Mixture Determination This routine computes the mixture quali ty and the vapor and liquid saturation concentrations given the pressure, temperature, and the original solution concentration. This routine is commonly employed to determ ine the mass flows and concentrations at the separator exit, states 5 and 6. It is also used at the rect ifier exit to determine states 7 and 11. //////////////// //////////////// //////////////// //////////////// //////////////// double PTXsatx(double P, double T, double xv) { // Check for single phase SatTemp(P,xv); if (T<=Tbub){ xq=0; return xq; } if (T>=Tdew){ xq=1; return xq; } // Must be two-phase, begin searching

PAGE 131

113 double xmin=0.01; // Search interval double xmax=1; double xLiq; double xGas; double xmid; double fxmid=1; SatTemp(P,xmax); double fxmax=Tbub-T; SatTemp(P,xmin); double fxmin=Tbub-T; int i=0; // Begin iterations to find xLiq, i.e. set T=Tbub while (sqrt(fxmid*fxmid) > .0001){ xmid=(xmin+xmax)/2; SatTemp(P,xmid); fxmid=Tbub-T; ++i; if (xmin 100){ xq=0; cout << "xLiq iteration limit" <
PAGE 132

114 xmax=1; xmin=0.01; SatTemp(P,xmax); fxmax=Tdew-T; SatTemp(P,xmin); fxmin=Tdew-T; i=0; // Begin iterations to find xGas, i.e. set T=Tdew while (sqrt(fxmid*fxmid) > .0001){ xmid=(xmin+xmax)/2; SatTemp(P,xmid); fxmid=Tdew-T; ++i; if (xmin 100){ xq=0; cout << "xGas iteration limit" <= 1){ xq=1; }

PAGE 133

115 else { xq=(xv-xLiq)/(xGas-xLiq); } return xq,xLiq,xGas; } //////////////// //////////////// //////////////// //////////////// //////////////// Saturated Liquid Concentration This routine computes the saturated li quid concentration given a pressure and temperature. It is used to determine if conditions are within the two phase region. //////////////// //////////////// //////////////// //////////////// //////////////// double PTSatLiq(double P, double T) { double xmin=0; // Search interval double xmax=1; double xmid; double fxmid=1; SatTemp(P,xmax); double fxmax=Tbub-T; SatTemp(P,xmin); double fxmin=Tbub-T; int i=0; while (sqrt(fxmid*fxmid) > .00001){ xmid=(xmin+xmax)/2; SatTemp(P,xmid); fxmid=Tbub-T; ++i; if (i > 100){ xsatL=0; return xsatL; } /*if ((sqrt(pow((xmax-xmin),2)))<0.0001){ xsatL=xmax; return xsatL; }*/ if ((fxmin*fxmid) <= 0){ // Root is in the left half interval xmax=xmid;

PAGE 134

116 fxmax=fxmid; } else { // Root is in the right half interval xmin=xmid; fxmin=fxmid; } } xsatL=xmid; return xsatL; } //////////////// //////////////// //////////////// //////////////// //////////////// Saturated Vapor Concentration This routine is the companion to the prev ious one, determining the vapor saturation pressure given a pressure and temperature. It is also used to determ ine if conditions have entered the two phase region, e.g. if th e exhaust from the expander has begun to condense. //////////////// //////////////// //////////////// //////////////// //////////////// double PTSatGas(double P, double T) { double xmin=0; // Search interval double xmax=1; double xmid; double fxmid=1; SatTemp(P,xmax); double fxmax=Tdew-T; SatTemp(P,xmin); double fxmin=Tdew-T; int i=0; while (sqrt(fxmid*fxmid) > .00001){ xmid=(xmin+xmax)/2; SatTemp(P,xmid); fxmid=Tdew-T; ++i;

PAGE 135

117 if (i > 100){ xsatG=0; return xsatG; } if ((sqrt(pow((xmax-xmin),2)))<0.0001){ xsatG=xmax; return xsatG; } if ((fxmin*fxmid) <= 0){ // Root is in the left half interval xmax=xmid; fxmax=fxmid; } else { // Root is in the right half interval xmin=xmid; fxmin=fxmid; } } xsatG=xmid; return xsatG; } //////////////// //////////////// //////////////// //////////////// //////////////// Two-Phase Mixture Enthalpy For those conditions that are in the two phase region, this ro utine computes the mixture enthalpy given a pressure temperature, and overall con centration. The quality of the fluid is also returned. //////////////// //////////////// //////////////// //////////////// //////////////// double mixh(double P, double T, double x) { PTXh(P,T,x); SatTemp(P,x); if (T<=Tbub){ hmix=hLm; xq=0; return hmix, xq; } if (T>=Tdew){

PAGE 136

118 hmix=hgm; xq=1; return hmix, xq; } else { PTXsatx(P,T,x); hmix=xq*hgm+(1-xq)*hLm; } return hmix,xq,hgm,hLm; } //////////////// //////////////// //////////////// //////////////// //////////////// Temperature Determination Using Enthalpy This subroutine determines the fluid temp erature from the input parameters of pressure, concentration, and enthalpy. This is used in situations where the enthalpy can be determined from an energy balance a nd the corresponding temp erature is needed. //////////////// //////////////// //////////////// //////////////// //////////////// double PXHmixT(double P, double x, double h) { double Tmin=100; // Search interval, T's in K double Tmax=800; double Tmid; double fTmid=1; mixh(P,Tmax,x); double fTmax=h-hmix; mixh(P,Tmin,x); double fTmin=h-hmix; int i=0; while (sqrt(fTmid*fTmid) > .0001){ Tmid=(Tmin+Tmax)/2; mixh(P,Tmid,x); fTmid=h-hmix; //cout << "Tmid << Tmid << "Tmax << Tmax << "Tmin << Tmin < 100){

PAGE 137

119 Tmix=0; //cout << "i>100 <
PAGE 138

120 PTXs(P,Tmax,x); double fTmax=s-sgm; PTXs(P,Tmin,x); double fTmin=s-sgm; int i=0; while (sqrt(fTmid*fTmid) > .00001){ Tmid=(Tmin+Tmax)/2; PTXs(P,Tmid,x); fTmid=s-sgm; ++i; if (i > 100){ TisenV=0; return TisenV; } if ((sqrt(pow((Tmax-Tmin),2)))<0.0001){ TisenV=Tmax; return TisenV; } if ((fTmin*fTmid) <= 0){ // Root is in the left half interval Tmax=Tmid; fTmax=fTmid; } else { // Root is in the right half interval Tmin=Tmid; fTmin=fTmid; } } TisenV=Tmid; return TisenV; } //////////////// //////////////// //////////////// //////////////// //////////////// Overall Cycle Calculation By employing many of the subroutines just presented, this routine computes the state point properties and ot her parameters given the boiling temperature, boiling pressure, basic solution concentr ation, absorption temperature, rectifier exit temperature,

PAGE 139

121 and component efficiencies. Returned para meters are any of the needed state point properties or manipulations thereof. //////////////// //////////////// //////////////// //////////////// //////////////// double CycCalc(double TB, double xs, double PB) { TXSatLiq(TA,xs); // Determine absorber pressure PA=PsatL; PTSatGas(PB,TB); // Determine vapor concentration xv=xsatG; if (xv<=xs || xv>=1){ null(); cout << "xv out of range" < PB" <=xs){ null(); cout << "xwb out of range" <
PAGE 140

122 h1=hLm; wp=ms*((h2s-h1)/Peff); h2=(wp/ms)+h1; PXHtempL(PB,xs,h2); T2=ThL; // Determine vapor and weak stream flowrate from boiler msd=ms; mv=((xs-xwb)/(xv-xwb))*msd; if (mv<=0){ null(); cout << "mv <= 0" <=Trect){ T7=Tdew; T11=Tdew; xvr=0.999; } else{ T7=Trect; T11=Trect; PTSatGas(PB,T11); xvr=xsatG; } TS=T11; PTXh(PB,T11,xvr); h11=hgm; PTSatLiq(PB,T7); xwr=xsatL; PTXh(PB,T7,xwr); h7=hLm;

PAGE 141

123 if (xvrxv){ null(); cout << "xwr out of range" << xwr << xv <qweak){ qmax=qweak; //cout << "qweak << qweak <
PAGE 142

124 } h3=(qmax*RecovEff)/msd+h2; h9=h8-(qmax*RecovEff)/mw; h10=h9; // Determine boiler heat input qB=mwb*h5+mv*h6-(msd*h3); // Determine superheater heat input PTXh(PB,TS,xvr); h12=hgm; qS=mvr*(h12-h11); // Determine turbine work out PTXs(PB,TS,xvr); s12=sgm; PTXv(PB,TS,xvr); v12=vgm; vin=v12; Q12=v12*mvr; PXStempV(PA,xvr,s12); SatTemp(PA,xvr); if (TisenV>=Tdew){ PTXh(PA,TisenV,xvr); h13s=hgm; T13s=TisenV; } else { PSXmixT(PA,Tbub,Tdew,xvr,s12); T13s=TisenM; PTXh(PA,T13s,xvr); h13s=xq*hgm+(1-xq)*hLm; } wt=Teff*mvr*(h12-h13s); h13=h12-wt/mvr; deltah=h12-h13s; if(wp>=wt){ null(); cout << "wp > wt" <
PAGE 143

125 return 0; } PXHmixT(PA,xvr,h13); T13=Tmix; xturb=xq; PTXv(PA,T13,xvr); v13=xq*vgm+(1-xq)*vLm; Q13=v13*mvr; // Determine refrigeration heat transfer mixh(PA,15+273.15,xvr); h14=hmix; qR=mvr*(h14-h13); // Determine absorber heat out qA=(mw*h10+mvr*h14)-ms*h1; // Determine efficiency CycEff=((wt-wp)/(qB+qS))*100; // Determine unrectified vapor work output and effective COP PTXs(PB,TB,xv); s12nr=sgm; PXStempV(PA,xv,s12nr); SatTemp(PA,xv); if (TisenV>=Tdew){ PTXh(PA,TisenV,xv); h13snr=hgm; T13snr=TisenV; } else { PSXmixT(PA,Tbub,Tdew,xv,s12nr); T13snr=TisenM; PTXh(PA,T13snr,xv); h13snr=xq*hgm+(1-xq)*hLm; } wtnr=Teff*mv*(h6-h13snr); if (wtnr
PAGE 144

126 null(); cout << "wtnr too small" <
PAGE 145

127 APPENDIX C EXPERIMENTAL DETAILS Instrument Settings The experimental procedures have been described in Chapter 5, this section provides the details needed to operate ancillary equipment. Data Acquisition System All of the data, aside from the manually recorded measurements, was collected with a DaqBook 200 acquisition system. For hardware information refer to the equipment list later in this Appendix, and for hardware configuration please refer to the operating manual for the DaqBook system. The acquisition software used is called DaqView. In this software each transducer is assigned a channel which can have its own calibration factors, i.e. for a linear fit, a slope and inte rcept. The factors used for this purpose were obtained by calibrating the transducers; these values are provided in the subsequent sections. As for data collection, a sampli ng rate of 25 scans per second for one second was used. The data was converted from bi nary-coded to ASCII format and saved as a text file. Gas Chromatograph A gas chromatograph (GC) was used to an alyze basic solution and weak solution samples during experiments. The startup and settings for the GC are as follows. 1. Verify that the correct column for a mmonia-water separation is installed. 2. The helium carrier gas supply tank is opened. All other valves leading to the GC are also opened. The regulator should be adjusted to approximately 60 psig. Gas flow past the thermal conductivity detector (TCD) can be verified by immersing the

PAGE 146

128 exhaust tube from the TCD oven, which is located inside the main oven, into a small vial of water and observing bubbles leaving the tube. 3. Power is turned on to the data acquisiti on PC and the PC-GC interface. With the computer on, the GC software PeakSimpl e is started and the appropriate control file is loaded, cycle.con. 4. The GC itself can now be turned on and th e settings verified. The on-board flow regulator should be set at 400. The TCD attenuator switch should be set to 1. Using the GCs digital readout the followi ng temperatures and pressures should be verified. 5. Carrier 1 pressure should be 50 psig 6. Head pressure 1 should be approximate ly 6-10 psig, pressures significantly different from this may require re placement of the injection septum. 7. Oven temperature set point of 100 C. 8. TCD set point of 125 C. 9. The filter bake switch on the GC is switched on. 10. The TCD current switch is sw itched to the high setting. The GC should idle for approximately one hour or until the TCD output stabilizes. Samples typically cleared the detector in unde r 8 minutes with these conditions, so this was chosen as the length of the data acquisition cycle. Shutdo wn of the GC is the reverse of startup. Uncertainty of Direct Measurements This section gives details of the measurem ent uncertainties and the computation of the error bounds for the experimental data. A description of the experimental system has been given in Chapter 5 and the equipmen t specification is provided later in this appendix. For the results of concern for this work, there are five types of fundamental measurements from which all other quanti ties are derived These are temperature, pressure, concentration, volume flow rate, and expander shaft speed. Other

PAGE 147

129 measurements were made when air-testing the expander, however, those measurements are treated in separately in Appendix D. Temperature All temperatures were measured with t ype T thermocouples interfaced with a DBK-82 thermocouple analog to digital card and a DaqBook 200 PC interface. A twopoint linear calibration (i ce point and boiling point) was performed with all thermocouples prior to installation. Tabl e C-1 presents the co mputed slopes and intercepts. Table C-1. Calibration factors for the thermocouples used in this work. ProbeLocationSlopeIntercept LetterDescription[-][C] EAbsorber Vapor In0.997008973-0.398803589 FAbsorber Weak In1.005025126-0.603015075 DAbsorber Strong Out1.006036217-0.704225352 CAbsorber Liquid Pool1.004016064-0.803212851 KBoiler Strong In1.004016064-0.702811245 LSeparator Strong In1.005025126-0.703517588 BAbsorber Coolant In1.005025126-0.804020101 AAbsorber Coolant Out1.004016064-0.702811245 MBoiler Heat Source In1.006036217-0.704225352 QBoiler Heat Source Out1.004016064-0.602409639 JRectifier Coolant In0.998003992-0.499001996 IRectifier Coolant Out1-0.6 RSeparator Weak Out1.005025126-0.703517588 SSeparator Vapor Out1.001001001-0.500500501 TExpander Inlet1.005025126-0.703517588 UExpander Exhaust1.003009027-0.702106319 WRectifier Vapor Out1-0.5 The accuracy of the thermocouple readings was improved by using the above calibration constants, however, the repeatability of the enti re measuring system was the largest contributor to measurement uncertain ty. Variations in measurement channels, ambient conditions, etc., contributed to a tota l uncertainty of 0.4 C in all thermocouple measurements.

PAGE 148

130 Pressure Pressure was measured with strain gage type transducers, the specific models are listed in the equipment specification secti on. Similar to the thermocouples, each transducer was calibrated before installation. An Ametek pr essure tester (model HL-24, serial HL-5886) was used to apply 5 differe nt calibration pressure s to each transducer across its expected range of operation. Calibration constants, Table C-2, for each transducer were determined from a linea r regression of the five data points. Table C-2. Pressure transducer calibration factors. Transducer Slope [psi/V] Intercept [psi] Absorber Exit, 2 18.79137 -33.6551 Vapor Throttle Exit, 3 18.74273 -33.8081 Absorber Vapor Inlet, 4 18.37891 -32.6564 Absorber Weak Inlet, 5 28.97704 -44.5912 Separator Entrance, 6 91.95402 -219.46 Turbine Inlet, 7 63.05169 -63.4615 Turbine Exhaust, 8 18.75117 -33.8843 Individual uncertainties in pressure measur ements were determined from the stated accuracy reported by the manufacturer for each transducer, Table C-3. Table C-3. Stated uncertainties for pressure transducers. Transducer Uncertainty Absorber Exit, 2 0.078 psi, (0.000538 MPa) Vapor Throttle Exit, 3 0.078 psi, (0.000538 MPa) Absorber Vapor Inlet, 4 0.24 psi, (0.00165 MPa) Absorber Weak Inlet, 5 0.13 psi, (0.000896 MPa) Separator Entrance, 6 0.325 psi, (0.00224 MPa) Turbine Inlet, 7 0.325 psi, (0.00224 MPa) Turbine Exhaust, 8 0.078 psi (0.000538 MPa) Volume Flow Rate Three volume flow rates were recorded and used to determine mass flows in this work, they are the basic solution flow, weak solution flow, and the turbine vapor flow. The first two were measured with variable area rotameters while for the latter a turbinetype flow meter was used.

PAGE 149

131 The basic and weak solution rotameters ha ve been calibrated fo r the volume flow of liquids with specific grav ities of 0.833 and 0.88 respectivel y. Equation C-1 is used to correct the indicated reading to the actual volume flow. floatactcal actind floatcalactQQ (C-1) Where the float density is the same for both meters, 8016 kg/m3, which corresponds to 316 stainless steel. The accuracy is stated as 1 % of full scale for the range 0.1 to 1.11 gpm for the basic solution meter and 1 % of full scale for the range 0.08 to 0.8 gpm for the weak solution meter. The turbine-type flow meter used for vapor flow measurements directly measures volume flow rate, therefore, no correction for calibrated conditions is needed. Accuracy is given as 1 % of reading. The output of the meter is a 4-20 mA signal proportional to the flow rate. The constants used in the data acquisition software were a slope of 6.1248569489 and an intercept of -6.250. Concentration Direct ammonia concentration measuremen ts were taken at two points in the system, the basic solution flow and the weak solution flow. A gas chromatograph (GC) with a thermal conductivity detector (TCD) calibrated for detecting ammonia-water was used to make these direct measurements. Bo th of these streams are liquids. This method proved unsuitable for saturated vapor concen tration measurements, therefore, vapor concentration is derived from pressure and temperature measurements, details in a later section.

PAGE 150

132 The GC was calibrated with a concentrati on standard to determine the calibration constants and measurement uncertainty. Repeat ed measurements were taken of a sample containing an ammonia mass frac tion of 28.76 %, per the certificate of analysis provided by the chemical supply company. The output of the GC is a plot of TCD voltage versus time; distinct peaks in voltage are recorded for each cons tituent component, ammonia and water. The quantitative output is the ratio of the area under the peak corresponding to ammonia to the entire area unde r all peaks and is related to the ammonia mass fraction through equation C-2. 3%100 11NHCx A Cx (C-2 The calibration constant, C, was comput ed as the average of the constants computed for each measurement of the standard solution using equation C-3. 3 3 2 31 0% 1 100%c cNH NH xx xc c HO NH x xxA A x C x A A (C-3) The average calibration constant comput ed from 15 measurements was 0.9794. The uncertainty in concentration measuremen ts using a 95 % confidence interval is 0.015. Shaft Speed The optical speed sensor has been descri bed previously. Measurement uncertainty based on the deviation of repeated measur ements is 94 rpm with a 95 % confidence interval. Due to the sensors design, it regi stered two pulses per revolution, therefore, a scale factor of 30 was used in the data ac quisition software to obtain readings in revolutions per minute.

PAGE 151

133 Uncertainty of Derived Measurements To determine the accumulated error in de rived measurements, those quantities that are functions of two or more primary measur ements, the square root of the sum of the squares of the individual contributing errors is used. A derived quantity y can be written in terms of the primary measurements x in general form as equation C-4. 12,,,nyfxxx (C-4) Assuming independent and symmetrical e rrors in the primary measurements, dx, the error in the derived parameter, dy, can be estimated from equation C-5. 2 22 12 12 n nyyy dydxdxdx xxx (C-5) Many of the derived quantities depend on th e fundamental measurements used with thermodynamic property data of ammonia-wate r. Since these relationships cannot be directly differentiated as indicated in equati on C-5 a finite difference numerical scheme is used to approximate them. The formulati on for the central-difference approximation, neglecting higher order terms, is presented as equation C-6. 1212,,,(),,,() 2nnnn nn f xxxdxfxxxdx y xdx (C-6) The following paragraphs describe the deri ved measurements used in this work, they present the dependence on direct measur ements and give the uncertainty values used. Vapor Concentration Syringe sampling and GC analysis proved unsuitable for vapor flow measurements because the vapor is saturated at the sampli ng location. Despite the precaution of heating

PAGE 152

134 the syringe, condensate inevitably formed and re mained in the needle after injection into the GC. On the other hand, there is always at least one location where the vapor is confidently considered saturated and separa ted from liquid, the separator exit and the rectifier exit when the rectifier is in operat ion. Therefore, it was decided to use the proximate temperature and pressure measur ements and an assumption of saturated conditions to determine concentration from thermodynamic property data. Vapor concentration = f (pressure, temperature) (C-7) Mass Flow Rates The strong and weak flow rates are determined from the direct measurements of the volume flow rate, temperature and concentr ation. The thermodynamic properties are used to compute a density which converts th e volume flow measurem ents to mass flow readings. For error analysis the functional dependence of the basic and weak mass flow rates is the following. Basic/Weak mass flow = f (volume flow, concentration, temperature) (C-8) Note that the pressure dependence on mass fl ow is neglected for these liquid flows. As for vapor mass flow determination, th e functional dependen ce of measurements directly related to the measured volum e flow is similar to the ones above. Vapor mass flow = f (vapor volume flow, pressure, temperature, concentration) (C-9) Note, however, that the errors due to uncerta inty in the pressure reading need to be accounted for in this case. Also, this flow has been generically termed the vapor mass flow, which is the flow measured by the vapor flowmeter. This is distinguished from the flow entering the rectifier (when in operation), which is not directly measured. Rather, it is derived from a mass balance of the rectifier, since both vapor inlet and outlet states are

PAGE 153

135 assumed saturated and the concentration is known. The functional dependence for this stream is then given as the following. Rectifier inlet vapor mass flow = f (vapor volume flow, pressu re, temperature, inlet and outlet concentrations) (C-10) Power Output For expander operation in the power-cooli ng setup, power output was estimated from the thermodynamic states at turbine inle t and exhaust. Mechanical measurements were used for some air testing, however, that analysis is treated separately in Appendix D. Power output is then dependent on the following parameters, of which the dependence on direct measurements has been discussed. Power output = f (vapor mass flow, vapor concen tration, inlet and exhaust temperatures and pressures) (C-11) Expander Efficiency Based on the just-described, derived measurement, the expander efficiency is determined from two power output meas urements as shown in the following. Expander efficiency = f (Computed power output, Idea l computed power based on measured conditions) (C-12) The resulting uncertainties along with a su mmary of the functional dependencies are presented in Table C-4.

PAGE 154

136 Table C-4. Derived measurement uncertainty summary. Parameter Error Dependence Uncertainty Rectifier Exit Vapor Conc. P, T 0.00031 kg/kg Separator Exit Vapor Conc. P, T 0.00067 kg/kg Basic Solution Mass Flow Q, T, X 0.00062 kg/s Vapor Mass Flow Q, T, P 0.00003 kg/s Rectifier Inlet Vapor Mass Flow Q, T (2), P (2) 0.000048 kg/s Power Output Q, T (2), P (2) 1.6 W Expander Efficiency Q, T, P (2) 5.4 % Equipment Specification This section details the measurement equi pment and expander used for this work. For details regarding the physical makeup of the rest of the experimental setup, which is not covered in detail here, the reader is referr ed to the description given in Chapter 5 or reference [70]. Instrumentation The pertinent descriptive information about the measurement equipment used is detailed in Table C-5. Table C-5. Detailed descriptions of the in strumentation and measurement equipment used for this work. Device Manufacturer Part Number/Model Vendor (if known) Data Acquisition PC Interface IoTech DaqBook 200 Ser. 219024 IoTech (440) 439-4091 Frequency Measurement Card IoTech DBK 7 IoTech (440) 439-4091 Current/Voltage Measurement Card IoTech DBK 15 IoTech (440) 439-4091 Thermocouple Measurement Card IoTech DBK 82 IoTech (440) 439-4091 All Temperatures T-type Thermocouples Omega TMQSS-125U-6 Omega (800) 826-6342 Pressure Transducers Absorber Exit, 2 Cole Parmer Model 68073-04 Ser. 1490231 Cole Parmer (800) 323-4340

PAGE 155

137 Table C-5. Continued. Device Manufacturer Part Number/Model Vendor (if known) Pressure Transducers Vapor Throttle Exit, 3 Cole Parmer Model 68073-04 Ser. 1490230 Cole Parmer (800) 323-4340 Absorber Vapor Inlet, 4 Cole Parmer Model 07356-13 Ser. 00205 Cole Parmer (800) 323-4340 Absorber Weak Inlet, 5 Cole Parmer Model 68073-06 Ser. 1853374 Cole Parmer (800) 323-4340 Separator Entrance, 6 Setra Part # SE256103 Model 256 Ser. 0903 1945647 Davis Inotek Instruments (800) 368-2516 Turbine Inlet, 7 Setra Part # SE256103 Model 256 Ser. 0903 1945648 Davis Inotek Instruments (800) 368-2516 Turbine Exhaust, 8 Cole Parmer Model A-68073-04 Ser. 1945762 Cole Parmer (800) 323-4340 Flowmeters Vapor, Turbine-type Hoffer Flow Controls, Inc. Model HO3/4X3/4-20-CB1MC3PAX8S-NPT Ser. 99628 Quinn Associates (813) 254-5211 Basic Solution, Rotameter Brooks Model 1110DJ33C4DAA Serial 0199120123644/001 Weak Solution, Rotameter Brooks Model 110CJ32CMDAA Serial 9812HC028152/1 Gas Chromatograph SRI Model 8610-50 Serial 1027 Column HAYSEP T Expander Details The expander used for these experiments wa s a converted turbine originally used in an aircraft air-cooling system. The units were manufactured by Airesearch manufacturing and have a factory part number of 56690. Components from three

PAGE 156

138 turbines were used in this work and they were obtained from C and H sales (800) 3259465, under the part number DCB2410. Some modifications were needed to adapt this device for service in the experimental setup. Most notably this i nvolved the manufacture of a new rear housing since the original had vents to the atmosphere. The drawings that follow give an overview of the parts manufactured (by the au thor) to form a sealed rear housing and adapt the turbine for service with ammonia-water working fluid. First is figure C-1 which shows the complete assembly. Figure C-1. View of the assembled rear housing. Figure C-2 is an exploded view of the asse mbly with the individual parts labeled. The spindle and the rotor were used as-rem oved from the original turbine. The back plate, housing, and front plate were all manufactured for this work.

PAGE 157

139 Figure C-2. Exploded view of the rear housing assembly. Back Plate Housing Spindle Front Plate Rotor

PAGE 158

140 APPENDIX D EXPANDER AIR TESTING It was mentioned in the earlier chapters of this work that the turbine used for the experiments was also tested with air to re solve some of the performance measurement discrepancies. Indeed, Figure 67 contains data from these tests and is used to eliminate some of the questionable results. This a ppendix provides the expe rimental details for those measurements based on air tests. Experimental Setup Figure D-1 is a schematic of the setup us ed for air testing. Compressed air was supplied by a standard shop air compressor at approximately 90 psig. An adjustable pressure regulator was used to control the flow and expander inlet pressure. On the inlet line two parallel rotameters were used to measure volume flow rate (two were used because the flow exceeded the capacity of a single meter), and a pressure transducer monitored inlet pressure. The same th ermocouples used in the ammonia-water experiments were used here to record in let and exhaust temperatures. The expander speed sensor was described in Chapter 5 a nd was used here. Also, the same data acquisition system used for the ammonia-water experiments was used for these tests. Note in Figure D-1 that two options were used to absorb the mechanical power produced by the expander. One is the ma gnetic brake device that was described in Chapter 5 and used for the ammonia-water e xperiments. Figure D-2 is a photograph of the rear of the expander showi ng the aluminum disk that wa s used in the magnetic brake arrangement.

PAGE 159

141 T T Expander From Air Compressor F Pressure Regulator P Exhaust to Atmosphere Magnetic Brake Gearbox Generator Torque Measurement Speed Meas. Figure D-1. Setup schematic used for the air testing. Figure D-2. Rear view of expa nder with cover removed. Th e aluminum disk used with the magnetic brake is visible.

PAGE 160

142 Figure D-3. Photograph of ge nerator loading arrangement. The other loading configurat ion is one that allowed a direct measurement of the power produced. It consisted of a DC permanent-magnet generator driven by the expander through a reducing planetary gearbox (6:1), Figure D-3. The generator was mounted in such a way that it was constraine d from rotating by a load cell link. This provided a measurement of the torque supplied to the generator and along with the shaft speed measurement was used to compute a torque-based power measurement (power = Load Cell Generator in reaction tor q ue frame Turbine

PAGE 161

143 torque angular momentum). The torque measurements were used to confirm the power computed from the thermodynamic measur ements. A photograph of the generator mounting arrangement is shown in Figure D3. Figure D-4 shows the gearbox mounted inside of the expander housing. Figure D-4. Photograph of gear box mounted on expander spindle. Test Results Tests were performed to verify the thermodynamic method of computing power and to provide a measure of the expanderÂ’s no-l oad power consumption. Initially, testing was performed in the same manner as with a mmonia-water. The pr essure regulator was Gearbox Spindle Housing

PAGE 162

144 used to maintain a desired inlet pressure, then beginning with th e no-load condition the turbine was progressively loaded in incremental steps. Measurements were made at each loading step. The results were used to compute the expander power via the two methods, thermal and torque-based. Results from a test with a nominal inlet pressure of 40 psig are shown in Figure D-5. 0 10 20 30 40 50 60 70 80 90 20000250003000035000400004500050000Shaft Speed [rpm]Turbine Power [W] Thermal Meas. Torque Meas. Figure D-5. Air testing results comparing the value of power that was computed by the thermal-based and torque-based measurements. As shown in Figure D-5 there is a fair ly constant difference between the two methods at lower shaft speeds (approximately 24 krpm to 32 krpm), while at higher speeds the difference grows dramatically with the thermal measurement indicating higher outputs. This difference has a plausible explanation which is, at higher speeds more of the power removed from the working fluid is cons umed within the turb ine and does not show

PAGE 163

145 up as torque on the generator. The internal loads of the turbine consist largely of the friction in the bearings and the fluid fricti on caused by the drag on the rotating parts of the expander. To test this theory, tests were perf ormed to measure the internal power consumption of the expander. This was accomplished by removing all external loading from the expander and measuring conditions at the no-load speed. Different speeds were obtained by adjusting the pressure regulator, which controls th e inlet pressure. For this case, the power consumed was computed based on the thermodynamic measurements. The no-load power results are plotted in Fi gure D-6 along with the difference between the power measurements of Figure D-5. 0 10 20 30 40 50 60 20000250003000035000400004500050000Shaft Speed [rpm]Turbine Power [W] Thermal-Torque Difference No Load Power Consumption Figure D-6. Comparison of the difference be tween the power measur ements of Figure D-5 and the no-load power measurements, which is essentially a measure of the internal power consumption.

PAGE 164

146 As Figure D-6 shows, the internal power consumption appears to explain much of the difference between the two methods of power measurement. This provides confidence in a couple of concepts. First, the agreement between th e thermal and torquebased power measurements indi cates that the thermal based power measurement can be a valid technique. Secondly, the measured no-load power consumption provides an estimate of the expanderÂ’s internal power c onsumption that can be compared to the ammonia-water test results. Granted, this co mparison is not entirely accurate since the windage loss of the expander is a function of the working fluidÂ’s kinematic viscosity, [50]. Nonetheless, this comparison was made in Chapter 6 and it was used to eliminate some questionable measurements of power production. Ideally all of the ammonia-water experi ments would have used the gearbox and generator and computed power fr om a torque-based perspectiv e, unfortunately there were many practical issues that prevented it from happ ening. The first is that unlike the testing with air, the supply of ammonia-water work ing fluid could not be switched on and off several times to tare the load cell while maintaining steady state operating conditions for the rest of the cycle. Secondly, because the generator was mounted outside of the expander housing a shaft seal was required and due to gearbox-generator alignment inaccuracy ammonia leakage became excessive. Also, this same misalignment eventually led to the failure of the gearbox. Experimental Details This section presents the supporting inform ation regarding the air-testing of the expander.

PAGE 165

147 Measurement Uncertainties The treatment of measurem ent uncertainty is dependent on the method used to compute the expander power. For the thermal-based power measurement the computation is dependent on the basic measurem ents of temperature (inlet and exhaust), inlet pressure, and volume flow rate. While the torque-based meas urements relied upon basic measurements of force and shaft speed. Table D-1 provides the details of the basic measurement uncertainty and the resulting uncertainty in power output. Table D-1. Thermal and torque-b ased measurement uncertainties. Parameter Uncertainty Thermal Measurements Flow Left [SLPM] 1.25 SLPM Flow Right [SLPM] 1.25 SLPM Inlet Temperature [C] 0.8 C Exhaust Temperature [C] 0.8 C Inlet Pressure [psig] 0.097 psig Power Determination [W] 4.3 W Torque Measurements Force [N] 0.2 N Frequency [Hz] 5 Hz Power Determination [W] 2.6 W Note that some of the uncertainties in Table D-1 are higher than the stated uncertainties used for equivalent measuremen ts during the ammoniawater testing, e.g. temperature. This is due to the fact th at the same calibration procedures were not performed for the air testing.

PAGE 166

148 Equipment Specification Table D-2 details the significant components used for this testing. Some components have been detailed in Appendix C and are not repeated here, e.g. the data acquisition system. Table D-2. Summary of the equipment a nd components used for the air tests. Device Manufacturer Part Number/Model Vendor (if known) Instrumentation Temperature T-type Thermocouples Omega TMQSS-125U-6 Omega (800) 826-6342 Pressure Cole Parmer Model A-68073-04 Ser. 1945762 Cole Parmer (800) 323-4340 Flow Rotameters Gilmont Model No. 15 Force Load Cell Interface Model SM-10 Ser. D02658 Part No. IF12123 Davis Inotek Instruments (800) 368-2516 Force Amplifier and Readout Cooper Instruments Model DFI 2555 Serial 952327 Misc. Components Generator (originally a PM, DC motor) Graupner Speed 700 BB Planetary Gearbox Model Motors VMGM 6.00:1

PAGE 167

149 REFERENCES 1. Goswami, D. Y., 1995, “Solar Thermal Power: Status of Technologies and Opportunities for Research,” Heat and Mass Transfer ’95, Proceedings of the 2nd ASME-ISHMT Heat and Mass Transfer Conference, Tata-McGraw Hill Publishers, New Delhi, India, pp. 57-60. 2. Lu, S., and D. Y. Goswami, 2003, “Optimization of a Novel Combined Power/Refrigeration Thermodynamic Cycle,” Journal of Solar Energy Engineering, 125, pp. 212-217. 3. Tamm, G., D. Y. Goswami, S. Lu, and A. A. Hasan, 2003, “Novel Combined Power and Cooling Thermodynamic Cycle for Low Temperature Heat Sources, Part I: Theoretical Investigation,” Journal of Solar Energy Engineering, 125, pp. 218-222. 4. Tamm, G., and D. Y. Goswami, 2003, “Novel Combined Power and Cooling Thermodynamic Cycle for Low Temperature Heat Sources, Part II: Experimental Investigation,” Journal of Solar Energy Engineering, 125, pp. 223-229. 5. Vijayaraghavan, S., 2003, “Thermodynamic Studies on Alternate Binary Working Fluid Combinations and Configurations for a Combined Power and Cooling Cycle,” Ph.D. dissertation, University of Florida, Gainesville, FL. 6. Dunn, S., 2000, “Micropower: The Next Electrical Era,” Worldwatch Paper, 151, July, 94p. 7. Thermally Activated Technologies: Technology Roadmap—Developing New Ways to Use Thermal Energy to Meet the Needs of Homes, Offices, Factories, and Communities, 2003, Office of Energy Effici ency and Renewable Energy, U. S. Dept. of Energy, 46p. Available at: http://www.eere.energy.gov/ de/pdfs/tat_roadmap.pdf last accessed October 23, 2004. 8. Butti, K., and J. Perlin, 1980, A Golden Thread, Van Nostrand Reinhold Co., New York. 9. Wahl, E. F., 1977, Geothermal Energy Utilization, John Wiley & Sons, New York. 10. Patel, P. S., and E. F. Doyle, 1976, “C ompounding the Truck Diesel Engine with an Organic Rankine-Cycle System,” SAE publication 760343, 12p.

PAGE 168

150 11. Lindsley, E. F., 1970, “New: Minto’s Un ique Steamless ‘Steam’ Car,” Popular Science, Oct., pp. 51-53. 12. Curran, H. M., 1988, “Mechanical Systems and Components,” Active Solar Systems, Vol. 6, G. Lf ed. MIT Press, Cambridge, MA. 13. Prigmore, D., and R. E. Barber, 1975, “Cooling with the Sun’s Heat: Design Considerations and Test Data for a Rankine Cycle Prototype,” Solar Energy, 17, pp. 185-192. 14. Barber, R. E., 1978, “Current Costs of Solar Powered Organic Rankine Cycle Engines,” Solar Energy, 20, pp. 1-6. 15. Yamamoto, T., T. Furuhata, N. Arai, a nd K. Mori, 2001, “Design and Testing of the Organic Rankine Cycle,” Energy, 26, pp. 239-251. 16. Angelino, M., M. Gaia, and E. Macchi, 1984, “A Review of Italian Activity in the Field of Organic Rankine Cycles ,” VDI Berichte 539, pp. 465-482. 17. Nguyen, T., P. Johnson, A. Akbarzadeh, K. Gibson, and M. Mochizuki, 1995, “Design, Manufacture and Testing of a Closed Cycle Thermosyphon Rankine Engine,” Heat Recovery Systems & CHP, 15 (4), pp. 333-346. 18. Kane, M., D. Larrain, D. Favrat, and Y. Allani, 2003, “Small Hybrid Solar Power System,” Energy, 28, pp. 1427-1443. 19. Larjola, J, 1995, “Electricity From Industrial Waste Heat Using High-Speed Organic Rankine Cycle (ORC),” International Journal of Production Economics, 41, pp. 227-235. 20. Smith, T. C. B., 2003, “Low Cost Organic Rankine Cycles for Grid Connected Power Generation,” Proceedings of the ISES Solar World Congress, Gteborg, Sweden, International Solar Energy Society, 8 p. 21. Wells, D. N., 2000. “Scroll Expansion Machines for Solar Power and Cooling Systems,” Proceedings of Solar 2000, Madison, WI, American Society of Mechanical Engineers, 7p. 22. Maloney, J. D., and R. C. Roberts on, 1953, “Thermodynamic Study of AmmoniaWater Heat Power Cycles,” ORNL Report CF-53-8-43, Oak Ridge, TN. 23. Kalina, A. I., 1984, “Combined Cycle System with Novel Bottoming Cycle,” ASME Journal of Engineering for Gas Turbines and Power, 106, pp. 737-742. 24. Marston, C. H., 1990, “Parametric Analysis of the Kalina Cycle,” Journal of Engineering for Gas Turbines and Power, 112, pp. 107-116.

PAGE 169

151 25. Ibrahim, O. M., and S. A. Klein, 1996, “Absorption Power Cycles,” Energy, 21, (1), pp. 21-27. 26. Xu, F., and D. Y. Goswami, 1999, “Thermodynamic Properties of Ammonia-Water Mixtures for Power-Cycle Applications,” Energy, 24, pp. 525-536. 27. Goswami, D. Y., and F. Xu, 1999, “Analysis of a New Thermodynamic Cycle for Combined Power and Cooling Using Low and Mid Temperature Solar Collectors,” Journal of Solar Energy Engineering, 121, pp. 91-97. 28. Xu, F., D. Y. Goswami, and S. S. Bh agwat, 2000, “A Combined Power/Cooling Cycle,” Energy, 25, pp. 233-246. 29. Lu, S., and D. Y. Goswami, 2002, “The oretical Analysis of Ammonia-Based Combined Power/Refrigeration Cycle at Low Refrigeration Temperatures,” Reno, NV, Proceedings of SOLAR 2002, American Society of Mechanical Engineers, 9p. 30. Vijayaraghavan, S., and D. Y. Goswami, 2002, “On Evaluating Efficiency of a Combined Power and Cooling Cycle,” New Orleans, LA, Proceedings of IMECE 2002, American Society of Mechanical Engineers, 9 p. 31. Hasan, A. A., D. Y. Goswami, and S. Vijayaraghavan, 2002, “First and Second Law Analysis of a New Power and Refr igeration Thermodynamic Cycle Using a Solar Heat Source,” Solar Energy, 73 (5), pp. 385-393. 32. Hasan, A. A., and D. Y. Goswami, 2003, “Exergy Analysis of a Combined Power and Refrigeration Thermodynamic Cycle Driven by a Solar Heat Source,” Journal of Solar Energy Engineering, 125, pp. 55-60. 33. Tamm, G., D. Y. Goswami, S. Lu, and A. A. Hasan, 2004, “Theoretical and Experimental Investigation of an A mmonia-Water Power and Refrigeration Thermodynamic Cycle,” Solar Energy, 76, pp. 217-228. 34. Tamm, G., D. Y. Goswami, S. Lu, and A. A. Hasan, 2003, “Novel Combined Power and Cooling Thermodynamic Cycle for Low Temperature Heat Sources, Part I: Theoretical Investigation,” Journal of Solar Energy Engineering, 125, pp. 218-222. 35. Badr, O., P. W. O’Callaghan, and S. D. Probert, 1984, “Performances of RankineCycle Engines as Functions of Their Expander’s Efficiencies,” Applied Energy, 18, pp. 15-27. 36. Tamm, G., D. Y. Goswami, S. Lu, and A. A. Hasan, 2001, “Theoretical and Experimental Investigation of an A mmonia-Water Power and Refrigeration Thermodynamic Cycle,” Adelaide, Australia, Proceedings of ISES 2001 Solar World Congress, International Solar En ergy Society, pp. 893-906.

PAGE 170

152 37. Vidal, A., R. Best, R. Rivero, and J. Ce rvantes, 2004, “Analysis of a Combined Power and Refrigeration Cycle by the Exergy Method,” Energy-Efficient, CostEffective, and EnvironmentallySustainable Systems and Processes (Proceedings of ECOS 2004), R. Rivero, L. Monroy, R. Puli do, and G. Tsatsaronis, eds., Instituto Mexicano del Petroleo, 3, pp. 1207-1218. 38. Oliveira, A. C., C. Afonso, J. Matos, S. Riffat, M. Nguyen, and P. Doherty, 2002, “A Combined Heat and Power System fo r Buildings Driven by Solar Energy and Gas,” Applied Thermal Engineering, 22, pp. 587-593. 39. Huang, B. J., J. M. Chang, V. A. Petre nko, and K. B. Zhuk, 1998, “A Solar Ejector Cooling System Using Refrigerant R141b,” Solar Energy, 64, (4-6), pp. 223-226. 40. Riffat, S. B., S. A. Omer, and P. S. D oherty, 2000, “Experimental Performance of an Ejector Refrigeration System Us ing Methanol as a Working Fluid,” International Journal of Ambient Energy, 21, (1), pp. 3-10. 41. Erickson, D. C., G. Anand, and I. K yung, 2004, “Heat-Activated Dual-Function Absorption Cycle,” ASHRAE Transactions, 110, (1), pp. 515-524. 42. Amano, Y., T. Suzuki, T. Hashizume, M. Akiba, Y. Tanzawa, and A. Usui, 2000, “A Hybrid Power Generation and Refr igeration Cycle with Ammonia-Water Mixture,” IJPGC2000-15058, Proceedings of 2000 Joint Power Generation Conf., ASME, 6 p. 43. Zhang, N., R. Cai, and N. Lior, 2004, “A Novel Ammonia-Water Cycle for Power and Refrigeration Cogeneration,” Paper IMECE2004-60692, Proceedings of IMECE 2004, ASME, 14 p. 44. Lozza, G. and E. Macchi. 1984. “Off-Design Performance and Adaptability to Energy Source Modifications of Wa ste Heat Recovery Engines,” Energy economics and management in industry: proceedings of the European Congress, Algarve, Portugal, 2, pp. 117-121. 45. Bliem, C. J., 1980, “Design and Off-Desi gn Operation of a Dual-Boiling Binary Geothermal Power Plant,” Heat Transfer—Orlando 1980, 76, (199), AIChE, pp. 163-172. 46. Herold, K. E., R. Radermacher, and S. A. Klein, 1996, Absorption Chillers and Heat Pumps, CRC Press, Boca Raton, FL. 47. ASHRAE, 1969, “Thermodynamic Properties of Refrigerants,” American Society of Heating, Refrigerating and Air-C onditioning Engineers, Inc., Atlanta. 48. Du Pont Chemicals, “Thermodynamic Pr operties of HCFC-123,” Document T123—ENG, Wilmington, DE.

PAGE 171

153 49. Barber, R., and D. Prigmore, 1981, Solar-Powered Heat Engines, Solar Energy Handbook, J. F. Kreider and F. Kreith eds. McGraw-Hill, New York. 50. Balje, O.E. 1981. Turbomachines A Guide to Design, Selection, and Theory. John Wiley & Sons, New York. 51. Macchi, E. and G. Lozza, 1985, Comparison of Partial vs. Full Admission for Small Turbines at Low Specific Speeds, Paper 85-GT-220, ASME Gas Turbine Conference and Exhibit, Houston, 8p. 52. Uehara, H. and Y. Ikegami, 1990, Optimization of a Closed-Cycle OTEC System, Journal of Solar Energy Engineering, 112, Nov., pp. 247-256. 53. Smith, R. W., J. Ranasinghe, D. Stats, and S. Dykas, 1996, Kalina Combined Cycle Performance and Operability, 1996 Joint Power Generation Conference, Vol. 2, ASME, pp. 701-727. 54. Leibowitz, H. M. and N. Zervos, 1992, Ins tallation and Early Te st Results of a 3 MW Kalina Cycle Demonstration Plant, Paper 929011, Proceedings of the 27th Intersociety Energy Conversion Engineering Conference, P-259, Vol. 3, SAE, pp. 3.35-3.42. 55. Mlcak, H. A., 2001, Design and Start-Up of the 2 MW Kalina Cycle Orkuveita Husavikur Geothermal Power Plant in Iceland, European Geothermal Energy Council, 2nd Business Seminar. 56. Amano, Y., K. Takeshita, T. Hashizuma, M. Akiba, A. Usui, and Y. Tanzawa, 2001, Experimental Results of an Amm onia-Water Mixture Turbine System, Proceedings of 2001 International Joint Power Generation Conference, ASME, pp. 69-75. 57. Maack, R. and P. Valdimarsson, Operating Experience with Kalina Power Plants, Available at: http://www.xorka.com/research_papers.htm last accessed October, 25, 2004. 58. Bloch, H. P., 1996, A Practical Guide to Steam Turbine Technology, McGraw-Hill, New York. 59. Guinn, G. R. and N. Turner, 1990, Design Manual for Small Steam Turbines, Tennessee Valley Authority, Muscle Shoals, AL. 60. Baily, F. G., 1987, Steam Turbines, Marks Standard Handbook for Mechanical Engineers 9th Edition, E. A. Avallone and T. Baumeister eds., McGraw-Hill, New York, pp. 9/39-9/60. 61. Badr, O., S. Naik, and P. W. OCalla ghan, 1991, Expansion Machine for a Low Power-Output Steam Rankine-Cycle Engine, Applied Energy, 39, pp. 93-116.

PAGE 172

154 62. Kim, H. J., I. Park, and P. C. Rha, 20 03, “Development of Scroll Expander for Power Generation,” IMechE Conference Transactions : International Conference on Compressors and Their Systems, pp. 359-368. 63. Oomori, H. and S. Ogino, 1993, “Waste Heat Recovery of Passenger Car Using a Combination of Rankine Bottoming Cycl e and Evaporative Engine Cooling System,” Paper 930880, SAE, pp. 159-164. 64. Badr, O., S. D. Probert, and P. W. O’Callaghan, 1986, “Optimal Design and Operating Conditions for a Multi-vane Expander,” Applied Energy, 24, pp. 1-27. 65. Larjola, J. 1986. “High-Speed Turb omachinery and Its A pplication in the Conversion of Energy,” ASME paper 86-WA/FE-4, 7p. 66. Rodgers, C., 1993, “Small (10-200 kW) Tur bogenerator Design Considerations,” ASME IGTI Vol. 8, pp. 525-542. 67. Batton, W. D., and R. E. Barber, 1981, “Rankine Engine Solar Power Generation, Part II – The Power Generation Module ,” ASME paper 81-WA/Sol-23, 8p. 68. Kremmer, M. and O. Okurounmu, 1965, “Condensation of Ammonia Vapor During Rapid Expansion,” Report No. 79, MIT Gas Turbine Laboratory, 32 p. 69. Kuwada, J. T., 1972, “Geothermal Power Plant Design,” Water, AIChE Symposium Series, pp. 439-444. 70. Tamm, G., 2003, “Experimental Investig ation of an Ammonia-Based Combined Power and Cooling Cycle,” Ph.D. dissertation, University of Florida, Gainesville, FL. 71. Ziegler, B. and C. Trepp, 1984, “Equation of State for Ammonia-Water Mixtures,” International Journal of Refrigeration, 7, (2), pp. 101-106. 72. Ibrahim, O. M. and S. A. Klein, 1993 “Thermodynamic Properties of AmmoniaWater Mixtures,” ASHRAE Transactions, 99, pp. 1495-1502. 73. El-Sayed, Y. M. and M. Tribus, 1985 “Thermodynamic Properties of WaterAmmonia Mixtures Theoretical Implementatio n for use in Power Cycles Analysis,” ASME Special Publication AES 1, New York, pp. 89-95.

PAGE 173

155 BIOGRAPHICAL SKETCH The author was born on November 22 in Ne w Orleans, Louisiana, approximately nine months after the 1974 Mardi Gras celeb ration. He completed high school in Mt. Holly, North Carolina, and then enrolled at the University of North Carolina at Charlotte. After receiving a bachelorÂ’s degree in mechanical engineer ing he moved to Gainesville, Florida and began graduate studi es at the University of Fl orida, where he received a masterÂ’s degree also in mechanical engi neering. After completing a short work experience he returned to th e University of Florida to pursue a doctoral degree. Upon graduation, the author would like to pursue engineering opportunitie s that will enhance individual independence.