Top Banner
Graduate School ETD Form 9 (Revised 12/07) PURDUE UNIVERSITY GRADUATE SCHOOL Thesis/Dissertation Acceptance This is to certify that the thesis/dissertation prepared By Entitled For the degree of Is approved by the final examining committee: Chair To the best of my knowledge and as understood by the student in the Research Integrity and Copyright Disclaimer (Graduate School Form 20), this thesis/dissertation adheres to the provisions of Purdue University’s “Policy on Integrity in Research” and the use of copyrighted material. Approved by Major Professor(s): ____________________________________ ____________________________________ Approved by: Head of the Graduate Program Date Abhinav Krishna ORGANIC RANKINE CYCLE WITH SOLUTION CIRCUIT FOR LOW-GRADE HEAT RECOVERY Master of Science in Mechanical Engineering Eckhard A. Groll Suresh V. Garimella James E. Braun W. Travis Horton Eckhard Groll David Anderson 07/25/2012
135

ORGANIC RANKINE CYCLE WITH SOLUTION CIRCUIT FOR LOW-GRADE HEAT RECOVERY

Sep 12, 2015

Download

Documents

soylashushu

ORGANIC RANKINE CYCLE WITH SOLUTION CIRCUIT FOR LOW-GRADE
HEAT RECOVERY
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
  • Graduate School ETD Form 9 (Revised 12/07)

    PURDUE UNIVERSITY GRADUATE SCHOOL

    Thesis/Dissertation Acceptance

    This is to certify that the thesis/dissertation prepared

    By

    Entitled

    For the degree of

    Is approved by the final examining committee:

    Chair

    To the best of my knowledge and as understood by the student in the Research Integrity and Copyright Disclaimer (Graduate School Form 20), this thesis/dissertation adheres to the provisions of Purdue Universitys Policy on Integrity in Research and the use of copyrighted material.

    Approved by Major Professor(s): ____________________________________

    ____________________________________

    Approved by: Head of the Graduate Program Date

    Abhinav Krishna

    ORGANIC RANKINE CYCLE WITH SOLUTION CIRCUIT FOR LOW-GRADE HEATRECOVERY

    Master of Science in Mechanical Engineering

    Eckhard A. Groll

    Suresh V. Garimella

    James E. Braun

    W. Travis Horton

    Eckhard Groll

    David Anderson 07/25/2012

  • Graduate School Form 20 (Revised 9/10)

    PURDUE UNIVERSITY GRADUATE SCHOOL

    Research Integrity and Copyright Disclaimer

    Title of Thesis/Dissertation:

    For the degree of Choose your degree

    I certify that in the preparation of this thesis, I have observed the provisions of Purdue University Executive Memorandum No. C-22, September 6, 1991, Policy on Integrity in Research.*

    Further, I certify that this work is free of plagiarism and all materials appearing in this thesis/dissertation have been properly quoted and attributed.

    I certify that all copyrighted material incorporated into this thesis/dissertation is in compliance with the United States copyright law and that I have received written permission from the copyright owners for my use of their work, which is beyond the scope of the law. I agree to indemnify and save harmless Purdue University from any and all claims that may be asserted or that may arise from any copyright violation.

    ______________________________________ Printed Name and Signature of Candidate

    ______________________________________ Date (month/day/year)

    *Located at http://www.purdue.edu/policies/pages/teach_res_outreach/c_22.html

    ORGANIC RANKINE CYCLE WITH SOLUTION CIRCUIT FOR LOW-GRADE HEATRECOVERY

    Master of Science in Mechanical Engineering

    Abhinav Krishna

    07/12/2012

  • ORGANIC RANKINE CYCLE WITH SOLUTION CIRCUIT FOR LOW-GRADE HEAT RECOVERY

    A Thesis

    Submitted to the Faculty

    of

    Purdue University

    by

    Abhinav Krishna

    In Partial Fulfillment of the

    Requirements for the Degree

    of

    Master of Science in Mechanical Engineering

    August 2012

    Purdue University

    West Lafayette, Indiana

  • All rights reserved

    INFORMATION TO ALL USERSThe quality of this reproduction is dependent upon the quality of the copy submitted.

    In the unlikely event that the author did not send a complete manuscriptand there are missing pages, these will be noted. Also, if material had to be removed,

    a note will indicate the deletion.

    Microform Edition ProQuest LLC.All rights reserved. This work is protected against

    unauthorized copying under Title 17, United States Code

    ProQuest LLC.789 East Eisenhower Parkway

    P.O. Box 1346Ann Arbor, MI 48106 - 1346

    UMI 1529707Published by ProQuest LLC (2012). Copyright in the Dissertation held by the Author.

    UMI Number: 1529707

  • ii

    ACKNOWLEDGEMENTS

    The contents of this work, and my personal growth during its development, is due

    in no small part to several people. My advisors Eckhard Groll and Suresh Garimella

    provided plenty of guidance and support, not to mention commensurate intellectual

    freedom, throughout my Masters studies. Jim Braun and Travis Horton thank you for

    all your useful insights during the course of this project.

    I would like to thank everyone that helped in the design and construction of the

    challenging experimental setup for this project. Frank, Bob and Gilbert thank you for

    your efforts. Special thanks to two tireless students, Philipp Danecker and Nick Czapla,

    for the countless hours you spent working on the setup. Your motivation and competence

    went a long way in raising my enthusiasm for this project.

    I have had the privilege of working with some of the best graduate students

    around. Brandon Woodland unassuming and soft spoken provided plenty of brain

    power for this work. Not that he represents a victory of substance over style, because he

    has plenty of both. Ian Bell and Craig Bradshaw provided quality mentorship and plenty

    of ideas. My other colleagues at Herrick Laboratories including Bryce, Christian, Dong

    Han, Howard, Huize, Simba, Stephen and several others have lent me their time, of

    which they had precious little. I have been lucky to foster several friendships that extend

    beyond professional interests during my time at Herrick Laboratories.

  • iii

    Finally, I would like to thank the Cooling Technologies Research Center, and

    TORAD Engineering for funding this research and providing customized equipment. I

    would also like to thank ASHRAE for providing me with financial support during my

    Masters studies.

  • iv

    TABLE OF CONTENTS

    Page

    LIST OF TABLES ............................................................................................................. vi LIST OF FIGURES .......................................................................................................... vii NOMENCLATURE .......................................................................................................... ix ABSTRACT ................................................................................................................. xii CHAPTER 1. INTRODUCTION ......................................................................................1

    1.1 Background ........................................................................................ 1 1.2 Motivation .......................................................................................... 2 1.3 Objective ............................................................................................ 3 1.4 Approach ............................................................................................ 6

    CHAPTER 2. CURRENT STATUS OF TECHNOLOGY ...............................................8 2.1 Organic Rankine Cycles .................................................................... 8 2.2 Absorption Power Cycles .................................................................. 9 2.3 Working Fluid Mixtures .................................................................. 11

    CHAPTER 3. THERMODYNAMIC MODEL DEVELOPMENT AND RESULTS .......................................................................................13

    3.1 Baseline Cycles ................................................................................ 13 3.1.1 Organic Rankine Cycle ............................................................ 13 3.1.2 Vapor Compression Cycle with Solution Circuit ..................... 14 3.1.3 Organic Rankine Cycle with Solution Circuit .......................... 16

    3.2 Thermodynamic Features of Binary Mixtures ................................. 18 3.2.1 Phase Equilibrium .................................................................... 18 3.2.2 Absorption / Desorption Process .............................................. 20 3.2.3 Temperature Glide and Capacity Control ................................ 22

    3.3 Cycle Model of an Organic Rankine Cycle with Solution Circuit .. 25 3.3.1 Mass Balance ............................................................................ 27 3.3.2 Energy Balance ......................................................................... 28

    3.4 Organic Rankine Cycle with Solution Circuit Model Results ......... 32 CHAPTER 4. DESIGN OF EXPERIMENTAL TEST SYSTEM ..................................45

    4.1 System Sizing................................................................................... 45 4.2 System Design and Layout .............................................................. 48 4.3 Design and Selection of Major Components ................................... 53

    4.3.1 Pump ......................................................................................... 53 4.3.2 Heat Exchangers ....................................................................... 57 4.3.3 Expander ................................................................................... 60 4.3.4 Separator ................................................................................... 64

  • v

    Page

    4.3.5 Receiver .................................................................................... 66 4.3.6 Expansion Valves ..................................................................... 66 4.3.7 Instrumentation and Data Acquisition ...................................... 67

    4.4 Experimental Error and Uncertainty ................................................ 68 CHAPTER 5. EXPERIMENTAL RESULTS AND ANALYSIS ...................................71

    5.1 Experimental Program Overview .................................................... 71 5.2 Experimental Performance Trends .................................................. 75 5.3 Experimental Program Summary ..................................................... 81

    CHAPTER 6. PERFORMANCE CHARACTERIZATION OF THE ORCSC SYSTEM ....................................................................................83

    6.1 Parametric Analysis ......................................................................... 83 CHAPTER 7. CONCLUSIONS AND FUTURE WORK ...............................................90

    7.1 Conclusion ....................................................................................... 90 7.2 Recommendations for Future Work................................................. 92

    LIST OF REFERENCES ...................................................................................................95 APPENDICES

    Appendix A Instrument Calibration Procedures .................................................. 98 Appendix B Wiring Diagram for Data Acquisition System............................... 105 Appendix C Procedures for Charging and Discharging the ORCSC System .... 106 Appendix D Procedures for Operating the ORCSC System .............................. 108 Appendix E Data from Experimental Testing .................................................... 112

  • vi

    LIST OF TABLES

    Table .............................................................................................................................. Page

    Table 3-1: Key assumptions used in the thermodynamic model. ..................................... 32 Table 4-1: Sample set of input parameters used as a basis for system design. ................. 46 Table 4-2: Sample property values at each state point in the

    ORCSC cycle using the LKP EOS (refer to Figure 3-3). ................................46 Table 4-3: Capacities, flow rates and cycle efficiencies used as a design

    basis for the ORCSC. .......................................................................................47 Table 4-4: Qualitative comparison for different pump types for an ORCSC

    application. .......................................................................................................54 Table 4-5: Primary specifications of CAT PUMP MODEL 1051.CO2. .......................... 56 Table 4-6: Data acquisition devices and their output signals. .......................................... 68 Table 4-7: Measurement uncertainties. ............................................................................. 70 Table 5-1: Experimental test matrix. ................................................................................ 72 Table 5-2: Key assumptions used in the thermodynamic model for

    experimental comparison. ................................................................................75 Table 6-1: Baseline conditions for parametric analysis. ................................................... 84 Table E-1: Experimental temperature data. .................................................................... 112 Table E-2: Experimental temperature data continued. ................................................... 113 Table E-3: Experimental pressure data. .......................................................................... 114 Table E-4: Experimental flow rate data. ......................................................................... 115 Table E-5: Experimental power, performance and heat transfer data. ........................... 116

  • vii

    LIST OF FIGURES

    Figure ............................................................................................................................. Page

    Figure 1-1: Heat sources, their temperature levels and heat recovery technologies. .......... 4 Figure 3-1: Schematic representation of the Organic Rankine Cycle. .............................. 13 Figure 3-2: Schematic representation of the Vapor Compression Cycle

    with Solution Circuit. .....................................................................................15 Figure 3-3: Schematic representation of the Organic Rankine Cycle

    with Solution Circuit. .....................................................................................17 Figure 3-4: p, T, diagram for an arbitrary mixture (modified from Kyle 1999). ........... 20 Figure 3-5: Variation of liquid and vapor composition with temperature

    for a binary mixture. .......................................................................................21 Figure 3-6: Irreversibilities in heat transfer processes (modified from Mulroy 1993). .... 24 Figure 3-7: Impact of ORCSC concentration pairings on efficiency for a

    source temperature of 100C and a sink temperature of 20C. ......................33 Figure 3-8: Variation of overall Second Law Efficiency as a function of

    Circulation Ratio. ...........................................................................................34 Figure 3-9: Impact of rich solution concentration on efficiency for various source

    temperatures. Plot is optimized for weak solution concentration. .................35 Figure 3-10: Second Law efficiency as a function of source temperature

    for a basic ORC. ............................................................................................36 Figure 3-11: Second Law efficiency as a function of source temperature

    for an ORC with internal regeneration at the expander outlet. .....................36 Figure 3-12: Second Law efficiency as a function of source temperature

    for the ORCSC. .............................................................................................37 Figure 3-12: Variation of component capacities as a function of Circulation Ratio. ....... 40 Figure 3-13: Variation of High-side system Pressure as a function of

    Circulation Ratio. ..........................................................................................41 Figure 3-14: Variation of Pressure Ratio as a function of Circulation Ratio.................... 41 Figure 4-1: Detailed schematic diagram of the secondary loop for heat input

    temperature control. ........................................................................................49 Figure 4-2: Detailed schematic diagram of the primary ORCSC system. ........................ 49 Figure 4-3: Preliminary CAD representation of ORCSC system arrangement. ............... 51 Figure 4-4: Top view of the primary ORCSC experimental load stand. .......................... 52 Figure 4-5: Front view of the primary ORCSC experimental load stand. ........................ 52 Figure 4-6: Secondary heat input loop load stand. ........................................................... 53 Figure 4-6: Sectional view of CAT PUMP MODEL 1051.CO2. ..................................... 56 Figure 4-7: Dual Coil Heat Exchangers selected for use in the ORCSC. ......................... 58

  • viii

    Figure ............................................................................................................................. Page

    Figure 4-8: Sanden TRS-105 scroll compressor housing. ................................................ 62 Figure 4-9: Front view of Torad rotary spool expander housing. ..................................... 63 Figure 4-10: Design of the Torad rotary spool expander. ................................................. 64 Figure 4-11: Design and layout of separator and sight glass. ........................................... 65 Figure 4-12: Receiver and sight glass selected for use in the ORCSC. ............................ 66 Figure 4-13: General overview of the data aquistion system. .......................................... 67 Figure 5-1: Comparison of theoretical and experimental Second Law

    efficiencies as a function of the Circulation Ratio. ........................................77 Figure 5-2: Comparison of theoretical and experimental Desorber Heat

    Input (System Capacity) as a function of the Circulation Ratio. ....................78 Figure 5-3: Comparison of theoretical and experimental Second Law Efficiencies

    as a function of the Desorber Exit Temperature. ............................................78 Figure 5-4: Comparison of theoretical and experimental Expander Pressure Ratio

    as a function of the Circulation Ratio. ............................................................79 Figure 5-5: Comparison of theoretical and experimental Pump Work Input as a

    function of the Circulation Ratio. ...................................................................80 Figure 5-6: Comparison of theoretical and experimental Expander Work Output as a

    function of the Circulation Ratio. .................................................................. 80 Figure 6-1: Net power output and Second Law Efficiency as a

    function of expander isentropic efficiency. ....................................................85 Figure 6-2: Net power output and Second Law Efficiency as a function of pump

    efficiency. .......................................................................................................86 Figure 6-3: Net power output and Second Law Efficiency as a function of the

    desorber pinch point temperature. ..................................................................87 Figure 6-4: Net power output and Second Law Efficiency as a function of the

    absorber pinch point temperature. ..................................................................88 Figure 6-5: Net power output and Second Law Efficiency as a function of the

    internal heat exchanger effectiveness. ............................................................89 Figure A 1: Sample calibration data for a thermocouple. ................................................. 99 Figure A 2: Sample calibration data for a pressure transducer. ...................................... 101 Figure A 3: Sample calibration data for a mass flow meter. .......................................... 102 Figure A 4: Top view of expander and torque cell. ........................................................ 103 Figure A 5: Calibration of the torque cell. ...................................................................... 104 Figure A 6: Sample calibration data for torque cell. ....................................................... 104

  • ix

    NOMENCLATURE

    Symbols

    A area CR Circulation Ratio,

    F degrees of freedom

    g gravitational acceleration

    enthalpy

    IHX Internal Heat Exchanger

    K number of components

    gm mass flow rate of glycol

    rm mass flow rate of rich solution

    wm mass flow rate of weak solution

    vm mass flow rate of vapor

    pressure

    Q capacity

    AQ absorber capacity

    DQ desorber capacity

    ,h wQ heating capacity for water

    ,h totQ total heating capacity

    T temperature V volumetric flow rate

    h

    p

  • x

    w overall uncertainty

    W work

    quality

    z measured quantity

    difference

    heat exchanger effectiveness

    efficiency

    relative humidity density

    number of phases

    concentration

    specific volume

    Tor torque

    rotational speed

    Subscripts

    a air

    A absorber

    c cold fluid

    car carnot

    comp compressor

    cond condensing

    crit critical

    D desorber

    h hot fluid

    i inlet

    IHX Internal Heat Exchanger

    meas measured

    o outlet

    x

  • xi

    pump pump

    pinch pinch-point temperature

    r rich solution

    ref refrigerant

    s isentropic

    sat saturation

    sub sub-cooling

    sup superheat

    turb turbine

    tot total

    w weak solution

    v vapor

    Acronyms

    DAQ Data Aquisition

    EOS equation of state

    EPDM Ethylene Propylene Diene Monomer

    LKP Lee-Kesler-Plcker equation of state

    ORC Organic Rankine Cycle

    ORCSC Organic Rankine Cycle with Solution Circuit

    PR Pressure Relief

    PS Pressure Switch

    PTFE Polytetrafluoroethylene

    VCCSC Vapor Compression Cycle with Solution Circuit

    VLE vapor-liquid-equilibrium

  • xii

    ABSTRACT

    Krishna, Abhinav. M.S.M.E., Purdue University, August 2012. Organic Rankine Cycle with Solution Circuit for Low-Grade Heat Recovery. Major Professors: Eckhard A. Groll, Suresh V. Garimella.

    Increasing interest in utilizing low-grade heat for power generation has prompted

    a search for ways in which the power conversion process may be enhanced. A novel

    Organic Rankine Cycle with Solution Circuit (ORCSC) using Carbon Dioxide / Acetone

    as the working fluid pair was studied for this purpose. A thermodynamic simulation

    model was developed and an experimental test stand was built to serve as a proof of

    concept for the technology.

    The thermodynamic model showed that the ORCSC using Carbon Dioxide /

    Acetone as the working pair offers no significant efficiency improvements over a

    conventional Organic Rankine Cycle (ORC) using only Carbon Dioxide as the working

    fluid. Furthermore, the ORCSC with a Carbon Dioxide / Acetone working pair has

    significantly lower performance than an ORC using conventional working fluids such as

    pentane or R245fa. This may render the ORCSC unattractive since the low-temperature

    heat sources mean that the theoretical (Carnot) efficiency limit is itself relatively low, and

    achieving cycle efficiencies as close to the Carnot limit as possible is necessary for

    ensuring the economic feasibility of the technology. However, the ORCSC was

  • xiii

    found to have significantly lower working pressures than an ORC, provides the ability to

    use temperature glide to match the temperature profiles of the source and sink fluids and

    facilitates intrinsic capacity control. This may lead to higher overall system efficiencies

    when coupled with sources that have varying heat input temperatures or loads. More

    application-specific studies that address the nature and capacity of the source and sink

    streams are required to identify where this ability may be most advantageous.

    The experimental tests showed good agreement with the simulation data when all

    the boundary conditions were matched. However, the efficiencies of the system were

    generally poor and many of the expected trends were skewed due to design shortcomings

    and the use of equipment that was not optimized for the ORCSC system. Isolation of

    individual parameters was an acute challenge due to the number of variables that need to

    be tightly controlled during system operation. Nevertheless, the experimental results

    provided a validation of the simulation model.

    The simulation model was expanded to include a parametric study of the various

    components on the overall system performance. It showed that the ORCSC is particularly

    sensitive to the performance of the expander and pump. Amongst the heat exchangers,

    the performance of the absorber had the greatest impact on the overall system

    performance.

    It is clear from this study that a range of practical considerations need to be taken

    into account and weighed together with the thermodynamic analysis when evaluating the

    feasibility of ORCSC technology. The ORCSC offers some potential practical advantages

    which may outweigh the added cost and complexity of these systems in certain

    applications. However, the maturity of the technology and associated body of literature is

  • xiv

    limited, and further work needs to be pursued in this area before widespread adoption of

    the technology is possible.

  • 1

    CHAPTER 1. INTRODUCTION

    1.1 Background

    The increasing cost of energy, coupled with the recent drive for energy security

    and climate change mitigation have provided the impetus for harnessing renewable

    energy sources as viable alternatives to conventional fossil fuels. However, several of

    these renewable energy sources, including geothermal, biomass and solar, intrinsically

    provide low-grade heat (at temperatures between 60C 300C). Furthermore,

    thermodynamic considerations mean that a large amount of low-grade heat is discharged

    from power plants and various other industrial processes. In fact, in the United States,

    over two-thirds of the primary energy supply is ultimately rejected as low-grade waste

    heat according to the World Energy Council (2006). Recovering low-grade heat,

    therefore, is increasingly becoming an economic and environmental imperative.

    Low-grade waste heat, largely at a temperature level between 30C 250C, is

    primarily a product of thermo-mechanical energy conversion losses. Thus far, waste heat

    recovery systems have mainly been designed for thermal heat recovery for process use

    using recuperative heat exchangers. Conversions from waste heat to higher forms of

    energy, such as shaft work and/or electricity, have not been viable due to technical and

    cost impediments. Nevertheless, due to increasing energy costs and available heat sources,

  • 2

    waste heat recovery systems are of increasing interest to designers, engineers and society

    at large (Little 2009).

    1.2 Motivation

    An increasing source of waste heat comes from growing technology needs. In the

    last twenty years, computer power consumption has increased exponentially. Of the total

    electricity consumption in the United States in 2006, more than 1 % was used to operate

    large data centers (Koomey 2007). According to ASHRAE (2005), datacom workload is

    expected to rise further at a 40 to 50 % compound growth rate. Additionally, the power

    density of datacom equipment is expected to reach up to 8 kW/m2 for computer servers

    and 15 kW/m2 for high density communication equipment by 2014. In large data centers,

    therefore, an enormous amount of electrical energy is consumed, which is directly

    converted into heat and ultimately rejected to the environment.

    The rejected heat represents a large energy stream that has already been paid for;

    however recovering and utilizing it represents a challenge. This is primarily due to the

    temperature level of the waste heat (50C 85C), which renders it inefficient to convert

    to a more usable form of an energy transport medium, such as electricity, using

    conventional power generation technology. Furthermore, several renewable energy

    sources face the same problem low source temperatures to the point where the

    application of steam Rankine Cycles, ubiquitous in power generation, is grossly

    inefficient and expensive.

    Currently, technologies attempting to provide low-grade heat recovery solutions

    have seen very limited commercialization. This is broadly due to two reasons: lack of

  • 3

    historical research and development in the area of waste heat recovery due to technical

    and cost impediments; and technical challenges associated with scaling the technology

    from utility scale to commercial scale, particularly with regard to expansion machines

    (turbines). However, due to rising primary energy costs and the environmental premium

    being placed on fossil fuels, the conversion from low-grade heat to electrical energy is a

    pressing societal challenge.

    1.3 Objective

    One way to recover low-grade heat, and use it in a power generating capacity, is

    to use Organic Rankine Cycle (ORC) technology. Organic Rankine Cycles differ from

    traditional steam Rankine Cycles in the use of an organic working fluid as opposed to

    water/steam as the working fluid, making them far better suited for low temperature heat

    sources. However, due to the low-temperature heat sources, the theoretical (Carnot)

    efficiency limit is relatively low. Therefore, achieving cycle efficiencies as close to the

    Carnot limit as possible remains a challenge, and is important for ensuring the economic

    feasibility of the technology. An overview of available heat sources and their

    temperature levels, as well as the technologies that could be used for a given temperature

    range to recover the available heat, is given in Figure 1-1.

  • 4

    Figure 1-1: Heat sources, their temperature levels and heat recovery technologies.

    In order to achieve cycle efficiencies as close to the Carnot limit as possible, it

    may be necessary to modify the Organic Rankine Cycle. One novel modification,

    proposed by Maloney and Robertson (1953) as well as Kalina (1983), is to introduce

    absorption technology for power generation. This modification is termed the Organic

    Rankine Cycle with Solution Circuit (ORCSC), also known as the Absorption-Rankine

    cycle. The ORCSC differs from the ORC primarily through the use of a zeotropic

    mixture, consisting of a refrigerant and an absorbent as the working fluids. The

    refrigerant and absorbent are characterized by a large boiling point difference. This

    enables the separation of the more volatile component (the refrigerant) in the vapor phase

  • 5

    from the absorbent solution in the liquid phase. The refrigerant vapor (and a small

    quantity of the solution) then flows through the expansion device, whereas the liquid

    absorbent forms a solution circuit.

    Aside from the possibility of higher exergetic conversion efficiency than an ORC,

    the ORCSC has several other inherent features that address critical issues related to the

    general applicability of low-grade heat recovery technology. For example, the solution in

    the ORCSC ensures significantly lower working pressures when compared to a

    conventional ORC. This allows for the use of high-pressure natural refrigerants, such as

    carbon dioxide, at moderate operating pressures which leads to significant cost savings.

    The use of natural refrigerants has numerous environmental benefits, including negligible

    Global Warming Potential (GWP), zero Ozone Depletion Potential (ODP), and non-

    toxicity. The use of a zeotropic mixture in the ORCSC also means that the system

    capacity can be easily adjusted by simply changing the concentration of the refrigerant

    used in the system. This offers a simple, cost-effective solution for adapting the system

    for peak and non-peak loads, which the system is bound to encounter in practice.

    Furthermore, as shown in Figure 1-1, the added system complexity of the ORCSC can be

    justified at two ends of the source temperature spectrum: at extremely low source

    temperatures (

  • 6

    temperature range is important because it fills an important technology gap where

    existing conversion systems cannot efficiently generate power. Numerous exhaust heat

    streams from industrial processes, as well as biomass and certain solar concentration

    techniques, fall into this spectrum.

    Despite significant advantages, there have been few ORCSC prototypes that have

    been built. The few prototypes that exist have relied on using an Ammonia / Water

    working pair, which renders them unattractive in many applications due to the corrosive

    properties and toxicity of Ammonia. In this work, the objective is to investigate an

    ORCSC cycle that uses a natural refrigerant in the working pair, and to demonstrate a

    proof-of-concept experimental system.

    1.4 Approach

    In order to fulfill the objectives of this project, the following approach was

    adopted in the given order:

    Literature and patent search of Organic Rankine Cycle technology.

    Identification of methods and measures from the literature to increase the

    efficiency and applicability of Organic Rankine Cycles, including novel cycle

    modifications.

    Evaluation of different working fluids.

    Thermodynamic simulations of possible cycles and assessment of their feasibility.

    Combination of the identified ideas for Organic Rankine Cycle modifications,

    including an Organic Rankine Cycle with Solution Circuit.

    Selection of the most promising working pair.

  • 7

    Detailed simulations and selection of the most promising concept.

    Design and construction of an experimental bread board system based on

    simulation results to serve as a proof of concept for the technology.

    Implementation of the measurement system and development of the test software.

    Experimental tests according to a predetermined test matrix, and determination of

    several key system characteristics such as efficiency, capacity, etc.

    Comparison of measured performance to simulation results.

    Model refinements based on test results.

    Feasibility assessment of the technology, and recommendations for improvements

    to the system based on experimental experience.

    Identification of future technology development needs.

  • 8

    CHAPTER 2. CURRENT STATUS OF TECHNOLOGY

    2.1 Organic Rankine Cycles

    Organic Rankine Cycles with low temperature heat input are relatively well

    known and have been investigated widely in the literature. Theoretical investigations

    were pursued as early as the 1970s by Davidson (1977) for integration with solar

    collectors, and further expanded by Probert et al. (1983). Experimental investigations

    were conducted by Monahan (1976), with reported First Law thermal efficiencies usually

    below 10% for small-scale systems. The experimental investigations identified that

    expansion turbines suitable for use in ORC systems have not been widely studied, and

    few commercial designs were available. In general, experimental investigations have

    largely involved the use of vane expanders (Badr et al. 1990, Davidson 1977).

    Hung et al. (1997) compared the efficiencies for various ORC working fluids such

    as benzene, ammonia, R11, R12, R134a and R113. The study established correlations

    between system efficiencies, source temperatures and system pressures. Of the fluids

    investigated, benzene was found to provide the highest efficiency, followed by R113,

    R11, R12, R134a, and Ammonia.

  • 9

    Despite the high Ozone Depletion Potential (ODP) and Global Warming Potential

    (GWP) of many of these fluids, the first commercial applications appeared in the late 70s

    and 80s with medium-scale power plants developed for geothermal and solar applications.

    Currently, over 300 ORC systems are in operation worldwide, with over 1800 MWe of

    installed capacity (and this number continues to grow at an ever increasing pace). The

    largest number of plants is installed for biomass Combined Heat and Power (CHP)

    applications, followed by geothermal plants and then Waste Heat Recovery (WHR)

    plants (Quoilin 2011). It should be noted, however, that the largest application in terms of

    installed power are geothermal applications. (Enertime 2011).

    2.2 Absorption Power Cycles

    The foundation for the Organic Rankine Cycle with Solution Circuit (ORCSC),

    also known as the Absorption-Rankine Cycle, is the Vapor Compression Cycle with

    Solution Circuit (VCCSC), first investigated by Altenkirch (1950). Groll and

    Radermacher (1994) carried out successful experimental testing to demonstrate the

    concept. Reversing the VCCSC creates a power generating cycle similar to the Rankine

    Cycle, but with higher potential efficiencies (Kalina 1983). While Absorption-Rankine

    Cycles have been known for more than 50 years, limited research and even more limited

    experimental investigations have been carried out in this field.

    One experimental investigation was performed by Maloney and Robertson (1953)

    using an Ammonia / Water pair as the working fluid. Their results showed that the

    absorption power cycle had no thermodynamic advantage over the Rankine Cycle.

  • 10

    However, the authors encouraged further investigations in this field, and proposed the use

    of other binary mixtures.

    Further investigations were conducted by Kalina (1983) with the same working

    fluid as Maloney and Robertson, but with a slightly different experimental setup. Kalina

    showed that the cycle has a thermal efficiency that is 30-60% greater than comparable

    steam power cycles at the same source temperature. In these studies, the cycle was

    coupled with relatively high turbine inlet temperatures of 180C.

    Due to the diversity of potential applications, the Kalina cycle was further studied

    for different purposes with slightly different configurations. Goswami and Xu (2000)

    proposed a simple combined cycle using solar energy as the heat source. Zheng et al.

    (2006) modified the Kalina cycle in order to produce power as well as provide

    refrigeration simultaneously, but the investigations carried out were numerical in nature,

    without experimental validation.

    Robbins and Garimella (2010) published theoretical investigations of an Organic

    Rankine Cycle with Solution Circuit using a novel binary mixture of Amyl-Acetate and

    Carbon Dioxide (CO2) as the working fluid. Their parametric theoretical results showed

    promising thermal efficiencies, but once again, experimental validation was not carried

    out.

    In summary, while there have been a few experimental investigations conducted

    for Absorption-Rankine Cycles using Ammonia / Water as the working pair, to the best

    of the authors knowledge, no investigations to date have considered novel working pairs

    in an experimental investigation of the Absorption-Rankine Cycle.

  • 11

    2.3 Working Fluid Mixtures

    The choice of a working fluid mixture for the ORCSC is defined by the

    requirement that the two fluids have a large boiling point difference. This enables the

    more volatile component (refrigerant) to easily separate from the liquid absorbent, and

    for the system to accommodate a large range of source and sink temperatures simply by

    adjusting the composition of the mixture. Generally, the working fluid mixtures found in

    absorption cycles may also be used in the ORCSC. Ammonia / Water mixtures have been

    studied extensively in the literature; however, ammonia has the obvious drawback of

    being toxic and corrosive, which limits its potential applications. An Amyl-Acetate /

    Carbon Dioxide mixture was studied for use in an ORCSC (Robbins and Garimella,

    2010); however, the operating pressures were found to be too high for this working pair.

    Groll and Radermacher (1994) studied the use of a Carbon Dioxide / Acetone mixture in

    a VCCSC application, and Carbon Dioxide was found to have the following advantages:

    Low Global Warming Potential and zero Ozone Depletion Potential as compared

    to conventional refrigerants. This is pertinent given that the application of this

    technology ultimately focuses on mitigating environmental impact.

    Non toxicity.

    Non-flammability in this working pair, CO2 can be considered an ideal fire

    extinguishing medium for the Acetone. In case of a leak, a large amount of CO2

    and a small quantity of Acetone would escape from the system due to the

    difference in the vapor pressures of the two components.

    Large volumetric heat capacity, which enables the use of smaller turbines and

    other components.

  • 12

    Compatibility with common component materials.

    Simplicity of operation.

    No recycling of working fluid required.

    Furthermore, Acetone was chosen as the absorbent solution for the following reasons:

    Higher overall thermodynamic efficiencies due to its ability to dissolve CO2.

    Wide availability at low cost.

    Lower flammability when compared to other hydrocarbon solutions.

    Given these advantages, a compelling case can be made for a Carbon Dioxide / Acetone

    working pair to be investigated in an ORCSC system.

  • 13

    CHAPTER 3. THERMODYNAMIC MODEL DEVELOPMENT AND RESULTS

    3.1 Baseline Cycles

    3.1.1 Organic Rankine Cycle

    The conceptual foundation for the ORCSC lies in a combination of a conventional

    Organic Rankine Cycle with a Vapor Compression Cycle with Solution Circuit (VCCSC).

    A simplified schematic diagram of an ORC is shown in Figure 3-1 below.

    Figure 3-1: Schematic representation of the Organic Rankine Cycle.

    The ORC is essentially made up of the four main components found in traditional

    steam Rankine Cycles: an evaporator, expander, condenser, and pump. However, unlike

    in steam Rankine cycles, there is usually no water-steam separation drum connected to

    Expander

    Pump Condenser

    Evaporator

    Low-Grade Heat Source TH

    Environment TL

    (1)

    (2)

    (3)

    (4)

  • 14

    the boiler, and one single heat exchanger is used to perform all three evaporation phases:

    preheating, vaporization and superheating. Due to a combination of cost impediments,

    system scale and thermodynamic properties of the chosen working fluid, reheating and

    turbine bleeding are generally not suitable for the ORC. However, an internal regenerator

    installed at the expander outlet is often used to preheat the liquid from the pump outlet.

    The working principle is as follows: the refrigerant leaves the evaporator as a

    supercritical gas (1), which then enters an expander. The thermal expansion through the

    expander produces mechanical shaft power (2), which can be converted to electrical

    energy in a generator. The refrigerant then enters the condenser, where it is converted to

    the liquid phase by rejecting heat to the ambient (3). The liquid refrigerant is then

    pumped back to the evaporator inlet to complete the cycle (4).

    3.1.2 Vapor Compression Cycle with Solution Circuit

    The VCCSC combines absorption and compression technology, and differs from the

    conventional vapor compression cycle primarily by employing a working fluid mixture

    consisting of a refrigerant and an absorbent, instead of pure components. A schematic

    diagram of the VCCSC is provided in Figure 3-2.

  • 15

    Figure 3-2: Schematic representation of the Vapor Compression Cycle with Solution Circuit.

    The mixture is evaporated in the desorber using a heat source (which may be the

    conditioned space); however, the evaporation of the mixture is not complete. Instead, a

    liquid and vapor exist at the same temperature and pressure (with differing concentrations)

    in the desorber. The refrigerant-rich vapor and low-concentration liquid (i.e. weak

    solution) are separated at the desorber outlet. Note that an additional separator may be

    necessary to complete the separation process. While the vapor (1) proceeds to the

    mechanically driven compressor, the weak solution liquid (6) is pumped to a recuperative

    internal heat exchanger (7) where it is used to preheat the rich solution from the absorber

    (4). At the other end of the cycle, the compressed refrigerant (2) and weak solution (8)

    enter the absorber. Since absorption is an exothermic process, a heat and mass exchange

    process takes place in which heat is rejected to the environment (or other appropriate heat

    sink) and the refrigerant is simultaneously resorbed into the absorbent, thereby forming a

    Heat Source TH

    Solution Pump

    Absorber

    Expansion Valve

    Heat Sink TL

    Compressor

    (2)

    (3)(4)

    (5)

    (6)

    (7)

    (1)

    Desorber

    Internal Heat Exchanger

    Weak Solution

    Rich Solution

    (8)

  • 16

    rich solution (3). An expansion valve is used to equalize the high-side and low-side

    pressures (5).

    3.1.3 Organic Rankine Cycle with Solution Circuit

    The ORCSC reverses the VCCSC and applies the operating principles of an ORC

    to create a power generating cycle based on the use of a binary mixture. Figure 3-3 shows

    a schematic diagram of the Organic Rankine Cycle with Solution Circuit. State (1)

    represents the outlet of the desorber (note that a separator may be used to separate the

    vapor and liquid streams at the desorber outlet). At this state, the heat source has heated

    the mixture, and the primary working fluid is desorbed from the solution. The CO2 vapor

    stream then enters the expander, where it is expanded to its low pressure state (2) while

    producing mechanical shaft power. State (3) represents the outlet of the absorber, where

    the CO2 has been resorbed into the solution to form a rich solution. Since absorption is an

    exothermic process, the absorber rejects heat to the environment during this process.

    Following this, the rich solution is pumped to the high pressure state (4) by means of a

    solution pump, and is subsequently preheated by an internal heat exchanger (5) before

    entering the desorber. State (6) represents the liquid phase weak solution at the desorber

    outlet. The weak solution stream is then subcooled (7) through the internal heat

    exchanger, and expanded to the low pressure state (8) by an expansion valve.

  • 17

    Figure 3-3: Schematic representation of the Organic Rankine Cycle with Solution Circuit.

    The main difference between the conventional vapor compression cycle and ORC

    when compared to the VCCSC and ORCSC, respectively, is the use of a zeotropic

    mixture with a large boiling point difference instead of a pure fluid. By introducing such

    a working fluid mixture, three important features are accomplished:

    1) Although desorption and absorption occur at constant pressures, the saturation

    temperatures are no longer constant but vary with the composition changes of the

    liquid and the vapor phases which occur during the phase change processes. This

    results in a temperature glide in the absorber and desorber. These temperature

    glides can be adjusted over a wide range or eliminated almost entirely.

    2) A change in the overall concentration of the mixture circulating through the cycle

    results in a change of the vapor pressures and densities at a given temperature,

    and therefore in a change of the capacity of the entire unit.

    Low-Grade Heat Source TH

    Pump

    Desorber

    Expansion Valve

    Environment TL

    Expander

    (2)

    (3)

    (4)

    (5)(6)

    (7)

    (8)

    (1)

    Absorber

    Internal Heat Exchanger

    Weak Solution

    Rich Solution

  • 18

    3) By introducing a solution in the cycle, the operating pressures of the cycle are

    significantly reduced. The solution allows the resorption temperature of the

    mixture to be higher than the critical temperature of the pure refrigerant.

    All of these features can be used to increase the overall COP and expand the

    flexibility of system operation. However, in order to understand these features more fully

    and evaluate their applicability in a power generating cycle, a detailed study of

    multiphase-multicomponent systems is necessary.

    3.2 Thermodynamic Features of Binary Mixtures

    This section provides a background on the thermodynamic treatment of binary

    mixtures. Since the proposed cycle utilizes a two component mixture, emphasis is placed

    on binary working fluids; however, the same concepts would hold if multicomponent

    mixtures were considered. A basic understanding of binary mixture behavior is essential

    to understand features such as temperature glide and capacity control that are integral

    facets of the ORCSC.

    3.2.1 Phase Equilibrium

    A single fluid is considered to be in phase equilibrium when successive pressure

    and temperature measurements of the liquid and vapor phase do not vary with time. For a

    binary mixture to be in phase equilibrium, in addition to pressure and temperature, the

    concentration of each component may not vary with time. In general, the Gibbs phase

    rule ( 3-1 ) gives the degrees of freedom of a system:

  • 19

    F K 2= + ( 3-1 )

    The degrees of freedom (F) depends on the number of components (K), and on the

    number of phases ( ) that are prevalent. When F number of intensive variables are

    specified, the system is determinate. Therefore, in order to visualize the phase behavior of

    a binary system, three intensive variables need to be considered when considering a

    single-phase region (two intensive variables are needed when considering the two-phase

    region). Typically, the variables chosen are the temperature (T), pressure (p) and the mass

    concentration ( ). In the context of this work, when considering a binary mixture

    consisting of a primary working fluid and an absorbent, the mass concentration may be

    defined as:

    liquidmass of primary working fluid in the liquid phase

    mass of both componets in the liquid phase = ( 3-2 )

    vapormass of primary working fluid in the vapor phase

    mass of both componets in the vapor phase =

    ( 3-3 )

    For the mass concentration of a binary mixture, the knowledge of either the vapor

    or liquid phase is sufficient, because the corresponding mass fractions can be derived

    from the knowledge of the other. Figure 3-4 represents a p, T, diagram for an arbitrary mixture.

  • 20

    Figure 3-4: p, T, diagram for an arbitrary mixture (modified from Kyle 1999).

    The thick, solid line represents the saturated liquid boundary; any point above the

    line would indicate that the binary mixture is in the subcooled region and a single liquid

    phase exists. Similarly, the dashed line corresponds to the saturated vapor boundary; any

    point below the line would indicate that the binary mixture is in the superheated phase.

    For a point located in between the two boundaries, an equilibrium between the liquid and

    vapor phases exists. Therefore, the pressure and temperature are the same for both phases,

    while concentrations differ in each phase. This is discussed in further detail below.

    3.2.2 Absorption / Desorption Process

    In order to understand the thermodynamic behavior of binary mixtures more fully,

    it is necessary to illustrate the absorption and desorption processes in detail. Figure 3-5

  • 21

    shows a two-dimensional rendition of Figure 3-4 where the pressure is held constant, i.e.,

    the horizontal plane in Figure 3-4.

    Figure 3-5: Variation of liquid and vapor composition with temperature for a binary mixture.

    Point 1 indicates a subcooled binary mixture in the liquid phase for a given mass

    concentration (1), temperature (T1) and pressure (p). As the solution is heated at constant

    pressure, point 1 moves up vertically until it reaches the saturated liquid line. The first

    vapor bubbles begin to form at temperature T2. At equilibrium, a horizontal line ties

    together the saturated liquid and vapor curves. These lines, called tie lines, reflect the

  • 22

    equilibrium between liquid and vapor compositions (, and ,) at a constant pressure and temperature for a binary mixture. At point 2, the mass fraction of the liquid phase is

    exactly the same as for the subcooled liquid at point 1(l). The mass fraction of the vapor

    phase is given by v,2, where the pressure and temperature are the same as the liquid

    phase as a condition for equilibrium. Note that the vapor concentration is rich in the

    primary fluid, because the boiling temperature ( boil,primary fluidT ) of the pure refrigerant is

    lower than the boiling point of the pure absorbent ( boil,absorbentT ). As the mixture is heated

    further at constant pressure to state point 3, the amount of primary fluid (l,3) remaining

    in liquid phase is less than at state point 2 (l). Accordingly, the vapor phase (v,3) also

    contains a higher mass fraction of absorbent and a lower mass concentration of the

    primary fluid compared to state point (2). Continued heating eventually moves the

    mixture to point 4, where the last remaining droplet is evaporated. At this point, the vapor

    has exactly the same mass concentration (v,4) as the sub cooled liquid (l). Further

    heating leads to a superheated vapor (5) at the same concentration.

    To have a better analytical understanding of two-phase behavior, the phase rule

    ( 3-1 ) can be applied. Given a binary mixture (K=2) and a two-phase region ( = 2) at constant pressure, only one degree of freedom remains. Therefore, either the temperature

    (T) or the mass concentration () can be selected to fix the equilibrium state.

    3.2.3 Temperature Glide and Capacity Control

    A key feature of utilizing a mixture-based cycle is that although desorption and

    absorption occur at constant pressures, the saturation temperatures are no longer constant

  • 23

    but vary with the composition changes of the liquid and the vapor phases that occur during

    the phase change processes. For example, during the evaporation process illustrated in

    Figure 3-5, the saturation temperature changes from T2 to T4. This temperature glide

    becomes larger as the difference in the boiling points of the pure components increases.

    The temperature glide also depends on the mass fraction and the shape of the vapor

    bubble, which is an intrinsic property of the working fluid mixture. In general, the

    temperature glide is larger for intermediate values of initial than for small or large

    values of . An important feature is that the temperature glide can be adjusted over a wide

    range; for example, for a value of 1 in Figure 3-5, the first bubbles form at a higher

    saturation temperature (2) when compared to the initial case of T2 and 1. Additionally,

    the evaporation process is completed at a higher temperature (4 ). As a result, the

    saturation temperatures are shifted to higher temperature levels with a smaller

    temperature glide simply by varying the mass concentration. The same effect can be

    accomplished by having an incomplete evaporation process in which a liquid and vapor

    exist at the same temperature and pressure, but at different concentrations.

    A key feature of the temperature glide is that it can reduce heat transfer

    irreversibilities in the heat exchangers. This is accomplished when the temperature profile

    of the evaporating and condensing mixture matches that of heat-source and heat-sink

    fluids in the counter flow desorber and absorber (Mulroy 1993). This is shown in Figure

    3-6, which illustrates the Carnot and Lorenz cycles operating with the same heat transfer

    fluid temperature profiles at two different source temperatures.

  • 24

    Figure 3-6: Irreversibilities in heat transfer processes (modified from Mulroy 1993).

    The Carnot cycle refers to a pure refrigerant, for which the evaporation and

    condensing process takes place at a constant temperature for a fixed pressure. In

    comparison to the Carnot cycle, the Lorenz cycle refers to the evaporation and

    condensation process of a mixture, with the temperature glide clearly evident during

    these phase change processes. If the profiles of the heat transfer fluid (HTF) flowing

    through a heat exchanger in counterflow is modeled with a non-zero slope, the shaded

    areas approximate the heat transfer irreversibilities. It can be seen that the irreversibilities

    are significantly lower for the Lorenz cycle due to the temperature glide. Furthermore, if

    the temperature of the heat source is increased for the same operating cycle, it can be

    seen from Figure 3-6 that the temperature profile of the HTF changes. The new profile

    can be matched by simply adjusting the mass concentration when using a binary mixture

  • 25

    (e.g. to a value of 1 shown in Figure 3-5). A change in the overall concentration of the

    mixture circulating through the cycle results in a change of the vapor pressures and

    densities at a given temperature, and therefore in a change of the capacity of the system.

    It also shows that for a capacity adjustment, the heat transfer irreversibilities remain

    nearly constant for the Lorenz Cycle, while they increase notably for the Carnot Cycle. It

    is important to note, however, that the saturation temperatures generally do not vary

    linearly as a function of enthalpy for zeotropic mixtures, so that Figure 3-6 can be

    regarded as a simplification.

    Nevertheless, the temperature glide and capacity control are two key features of a

    mixture-based cycle. The temperature glide can be used to reduce heat transfer

    irreversibilities when the source or sink fluids are not approximated as reservoirs, i.e.,

    they have temperature profiles with non-zero slopes through the heat exchange processes.

    Furthermore, since the temperature glide can be adjusted over a wide range of heat-

    source and heat-sink temperatures simply by varying the concentration of` the mixture,

    the capacity of the system can be adjusted for peak and non-peak operation.

    3.3 Cycle Model of an Organic Rankine Cycle with Solution Circuit

    In order to quantify the performance of the ORCSC, it is necessary to build a

    thermodynamic cycle model. To simplify the analysis of the cycle, the following general

    assumptions are made:

    The pressure drop due to frictional losses through the piping and fittings is

    negligible. Therefore, the only pressure drops in the system are across the turbine,

    pump and expansion valve.

  • 26

    The rich solution exiting the absorber and the weak solution exiting the desorber

    are both saturated liquids.

    The vapor stream exiting the desorber is a saturated vapor.

    There is thermodynamic equilibrium between the vapor and liquid phases during

    the absorption and desorption processes.

    All the piping in the system is perfectly insulated.

    There is no oil present in the cycle.

    A simulation model was developed to compute the thermodynamic properties at

    each state point. Since the working fluid is a binary mixture, an equation of state (EOS) is

    required to obtain extensive properties (enthalpy, entropy, etc.) of the mixture at each

    state point. Two equations of state were considered: the correspondence method given by

    the Lee-Kesler-Plcker (LKP) (Plcker et al. 1977), and the Wide Range Equation of

    State by Kunz and Wagner (Kunz et al. 2010). The LKP EOS is given in a form that is

    easily modifiable for several working pairs given appropriate interaction parameters. It

    was found to perform well for a Carbon Dioxide / Acetone mixture in studies conducted

    by Groll and Radermacher (1994). However, the LKP EOS was found to have several

    limitations as listed below:

    Inability to calculate fluid properties accurately when given a two-component

    mixture in the two-phase region. This is particularly important when fixing the

    expander and expansion valve outlet states in the ORCSC cycle.

    Inability to calculate fluid properties accurately in the superheated or subcooled

    regions. This necessitated the assumption of either saturated liquid or saturated

    vapor at each state point.

  • 27

    Limitations in the range of pressures for which property data were available

    (accuracy was limited beyond 70 bars for the CO2 / Acetone mixture).

    Limitations in the range of concentrations for which property data were available

    (accuracy was limited outside the 0.15-0.6 [kgCO2/kgmixture] range for the CO2 /

    Acetone mixture).

    Limitations in the range of temperatures for which property data were available

    (accuracy was limited beyond 150 C for the CO2 / Acetone mixture).

    Due to these limitations, it was found that the cycle performance was significantly over-

    predicted by the LKP EOS.

    The Kunz and Wagner EOS was found to have a greater flexibility of application,

    particularly with regard to the range of temperatures, pressures and concentrations for

    which useful vapor-liquid-equilibrium (VLE) data were available. Furthermore, the Kunz

    and Wagner EOS is integrated with Refprop 9.0 (Lemmon et al., 2012), allowing for easy

    retrieval of the thermodynamic properties. The details of the thermodynamic model are

    given below.

    3.3.1 Mass Balance

    The mass balance for the ORCSC relates the mass flow rates of the rich solution,

    rm , weak solution, wm , and the vapor stream vm .

    r w vm m m= + ( 3-4 )

    Based on the mass concentrations of the rich and weak solutions, the following mass

    balance may be obtained:

  • 28

    r r w w v vm m m = + ( 3-5 )

    A circulation ratio is defined as the ratio of the mass flow rates of the rich solution and

    vapor:

    rv

    v wk

    r wk

    mCRm

    = =

    ( 3-6 )

    3.3.2 Energy Balance

    With reference to Figure 3-3, the heat capacity of the desorber can be calculated using the

    specific enthalpies evaluated at state points 1, 5 and 6:

    ( ) 6 1 51D vQ m CR h h CRh= + ( 3-7 )

    Similarly, the heat capacity of the absorber can be calculated using the specific enthalpies

    evaluated at state points 2, 3 and 8:

    ( )2 8 31A vQ m h CR h CRh= + ( 3-8 )

    The expansion across the expansion valve from state point 7 to 8 is assumed to be an

    isenthalpic process:

    7 8h h= ( 3-9 )

    The energy balance across the internal heat exchanger gives the following equation:

    ( )( )5 4 7 61v v vm CRh m CRh m CR h h= ( 3-10 )

    Based on the isentropic turbine efficiency, turb , the enthalpy at the turbine outlet can be

    calculated as follows:

    h2 = h1 turb h1 h2s( ) ( 3-11 )

    The power output from the turbine is given by:

  • 29

    ( )1 2turb vW m h h= ( 3-12 )

    Similarly, based on the isentropic pump efficiency, pump , the enthalpy at the pump outlet

    can be calculated as follows:

    4 34 3spump

    h hh h

    = + ( 3-13 )

    The power input to the pump is given by:

    ( )4 3pump rW m h h= ( 3-14 )

    Finally, the equations used for the thermal (first law) and second law efficiencies are

    given below:

    turb pumpthermalD

    W WW

    = ( 3-15 )

    sec

    1

    thermalond law

    source

    sink

    TT

    =

    ( 3-16 )

    Note that sourceT and sinkT refer to the temperatures of the heat-source and heat sink

    temperature, respectively, assuming that they can be approximated as constant

    temperature reservoirs. The heat-source and heat-sink temperatures cannot be reached

    because of an incomplete heat transfer in the heat exchangers. Given that the absorber

    and desorber incorporate both heat and mass transfer, and the fact that the working fluid

    may traverse through the subcooled, two-phase and superheated regions in these heat

    exchangers, a traditional definition of heat exchanger effectiveness may not be applied.

    Consequently, pinch-point temperatures were used to predict the heat exchanger outlet

    temperatures. For the internal heat exchanger, however, the weak solution and rich

  • 30

    solution streams remain in the liquid phase, and the traditional heat exchanger

    effectiveness definition is used.

    With reference to Figure 3-3, the following steps were followed to obtain the

    requisite thermodynamic properties at each state point:

    State point 6 is set by assuming a saturated liquid mixture at the desorber outlet

    (quality of zero), desorber outlet temperature, and a weak solution concentration at

    the desorber outlet. Since state points 6 and 1 are in equilibrium in the two-phase

    region, providing a temperature and the weak solution concentration leaving the

    desorber fixes the state. The vapor concentration is returned as part of the outputs.

    Note that a pinch-point is set between the heat source inlet temperature and

    maximum working fluid temperature:

    1,6 1,6 source pinch source pinchT T T T T T= + = ( 3-17 )

    Initially, this pinch-point temperature is set to 10 K.

    State point 1 is set by assuming equilibrium with state point 6. Given the two-

    component two-phase condition that exists at the desorber, only two intensive

    properties need to be given to fix state point 1. These may be the same temperature

    and pressure as state point 6. The vapor concentration is returned by assuming it to

    be a saturated vapor in equilibrium with the saturated liquid weak solution.

    State point 3 is set by initially assuming a saturated liquid mixture at the absorber

    outlet, absorber outlet temperature, and a rich solution concentration. This gives the

    saturation pressure as one of the outputs. Using the saturation pressure, the saturated

    condition may be relaxed, and the state point can be adjusted to incorporate some

    subcooling at the same saturation pressure. Once again, a pinch-point temperature is

  • 31

    applied between the heat sink inlet temperature and minimum working fluid

    temperature.

    3, sin 3 sin sat k pinch sub k pinchT T T T T T T= += + + ( 3-18 )

    Initially, the pinch-point and subcooling temperatures are set to 5 K each.

    State point 4 is set by assuming the same high side pressure returned by state point 6,

    the rich solution concentration as stated above, and the definition of isentropic

    efficiency given by equation ( 3-13 ).

    State point 7 is set by assuming a saturated liquid mixture at the internal heat

    exchanger outlet, weak solution concentration as given by state point 6, and the same

    high side pressure given by state point 6.

    State point 5 may be calculated by using the high side pressure given by state point 6,

    rich solution concentration calculated by the mass balance, and the enthalpy given by

    the internal heat exchanger energy balance in equation (7). Note that an internal heat

    exchanger effectiveness is defined based on fluid enthalpies:

    ( )

    ( ), ,

    max , , min

    h h i h o

    h i c i

    h hQQ h h

    m

    m

    = =

    ( 3-19 )

    Initially, the heat exchanger effectiveness is set to 0.95.

    State point 8 may be fixed by assuming the same enthalpy as state point 7, weak

    solution concentration as given by state point 6, and the same pressure low side

    pressure given by state point 3.

  • 32

    State point 2 is set by assuming the low side pressure set by state point 3, the

    concentration and the entropy given by state point 1, and the definition of isentropic

    efficiency given by equation ( 3-11 ).

    By combining the above steps to calculate the thermodynamic properties with the cycle

    mass and energy balance equations, the properties at each state point as well as an overall

    solution to the cycle model is achieved.

    3.4 Organic Rankine Cycle with Solution Circuit Model Results

    This section presents some of the key results from the thermodynamic cycle

    model. Table 3-1 summarizes some of the key assumptions mentioned above. Unless

    otherwise stated, these assumptions were applied when generating the results described in

    this section.

    Table 3-1: Key assumptions used in the thermodynamic model. Description Value

    Condenser, absorber outlet subcooling 1. 5 C Temperature difference between heat sink and condenser/absorber outlet

    2. 5 C

    Heat sink temperature 3. 20 C Temperature difference between heat source and evaporator/desorber outlet

    4. 10 C

    Regenerator/internal heat exchanger effectiveness 5. 0.95 Pump isentropic efficiency 6. 0.6 Expander isentropic efficiency 7. 0.8 Negligible pressure drop in lines, separators, and heat exchangers 8. Negligible heat loss in lines, mixer, separator, pumps, and expander 9. Complete separation of liquid and gas phases 10.

    Figure 3-7 shows that for a fixed source temperature, the ORCSC has an ideal

    concentration pairing which maximizes the efficiency. This means that for a chosen rich

  • 33

    solution concentration, there exists an optimal weak solution concentration for which the

    cycle performs the best. This is further illustrated in Figure 3-8. Furthermore, as expected,

    the trend shows that higher rich solution concentrations lead to higher efficiencies.

    However, there are other tradeoffs such as high working pressures that may constrain the

    choice of concentration pairings.

    Figure 3-7: Impact of ORCSC concentration pairings on efficiency for a source temperature of 100C and a sink temperature of 20C.

    Figure 3-8 clearly shows that there exists an optimal Circulation Ratio for which

    the cycle performs the best. Based on equation ( 3-6 ), it is clear the Circulation Ratio is

    analogous to the concentration pairing, and therefore confirms the results shown in

    Figure 3-7. The Circulation Ratio, however, is an important physical parameter that is

    essential to the control of an ORCSC system. It fixes the relative mass flow rates of the

    rich solution and vapor, and therefore fixes the relative speeds of the pump and expander

    in the system. By extension, it fixes the concentration pairing of the system. Therefore

  • 34

    the term Circulation Ratio is used in this text as opposed to concentration pairing,

    particularly when discussing the physical ORCSC system and the experimental study.

    Figure 3-8: Variation of overall Second Law Efficiency as a function of Circulation Ratio.

    Figure 3-9 illustrates the effect of rich solution concentration on overall cycle

    Second Law Efficiency for a Carbon Dioxide / Acetone working pair. Note that the

    chosen points have been optimized with respect to weak solution concentration using

    Engineering Equation Solver (EES) (Klein, 2012). From the results shown in Figure 3-7,

    and from a review of absorption studies in the literature, an expected trend would show

    that higher rich solution concentrations lead to higher the efficiencies. The trend in Figure

    3-9 matches this expected outcome for high concentration values. Note that a peak value

    may not occur in the high concentration region because the two-phase concentration

    range shrinks as the mixture critical point is approached. This means that at extremely

    high concentrations; it may not be possible to operate a solution circuit. Furthermore, an

    interesting overall result exists in which the peak efficiency is achieved at extremely low

  • 35

    concentration values. This means that a large fraction of the absorbent, in this case

    Acetone, participates in the expansion process. One reason for this phenomenon, shown

    in Figure 3-10, is that at extremely low concentrations the absorbent, Acetone, is a better

    working fluid than the absorbate, Carbon Dioxide as pure fluid. Therefore, at extremely

    low concentrations, the highest efficiencies are achieved because the component with

    better thermodynamic properties dominates the mixture, while the overall cycle

    simultaneously takes advantage of the internal regeneration intrinsic to the solution

    circuit. Achieving this optimal efficiency in the low-concentration region, however, is a

    practical challenge. Extremely tight tolerances in solution concentrations are required to

    realize these benefits. Nevertheless, the peak efficiencies that occur at extremely low

    concentrations have not been investigated in the literature, and merits further

    investigation in the future.

    Figure 3-9: Impact of rich solution concentration on efficiency for various source temperatures. Plot is optimized for weak solution concentration.

  • 36

    Figure 3-10: Second Law efficiency as a function of source temperature for a basic ORC.

    Figure 3-11: Second Law efficiency as a function of source temperature for an ORC with internal regeneration at the expander outlet.

  • 37

    Figure 3-12: Second Law efficiency as a function of source temperature for the ORCSC.

    Figure 3-10 and Figure 3-11 compare the optimized performance of various

    working fluids for a basic ORC and an ORC with internal regeneration, respectively,

    using the parameters listed in Table 3-1. Figure 3-11 shows that the addition of a

    regenerator results in better performance for dry working fluids, such as R245fa and

    pentane, at high temperatures because the availability in the expander exhaust stream is

    not wasted. However, wet fluids like water and ammonia show no improvement because

    the expander exhaust temperature is not significantly higher than the pump discharge

    temperature.

    Figure 3-12 shows the optimum performance of an ORCSC with Ammonia /

    Water and CO2 / Acetone working fluid pairs at two different concentration ranges: at

    extremely low absorbate concentrations (less than 1% by mass), in which case a large

    fraction of saturated water vapor or Acetone vapor participates in the expansion process;

    and at high concentrations (above 50% by mass), where only small amounts of saturated

  • 38

    Water vapor or Acetone vapor participates in the expansion process. However, at a

    200 C source temperature for the CO2 / Acetone working pair, it was not possible to

    operate at rich solution concentrations above 10% because the two-phase concentration

    range shrinks as the mixture critical point is approached. In this case, the highest possible

    concentration of 10% was used.

    Note that in addition to the assumptions shown in Table 3-1, there were no

    restrictions on vapor quality at the expander exhaust, expander pressure ratios,

    Circulation Ratios, amount of charge required in the system for a given capacity, or

    specific volume of the working fluids at the expander discharge. Therefore, the results

    presented in Figure 3-10, Figure 3-11 and Figure 3-12 are highly idealized and only

    represent what is possible in the thermodynamic design space. For example, Water

    requires a pressure ratio of 296 for optimum efficiency at 200C. It also has a specific

    volume at the expander discharge of nearly 50 m3/kg at the same design point. Designing

    an expansion machine with multiple stages to accommodate the high pressure ratios,

    along with the cost of large components to accommodate the low density renders Water

    an impractical working fluid. Several of these practical concerns are mitigated in the

    ORCSC system.

    The ORCSCs with high concentrations of ammonia or CO2 are generally less

    efficient than traditional ORCs using pure ammonia or CO2 respectively. In these cases

    the temperature glide is detrimental to the Second Law efficiency because it results in

    higher irreversibilities against the assumed constant temperature heat source. A real heat

    source fluid will begin to cool as it heats the cycle working fluid. This could severely

    limit the maximum evaporating temperature of a traditional ORC as the heat source

  • 39

    stream reaches a pinch point against the liquid to two-phase transition of the working

    fluid. In contrast, the ORCSC may be able to reach higher average working fluid

    temperatures by being better able to match the temperature profile of the heat source

    stream with its two-phase temperature glide. A similar argument can be made between

    the working fluid and the heat sink fluid. Exceptions occur for ammonia-water above

    source temperatures of 160 C and for CO2-acetone at 200 C, where the CO2

    concentration is relatively low. The ORCSCs with low concentrations of ammonia and

    CO2 do not perform as well as traditional ORCs with pure water and pure acetone

    respectively (Woodland et al. 2012).

    Figure 3-13 shows the desorber capacity along with the pump power input as a

    function of the Circulation Ratio for a fixed rich solution concentration. Note that the

    desorber capacity and pump power input have been normalized for a 1 kW turbine design.

    As the Circulation Ratio increases, the amount of CO2 in the weak solution is increased

    and the pump input power increases. More important, however, is the fact that the

    desorber capacity decreases as the Circulation Ratio is increased. This illustrates a

    solution for capacity control in this system during non-peak loads, the Circulation Ratio

    may be increased either by a concentration adjustment or, to lesser extent, by varying the

    relative speeds of the pump and expander to accommodate a smaller heat input at the

    source (Krishna et al. 2011).

  • 40

    Figure 3-13: Variation of component capacities as a function of Circulation Ratio.

    Figure 3-14 indicates that the system pressures increase considerably as the

    Circulation Ratio is increased. Therefore, with reference to Figure 3-8, it is clear that

    there is a tradeoff between the efficiency and the operating pressures seen in the system.

    With reference to Figure 3-13, it can also be deduced that there is a relationship between

    capacity control and the system pressures. As the capacity is lowered by increasing the

    Circulation Ratio, the system sees higher pressures. Figure 3-15 shows the linear

    variation of the Pressure Ratio with the Circulation Ratio. This is an important design

    consideration, particularly with regard to the expansion machine. Depending on the

    desired capacity range of the system, it would be advisable to design and optimize the

    expander for the corresponding Pressure Ratio.

  • 41

    Figure 3-14: Variation of High-side system Pressure as a function of Circulation Ratio.

    Figure 3-15: Variation of Pressure Ratio as a function of Circulation Ratio.

    Based on the results shown in Figure 3-10 and Figure 3-12, it may be tempting to

    conclude that the ORCSC does not offer sufficient benefits, and a basic ORC with the

    best performing working fluid would be the ideal choice due to a combination of

  • 42

    simplicity and performance. However, it is important to note that several practical

    considerations should be weighed together with the thermodynamic results presented

    above. For example, the optimum efficiency for a basic ORC with water as the working

    fluid and a source temperature of 200 C is significantly higher than the theoretical

    efficiencies of the ORCSC. However, the use of water requires a low vapor quality and

    low density at the expander exhaust, vacuum pressure condensation, and extremely high

    expander pressure ratios. High pressure ratios require multiple expansion stages. The sub-

    atmospheric pressure makes air leakage into the system a challenge. The low working

    fluid density is a capacity concern, requiring large diameter piping and a large expander

    to achieve a capacity comparable to denser working fluids. Therefore, despite the

    promising efficiency of water as a working fluid for the source temperatures considered,

    the practical issues associated with its use may be prohibitive.

    The hydrocarbons acetone and pentane are very efficient, but they are highly

    flammable and pose challenges for sealing materials. This leads to the widely accepted

    view that R245fa is perhaps the best working fluid in an ORC. However, the ORCSC

    offers some potential practical advantages over an ORC with regeneration using R245fa

    as the working fluid that may outweigh the added complexity of the system. The ORCSC

    using ammonia-water as the working fluid at high concentrations may have marginally

    lower efficiencies than a R245fa ORC with regeneration, but the use of a zeotropic

    mixture in the ORCSC provides the ability to use the temperature glide to match the

    temperature profiles of the source and sink fluids. As shown in Figure 3-13, the zeotropic

    mixture also allows for capacity control simply by changing the concentration of the

    working fluid mixture. This may lead to higher overall system efficiencies when coupled

  • 43

    with sources that have varying heat input temperatures or loads, such as waste heat

    streams from diesel engines, power plants or other industrial processes. The use of a

    Carbon Dioxide / Acetone working pair is difficult to justify from a performance

    standpoint; its main advantage lies in the fact that aside from CO2 being an inexpensive,

    natural, non-flammable refrigerant, its high volumetric heat capacity allows for the use of

    smaller components which may lead to significant weight savings in mobile applications.

    The ORCSC is not without its challenges. The highest efficiencies for the ORCSC

    using Ammonia-Water require significantly low vapor quality in the expander exhaust.

    This is because the desorber outlet state point is modeled as a saturated state, with liquid

    and vapor existing at the same temperature and pressure (with differing concentrations).

    Therefore, since the vapor is at a saturated condition at the expander inlet, the expansion

    process has a tendency to generate significant quality at the outlet condition. This is not

    an issue with the Carbon Dioxide / Acetone working pair because both Carbon Dioxide

    and Acetone are dry working fluids. Therefore, the shape of the vapor dome dictates that

    the expander outlet condition will not fall into the two-phase region. Control related

    issues, which become significantly more complex with the ORCSC compared to the

    traditional ORC, have not been investigated in detail in the literature, and require

    rigorous examination.

    Thus, a truly optimal choice of ORC and working fluid may be highly

    application-specific. Only two working fluid pairs were studied out of a broad range of

    possible binary mixtures. More binary mixture data is needed to study the wide range of

    working fluid combinations that may be possible with the ORCSC. Appropriate system

    selection, therefore, requires a much more thorough investigation of the tradeoffs

  • 44

    between efficiency, ability to match the heat source temperature profile, cost of

    implementation, and size constraints. Investigating these tradeoffs, however, requires

    experimental study. In this regard, an ORCSC experimental test stand was constructed in

    order to provide insight into the feasibility of the technology, and further develop the

    results of this study.

  • 45

    CHAPTER 4. DESIGN OF EXPERIMENTAL TEST SYSTEM

    An experimental test setup (breadboard system) was constructed to serve as a

    proof of concept for ORCSC technology. The system was fabricated largely using off-

    the-shelf parts harvested from the Liquid-Flooded Ericsson Cooler experimental test

    setup (Hugenroth 2006) at the Herrick Laboratories. Since the breadboard system was

    constructed before the computer modeling was complete, several estimates pertaining to

    system design were necessary. The timing for the design and construction of the system

    were largely dictated by needs of the project sponsor. In addition to proof of concept

    testing, the breadboard system was intended to validate the results of the thermodynamic

    cycle model, and to gain design-related experience.

    4.1 System Sizing

    The experimental breadboard system was designed for a nominal 1 kW power

    output. Table 4-1 shows the key input parameters used in the simulation model. The

    values reflect initial expected performance of the equipment, and were used as a basis for

    system design and component selection.

  • 46

    Table 4-1: Sample set of input parameters used as a basis for system design. Symbol Value Description Unit

    CR 5 Circulation Ratio [-] , 0.7 Isentropic turbine efficiency [-] 0.5 Mechanical pump efficiency [-] 80 Temperature level of heat-sink [ C] 20 Temperature level of heat-source [ C] 5 Pinch Point temperature [ C] 0.21 Weak Solution Mass Concentration [-] 0.04 CO2 vapor mass flow rate [kg/s] 1.288 Heat Capacity ratio of the working fluid [-]

    Based on the data in Table 4-1, the conditions at each state point in the cycle were

    calculated as shown in Table 4-2. Note that these initial calculations were performed

    using the LKP EOS and contain significant sources of error as described in Section 3.3.

    However, these were used as the initial design basis and are presented here to illustrate

    the parameters taken into account when constructing the experimental bread board

    system.

    Table 4-2: Sample property values at each state point in the ORCSC cycle using the LKP EOS (refer to Figure 3-3).

    State point T [ C] p [bar] h [kJ/kg] s [kJ/kg/K] [kg/m3]

    1 75.00 29.85 579.0 2.4 49.7 2s 49.41 21.68 560.2 2.4 39.0 2 41.57 21.68 552.