<%BANNER%>

Modeling and Control of a Novel Semi-Closed Gas Turbine-Absorption Combined Cycle

Permanent Link: http://ufdc.ufl.edu/UFE0042893/00001

Material Information

Title: Modeling and Control of a Novel Semi-Closed Gas Turbine-Absorption Combined Cycle
Physical Description: 1 online resource (239 p.)
Language: english
Creator: RYU,CHOON JAE
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: ABSORPTION -- AIRCONDITIONING -- AMMONIA -- DYNAMIC -- ICEMAKING -- MOVINGBOUNDARY -- REFRIGERATION -- WATER
Mechanical and Aerospace Engineering -- Dissertations, Academic -- UF
Genre: Mechanical Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The Power, Water Extraction, and Refrigeration (PoWER) engine has been investigated for several years as a distribute energy (DE) system, among other applications, for civilian or military use. Previous literature describing its modeling and experimental demonstration have indicated several benefits, especially when the underlying semi-closed cycle gas turbine is combined with a vapor absorption refrigeration system (VARS), the PoWER system described herein. The benefits include increased efficiency, high part-power efficiency, small lapse rate, compactness, decreased emissions, reduced air and exhaust flows (which decrease filtration and duct size) and condensation of fresh water. In this research the preliminary design and modeling of a modified version of this system as applied to DE is the focus, a system especially useful in regions prone to major grid interruptions due to hurricanes, undercapacity, or terrorism. In such cases, the DE system should support most or all services within an isolated service island, including ice production, so that the influence of the power outage is limited in scope. This research also describes the rather straightforward system modifications necessary for ice production and the use of this ice-making capacity to achieve significant load-leveling during the summer utility peak, hence reducing the electrical capacity requirements for the grid Load-leveling strategies are also discussed. For this complex system, maximum efficiency and safe operation at any instant of time is the ultimate goal of an advanced control algorithm. To develop such advanced system controls, dynamic modeling is necessary. As part of this research, thermodynamic properties of ammonia-water mixtures and their behavior in a VARS are improved and applied to dynamic modeling of a VARS unit. A conventional moving boundary dynamic heat exchanger model is extended for two-component two-phase flow which exists in the VARS.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by CHOON JAE RYU.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Lear, William E.
Local: Co-adviser: Crisalle, Oscar D.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0042893:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042893/00001

Material Information

Title: Modeling and Control of a Novel Semi-Closed Gas Turbine-Absorption Combined Cycle
Physical Description: 1 online resource (239 p.)
Language: english
Creator: RYU,CHOON JAE
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: ABSORPTION -- AIRCONDITIONING -- AMMONIA -- DYNAMIC -- ICEMAKING -- MOVINGBOUNDARY -- REFRIGERATION -- WATER
Mechanical and Aerospace Engineering -- Dissertations, Academic -- UF
Genre: Mechanical Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The Power, Water Extraction, and Refrigeration (PoWER) engine has been investigated for several years as a distribute energy (DE) system, among other applications, for civilian or military use. Previous literature describing its modeling and experimental demonstration have indicated several benefits, especially when the underlying semi-closed cycle gas turbine is combined with a vapor absorption refrigeration system (VARS), the PoWER system described herein. The benefits include increased efficiency, high part-power efficiency, small lapse rate, compactness, decreased emissions, reduced air and exhaust flows (which decrease filtration and duct size) and condensation of fresh water. In this research the preliminary design and modeling of a modified version of this system as applied to DE is the focus, a system especially useful in regions prone to major grid interruptions due to hurricanes, undercapacity, or terrorism. In such cases, the DE system should support most or all services within an isolated service island, including ice production, so that the influence of the power outage is limited in scope. This research also describes the rather straightforward system modifications necessary for ice production and the use of this ice-making capacity to achieve significant load-leveling during the summer utility peak, hence reducing the electrical capacity requirements for the grid Load-leveling strategies are also discussed. For this complex system, maximum efficiency and safe operation at any instant of time is the ultimate goal of an advanced control algorithm. To develop such advanced system controls, dynamic modeling is necessary. As part of this research, thermodynamic properties of ammonia-water mixtures and their behavior in a VARS are improved and applied to dynamic modeling of a VARS unit. A conventional moving boundary dynamic heat exchanger model is extended for two-component two-phase flow which exists in the VARS.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by CHOON JAE RYU.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Lear, William E.
Local: Co-adviser: Crisalle, Oscar D.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0042893:00001


This item has the following downloads:


Full Text

PAGE 1

1 MODELING AND CONTROL OF A NOVEL SEMI CLOSED GAS TURBINE ABSORPTION COMBINED CYCLE By CHOON JAE RYU A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FO R THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011

PAGE 2

2 2011 Choon Jae Ryu

PAGE 3

3 To my parents, Jiyeon, and Hahnerl

PAGE 4

4 ACKNOWLEDGMENTS I acknowledge the support and guidance of my doctoral advisor, Dr. William E. Lear Jr. He was helpful throughout my time at the University of Florida, encouraging me to think in all different directions and study challenging research topics with the right path. I would like to thank my other doctoral committee members for their time and patience spent wit h me. I appreciate the dynamic and control engineering knowledge Dr. Oscar Crisalle has given me. In addition, I must thank Dr. Sherif, Dr. Ingley, and Dr. Svoronos for their willingness to participate in my doctoral review process. I thank many colleagues working with me in Dr. Lear s lab. Dr. Khan and Dr. Singh taught me many important lab procedures during my first days in the lab. David was always helpful to work together in the lab and he was a good English teacher. Minki was supportive during my last days in the lab. Additionally, I appreciate a friendship of long standing with Cheng Chan and Sydni in fuel cell lab. I also thank my family for their support and understanding. My parents always encourage me with their constant trust. My brother and sister in law have always helped me. I must thank my parents in law for their support to my wife and me during out stay in USA. I really appreciate patient understanding of my dear wife, Jiyeon Lee. She has always inspired me and been helpful. She is also my good discussion partner. I thank her for her constant love and support. Lastly, I thank my precious and adorable baby, Hahnerl.

PAGE 5

5 TABLE OF CONTENTS ACKNOWLEDGMENTS .................................................................................................. 4 page LIST OF TABLES ............................................................................................................ 8 LIST OF FIGURES .......................................................................................................... 9 NOMENCLATURE ........................................................................................................ 13 ABSTRACT ................................................................................................................... 19 CHAPTER 1 INTRODUCTION .................................................................................................... 21 Semi Closed Cycle System .................................................................................... 21 Vapor Absorption Refrigeration System .................................................................. 22 Statement of the Research Problem ....................................................................... 23 Thermodynamic Properties of AmmoniaWater Mixtures ................................. 23 Process Models ................................................................................................ 24 Organization of This Document .............................................................................. 24 2 LITERATURE REVIEW .......................................................................................... 26 Thermodynamic Properties of AmmoniaWater Mixtures ........................................ 26 Process Models ...................................................................................................... 28 3 THERMODYNAMIC PROPERTIES OF AMMONIA WAT ER MIXTURES .............. 30 Gibbs Free Energy for AmmoniaWater Mixture ..................................................... 30 Gibbs Free Energy for Pure Components ........................................................ 30 Vapor and Liquid Gibbs Free Energy for the Binary Mixtures ........................... 32 Pressure Temperature Composition Properties of the Binary Mixtures .................. 33 Results and Discussion ........................................................................................... 35 Comparison with Gillespies Data ..................................................................... 35 Comparison with IGT Experimental Data ......................................................... 36 Saturated liqui d phase ............................................................................... 36 Saturated vapor phase ............................................................................... 36 Comparison with IGT Smoothed Data .............................................................. 37 Temperature Discrepancies for Pure Components .......................................... 38 Summary ................................................................................................................ 39 4 SYSTEM DESIGN OF A REFRIGERATION SYSTEM ........................................... 48 Second Law Analysis Using Constant Internal Temperatures ................................ 48

PAGE 6

6 Thermodynamic Model ..................................................................................... 48 Solution Method ............................................................................................... 51 Results and Discussion .................................................................................... 53 Summary .......................................................................................................... 56 Second Law Analysis Using Variable External Temperatures ................................ 56 Thermodynamic Model ..................................................................................... 57 Solution Method ............................................................................................... 62 Results and Discussion .................................................................................... 64 Summary .......................................................................................................... 67 5 DYNAMIC MODELING WITH CONSTANT AMMONIA MASS F RACTION APPROACH ............................................................................................................ 89 Thermodynamic Model ........................................................................................... 89 Heat Recovery Vapor Generator ...................................................................... 91 Rectifier ............................................................................................................ 91 Condenser, Absorber, and Solution Heat Exch anger ....................................... 92 Evaporator ........................................................................................................ 92 Pump ................................................................................................................ 9 3 Thermal Expansion Valve ................................................................................. 93 Refrigerant Heat Exchangers ........................................................................... 93 Ice Maker and Storage ..................................................................................... 96 Solution Method ...................................................................................................... 96 Results and Discussion ........................................................................................... 98 Consistency of First Law and Second Law Models ............................................... 100 Summary .............................................................................................................. 101 6 DYNAMIC MODELING FOR TWO COMPONENT FLUIDS ................................. 113 Introduction to Dynamic Heat Exchanger Modeling .............................................. 113 Movin g Boundary Model for Single Component Fluid .................................... 113 Two C omponent F luids .................................................................................. 117 Two Phase Model Options ............................................................................. 118 Thermodynamic M odels of VARS for T wo C omponent F luids .............................. 120 Condenser ...................................................................................................... 121 Absorber ......................................................................................................... 128 Generator ....................................................................................................... 133 Evaporator ...................................................................................................... 137 Solution H eat E xchanger ................................................................................ 140 Results and Discussion ......................................................................................... 142 Condenser ...................................................................................................... 143 Absorber ......................................................................................................... 144 Generator ....................................................................................................... 145 Evaporator ...................................................................................................... 145 Solution Heat Exchanger ................................................................................ 146 Summary .............................................................................................................. 147

PAGE 7

7 7 CONCLUSIONS ................................................................................................... 162 Thermodynamic Properties of AmmoniaWater Mixtures ...................................... 162 Preliminary Design of VARS ................................................................................. 163 Dynamic Modeling of Heat Exchangers ................................................................ 164 Future Work .......................................................................................................... 165 APPENDIX A COMPUTER CODE FOR THEMODYNAMIC PROPERTIES OF AMMONIA WATER MIXTURES ............................................................................................. 167 B COMPUTER CODE FOR DYNAMIC MODELING WITH CONSTANT AMMONIA MASS FRACTION APPROACH ......................................................... 175 C GOVERNING EQUATIONS AND COMPUTER CODE FOR MOVING BOUNDARY MODEL ............................................................................................ 178 Governing Equations for Moving Boundary Model ................................................ 178 Derivations and Coefficients of Single Component Evaporator Model ........... 178 Coefficients of Two Component Condenser Model ........................................ 183 Coefficients of Two Component Absorber Model ........................................... 185 Coefficie nts of Two Component Generator Model ......................................... 186 Derivations and Coefficients of Two Component Evaporator Model .............. 188 Coefficients of Two Component SHX Model .................................................. 189 Computer Code for Moving Boundary Model ........................................................ 190 Condenser ...................................................................................................... 191 Absorber ......................................................................................................... 200 Generator ....................................................................................................... 210 Evaporator ...................................................................................................... 219 Solution Heat Exchanger ................................................................................ 227 LIST OF REFERENCES ............................................................................................. 235 BIOGRAPHICAL SKETCH .......................................................................................... 239

PAGE 8

8 LIST OF TABLES Table page 3 1 Coefficients for equations (31) and (32) for pure water and ammonia ............. 40 3 2 Coefficient for Gibbs excess energy function ...................................................... 40 3 3 Comparison of dew point te mperatures .............................................................. 40 4 1 Input/Output parameters of the HPRTE system. ................................................ 69 4 2 Comparison of results ......................................................................................... 69 4 3 Input parameters of the VARS ............................................................................ 70 4 4 Input/output parameters of the HPRTE system. ................................................. 70 4 5 Ice produced a day [Ton/day (L/day)] ................................................................ 70 5 1 Design values or states at each state point ...................................................... 103 5 2 Results for higher air condit ioning load ............................................................. 104 5 3 Results for lower air conditioning load .............................................................. 105 5 4 Consistency of the results between Chapters 4 and 5. ..................................... 106 6 1 Parameters of the condenser ........................................................................... 148 6 2 Parameters of the absorber .............................................................................. 148 6 3 Parameters of the generator ............................................................................. 148 6 4 Parameters of the evaporator ........................................................................... 149 6 5 Parameters of SHX ........................................................................................... 149 A 1 Matlab files for thermodynamic properties of ammoniawater mixtures ............ 167 B 1 Matlab and Simulink files for dynamic modeling with constant amm onia mass fraction approach. ............................................................................................. 175

PAGE 9

9 LIST OF FIGURES Figure page 3 1 Comparison P T x relations with Gillespies data at 313.15K ............................. 41 3 2 Comparison P T x relations with Gillespies data at 405.9K ............................... 41 3 3 Comparison for saturated liquid enthalpy ........................................................... 42 3 4 Comparison with IGT data for saturated vapor temperature at 1654.74kPa ....... 42 3 5 Comparison with IGT data for saturated vapor enthalpy at 1654. 74kPa ............ 43 3 6 Comparison with IGT data for saturated vapor and liquid temperature .............. 43 3 7 Comparison with IGT data for saturated vapor and l iquid enthalpy .................... 44 3 8 Saturated temperatures for high ammonia mass fraction at 344.74kPa ............. 44 3 9 Saturated temperatures for hig h ammonia mass fraction at 1034.21kPa ........... 45 3 10 Saturated temperatures for high ammonia mass fraction at 1723.69kPa ........... 45 3 11 Temperature discrepancies at pure components ................................................ 46 3 12 Discrepancy and oscillation of El Sayed and Tribuss P T x .............................. 46 3 13 Discrepanc y and oscillation of Ptek and Klomfar s P T x ................................. 47 4 1 Gas turbine flowpath of the PoWER system ....................................................... 71 4 2 The multi stage vapor absorption refrigeration system ....................................... 71 4 3 The multi stage simple vapor absorption refrigeration system ........................... 72 4 4 Air conditioner hourl y load for typical summer day ............................................. 72 4 5 Typical variations of outdoor air temperature at 42 north latitude on July 1 ...... 73 4 6 Carn ot COP variations for the three evaporators ................................................ 73 4 7 COP variations for three evaporators at R=0.5 .................................................. 74 4 8 COP variations for three evaporators at R=0.7 .................................................. 74 4 9 Total COP variations of the system at R=0.5 ..................................................... 75 4 10 Total COP variations of the system at R=0.7 ..................................................... 75

PAGE 10

10 4 11 Heat removed from surroundings at R=0.5, Revp=0.3 ......................................... 76 4 12 Allocated heat supply from generator at R=0.5, Revp=0.3 ................................... 76 4 13 Heat removed from surroundings at R=0.7, Revp=0.3 ......................................... 77 4 14 Allocated heat supply from generator at R=0.7, Revp=0.3 ................................... 77 4 15 Heat removed from surroundings at R=0.5, Revp=0.5 ......................................... 78 4 16 Heat removed from surroundings at R=0.7, Revp=0.5 ......................................... 78 4 17 Allocated heat supply from generator at R=0.5, Revp=0.5 ................................... 79 4 18 Allocated heat supply from generator at R=0.7, Revp=0.5 ................................... 79 4 19 Heat removed from surroundings at R=0.5, Revp=0.7 ......................................... 80 4 20 Heat removed from surroundings at R=0.7, Revp=0.7 ......................................... 80 4 21 Allocated heat supply from generator at R=0.5, Revp=0.7 ................................... 8 1 4 22 Allocated heat supply from generator at R=0.7, Revp=0.7 ................................... 81 4 23 Ice produced hourly for various Revp at R=0.5 ..................................................... 82 4 24 Ice produced hourly for various Revp at R=0.7 ..................................................... 82 4 25 The multi temperature simple vapor absorption refrigeration system ................. 83 4 26 Four cases of external temperatures .................................................................. 83 4 27 COP for a small control volume when A) generator t emperature varies and B) Evaporator 1 temperature varies ........................................................................ 83 4 28 Carnot COP for Evaporator 1 ............................................................................. 84 4 29 Carnot COP for Evaporator 2 ............................................................................. 84 4 30 Carnot COP for Evaporator 3 ............................................................................. 85 4 31 Heat removed at Evaporator 2 ............................................................................ 85 4 32 Heat removed at Evaporator 3 ............................................................................ 86 4 33 Heat required for Subsystem 1 .......................................................................... 86 4 34 Heat required for Sub system 2 .......................................................................... 87 4 35 Heat required for Sub system 3 .......................................................................... 87

PAGE 11

11 4 36 Efficiency ratio based on Case (a) efficiency ...................................................... 88 5 1 Extended vapor absor ption refrigeration system of the PoWER system .......... 107 5 2 Simple shell and tube heat exchanger .............................................................. 107 5 3 Lumped wall model ........................................................................................... 108 5 4 Control volumes for A) RHX1 and B) RHX2 ..................................................... 108 5 5 Extended VARS Simulink model ...................................................................... 108 5 6 Air conditioning hourly load .............................................................................. 109 5 7 Mass change of the stored ice [kg] ................................................................... 109 5 8 Chilled water use [kg/ s] .................................................................................... 110 5 9 Heat gained from surroundings at three evaporators ....................................... 110 5 10 Coefficient of performance (COP) .................................................................... 111 5 11 Temperature of RHX1 ...................................................................................... 111 5 12 Temperature of RHX2 ...................................................................................... 112 6 1 A schematic of an evaporator model for single component flow ....................... 150 6 2 Two phase flow t emperature distribution in an evaporator ............................... 150 6 3 Saturated vapor a mmonia mass fraction in an evaporator ............................... 151 6 4 Saturated liquid ammonia mass fraction in an evaporator ................................ 151 6 5 Enthalpy Ammonia mass fraction ( h x ) diagram ............................................... 152 6 6 A s chematic of a condenser ............................................................................. 152 6 7 A s chematic of an absorber .............................................................................. 153 6 8 A schematic of a generator ............................................................................... 153 6 9 A s chematic of an evaporator ........................................................................... 154 6 1 0 A schematic of SHX .......................................................................................... 154 6 11 Inlet ammonia mass fraction step increase for the condenser .......................... 155 6 12 Inlet mass flow rate step increase for the condenser ........................................ 155

PAGE 12

12 6 13 Secondary fluid inlet mass flow rate step increase for the condenser .............. 156 6 14 Inlet ammonia mass fraction step increase for the absorber ............................ 156 6 15 Inlet enthalpy step increase for the absorber .................................................... 157 6 16 Inlet mass flow r ate step increase for the absorber .......................................... 157 6 17 Secondary fluid inlet mass flow rate step increase for the absorber ................. 158 6 18 Inlet ammonia mass fraction step increase for the generator ........................... 158 6 19 Inlet mass flow rate step increase for the generator ......................................... 159 6 20 Inlet ammonia mass fraction step increase for the evaporator ......................... 159 6 21 Inlet enthalpy step increase for the evaporator ................................................. 160 6 22 Inlet mass flow rate step increase for the evaporator ....................................... 160 6 23 Colder side inlet ammonia mass fraction step increase for SHX ...................... 161 6 24 Colder side inlet mass flow rate step increase for SHX .................................... 161 B 1 Simulink model for Chapter 5. .......................................................................... 177 C 1 Condenser Simuli nk model ............................................................................... 191

PAGE 13

13 NOMENCLATURE A Area [ m2] pC Specific heat capacity at constant pressure [ kJ/kg K ] carnotCOP Carnot COP D D iameter of tube [ m ] G Specific Gibbs free energy [kJ/kg] H Heat transfer coefficient between tube wall and fluid [ kW/ m2] L Heat exchanger length [ m ] *L Imaginar y moving boundary position [ m ] P Pressure [ kPa ] Q Heat transfer rate [ kW ] Q Heat transfer rate per area [ kW/m2] R Gas constant [kJ/kmole K] evpR Refrigeration ratio T Temperature [ K ] or T ranspose of matrix UA Area universal heat transfer coefficient [ kW/K ] UD Length universal heat transfer coefficient [ kW/K m ] V Volume [ m3] pW Pump work [ kW ] c Concentration d Vapor quality or dryness f Body force [ N ] h Spec ific enthalpy [ kJ/kg ] l Interface position or length [ m ]

PAGE 14

14 m Mass [ kg ] m Mass flow rate [ kg/s ] latentq Latent heat of fusion of water [ kJ/kg ] s Specific entropy [kJ/kg K] t Time [ sec ] u Fluid velocity [ m/s ] u Input matrix v Specific volume [m3/kg] x Total ammo nia mass fraction or Ammonia mass fraction in liquid x State variable matrix y Ammonia mass fraction in vapor y Output matrix z Length [ m ] Greek Stress tensor [ N / m2] i Heat transfer coefficient bet ween tube wall and refrigerant [ kW/ m2] o Heat transfer coefficient between tube wall and ambient [ kW/ m2] Density [ kg/ m3] p Pump efficiency R Second law efficiency th Thermal efficiency Characteristic time [ sec ] or Shear stress tensor [ N / m2]

PAGE 15

15 Superscript s Non dimensional variables or imaginary moving boundary L Liquid phase G Gas phase E Excess energy Subscripts AC Air condit ioning E Evaporator ICE Ice storage R Refrigeration a Ambient, A bsorber or A mmonia 12 a Interface in absorber aa Ambient for absorber ai Absorber inlet ao Absorber outlet ar Refrigerant in absorber avg Length average b Bubble point 12 c Interface in condenser c Condenser or Carnot ca Ambient for condenser ci Condenser inlet co Condenser outlet cr Refrigerant in condenser

PAGE 16

16 cw Critical point for water or Chilled water d Dew point e Equilibrium or E vaporator ea Ambient for evaporator eg Saturated vapor for thermodynamic equilibrium ei Evaporator inlet el Saturated liquid for thermodynamic equilibrium eo Evaporator outlet er Refrigerant in evaporator g Gas phase or Generator 12 g Interface in generator ga Ambient for generator gi Generator inlet go Genera tor outlet gr Refrigerant in generator i Inlet int Interface between two phase and singlephase m Mixture o Reference values, Out let, or Ambient p Pump r Refrigerant or Reduced values rc Colder side in RHX rci Colder side inlet in RHX rco Colder side outlet in RHX

PAGE 17

17 rh Hotter side in RHX rhi Hotter side inlet in RHX rho Hotter side outlet in RHX s SHX or Steady state sc Colder side in SHX sci Colder side inlet in SHX sco Colder side outlet in SHX sg Saturated vapor sh Hotter side in SHX shi Hotter side inlet in SHX sho Hotter side outlet in SHX sl Saturated liquid w Wall or W ater Abbreviations ABS Absorber AC Air Condition ing CGC Cold Gas Cooler CM Combustor COND Condenser COP Coefficient of Perform ance CW Chilled Water DE Distributed Energy EVAP Evaporator GP Gas Path

PAGE 18

18 HGC Hot Gas Cooler HPC High Pressure Compressor HPRTE High Pressure Regenerative Turbine Engine HPT High Pressure Turbine HRVG Heat Recovery Vapor Generator h x Enthalpy Ammonia Mass F raction IGT Institute of Gas Technology LPC Low Pressure Compressor LPT Low Pressure Turbine ODE Ordinary Differential Equation P Pump PDE Partial Differential Equation PoWER Power, Water Extraction, and Refrigeration P T x Pressure Temperature Composition RECT Rectifier RHX Refrigerant Heat Exchanger SHX Solution Heat Exchanger SOFC Solid Oxide Fuel Cell TXV Thermal Expansion Valve VARS Vapor Absorption Refrigeration System WGC Warm Gas Cooler

PAGE 19

19 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 MODELING AND CONTROL OF A NOVEL SEMI CLOSED GAS TURBINE ABSORPTION COMBINED CYCLE By Choon Jae Ryu M ay 2011 Chair: William E. Le ar Jr. Cochair: Oscar Crisalle Major: Mechanical Engineering The Power, Water Extraction, and Refrigeration (PoWER) engine has been investigated for several years as a distribute energy (DE) system, among other applications, for civilian or military use. Previous literature describing its modeling and experimental demonstration have indicated several benefits, especially when the underlying semi closed cycle gas turbine is combined with a vapor absorption refrigeration system (VARS), the PoWER system descr ibed herein. The benefits include increased efficiency, high part power efficiency, small lapse rate, compactness, decreased emission s, reduced air and exhaust flows (which decrease filtration and duct size) and condensation of fresh water. In this resear ch the preliminary design and modeling of a modified version of this system as applied to DE is the focus a system es pecially useful in regions prone to major grid interru ptions due to hurricanes, undercapacity, or terrorism. In such cases, the DE system should support most or all services within an isolated service island, including ice production, so that the influence of the power outage is limited in scope. This research also describes the rather straightforward system modifications necessary

PAGE 20

20 for ice production and the use of this icemaking capacity to achieve significant loadleveling during the summer utility peak, hence reducing the electrical capacity require m e n ts for the grid L oad le veling strategies are also discussed. For this complex system m aximum efficiency and safe operation at any instant of time is the ultimate goal of an advanced control algorithm To develop such advanced system control s, dynamic modeling is necessary. As part of this research, thermodynamic properties of ammoniawater mixtures and their behavior in a VARS are improved and applied to dynamic modeling of a VARS unit. A c onventional m oving boundary dynamic heat exchanger model is extended for two component two phase flow which exists in the VARS

PAGE 21

21 CHAPTER 1 INTRODUCTION Semi Closed Cycle System Currently most countries generate their electricity in large centralized facilities such as coal, nuclear, hydropower or gas powered plants. These plants have excellent economies of scale, but usually transmit electricity long di stances and /or are often considered to be too far away for their waste heat to be used effectively To avoid these problems, construction of intermediatesize district heating plants near a city is one possible solution, often employed in northern European countries. Distributed generation is another popular approach in which small distributed energy resources such as wind turbines, fuel cell s, solar cell s, reciprocating engines and microturbines installed near loads can produce electricity without signi ficant transmission los s. Some of these can also take advantage of waste heat to meet local heating requirements Of the various distributed energy systems, gas turbines are more reliable and require less maintenance than others. They are compact and produce low emissions and high quality heat but are expensive and the efficienc ies of small gas turbines are typically lower than some other options C ombined cycle power generation and/or cogeneration are used in larger plants to increase efficiency. The waste heat can drive steam turbines, absorption chillers, or other cycles. The current paper deals with a novel type of semi closed cycle gas turbine engine which inherently favors coupling to bottoming cycles or to cogeneration. Anxionnaz [ 1 2 ] proposed semi closed gas turbine cycles due to their potential advantages in low fuel consumption over their power range and significantly reduced airflow requirements. Lear and Sherif [ 3 ] patented the concept of integrating a semi -

PAGE 22

22 closed gas turbine system, called High Pressure Regenerative Turbine Engine (HPRTE), with a Vapor Absorption Refrigeration System (VARS) to generate the Power, Water Extraction, and Refrigeration (PoWER) system This system is the focus of the current research study. In order to understand thi s complex system and control it optimally accurate process models for each component must be developed and physical properties of working fluids must be well understood. After the dynamic behavior of the coupled VARS and HPRTE is properly modeled, various control logic can be applied to the integrated system. Vapor Absorption Refrigeration System VARS units using ammoniawater as the working fluid were popular and widely used before the development of the vapor compression refrigeration system, which has a higher coefficient of performance (COP) than the VARS. The VARS however, can use low grade waste heat energy rather than electricity, so the lower COP is unsurprising. Solar or geothermal energy or waste heat from power plants is of sufficient ly high tem perature to drive a VARS unit. Various refrigerant absorbent pairs can be used for the VARS refrigeration system, depending on the temperatures to be controlled. Two common working fluids for the VARS are lithium bromide water (LiBr H2O) and ammoniawater (NH3H20) mixtures. Lithium bromide water refrigeration systems use water as their refrigerant. Therefore, these systems cannot be used below freezing, but are suitable for air conditioning or other systems operated above freezing temperatures. For ammonia water refrigeration system, ammonia is the refrigerant and water is the absorbent. Ammonia is an excellent refrigerant not only for air conditioning but also for freezing. Since the absorbent, water,

PAGE 23

23 is volatile, distillation columns need to be included i n the cycle in order to purify the ammonia sufficiently to attain high efficiency. Even though the ammoniawater VARS has been under development for over 100 years, there is still the potential to improve performance. Better understanding of the ammoniawa ter mixture behavior and a thermodynamically consistent property model would significantly improve the accuracy and robustness of process models. These mathematical models are necessary in order to create an advanced controller for the multivariate system. Statement of the Research Problem In this research, dynamic modeling of a VARS and the control the system, thermodynamic properties of ammoniawater mixtures, modeling methods and control scheme s have been explored. All aspects of the research are in sup port of the development of advanced physics based control for the PoWER integrated system, including consideration of multiple applications with timevarying objective functions Thermodynamic Properties of Ammonia Water Mixtures There are various mathemat ical models for ammoniawater mixtures in common use for research or design purposes Fugacity and chemical potential based methods combined with the energy equation, have been used to obtain pressuretemperaturecomposition (P T x) relationships, which are the bubble and dew point temperatures. Simpler curvefits for P T x are also widely used to save computing time, suitable for design. However, fugacity and chemical potential based methods need timeconsuming iteration and curvefit models using polynomial equations have oscillatory behavior so that they are t hermodynamically inconsistent in some r egimes

PAGE 24

24 For dynamic analysis or accurate iteration in the solution of ammonia water system s, a faster and smoother property model is necessary. A set of tabulat ed data with an interpolation method for P T x are chosen to overcome the problems from the abovementioned relationships. Process Models Since terrestrial gas turbine systems gained widespread use, utilizing waste heat has been an important option for inc reasing the system efficiency. S tea m turbines, absorption chillers or cogeneration can be integrated, depending on the local demands. Absorption chillers are appropriate when the ambient temperature is always high enough to run air conditioners or for the food processing or storage industry A semi closed microturbine system is combined with a double expansion VARS, which has three absorbers for cooling hot air or gas and making ice. T h is system, however, has not yet been studied in the literature. In this research, the VARS unit has been conceptually modeled at multiple levels and analyzed. Thermodynamic models based on the First Law and on the Second Law have been developed in this work. To model twophase two component flow in heat exchangers, the conventional moving boundary heat exchanger model has been exten ded to enforce species balances and phase equilibrium Organization of This Document The literature for the relevant research areas is reviewed in Chapter 2. In Chapter 3, a study of thermodynamic properties of ammoniawater mixtures is presented, in which existing commonly used mathematical models and an improved interpolation approach are introduced and compared with experimental data.

PAGE 25

25 Chapter s 4 to 6 address thermodynamic modeling of semi closed c ycle systems including both the VARS and HPRTE A second law VARS model is first discussed in Chapter 4, and dynamic modeling of single effect systems for constant and variable ammonia mass fraction are presented in Chapter s 5 and 6, respectively The over all results are summarized in Chapter 7 including recommendations for future research

PAGE 26

26 CHAPTER 2 LITERATURE REVIEW This section discusses the literature on thermodynamic properties of ammoniawater mixtures, process models. Thermodynamic Properties of Amm onia Water Mixtures Ammoniawater mixtures are considered as a highly effective working fluid for vapor absorption refrigeration operations, which take advantage of waste heat, and for other energy applications such as Kalina cycle power generation plants [ 4 ] which can be driven by low temperature heat sources. To design these systems for highefficiency operation, it is important to have access to the thermodynamic properties of vapor liquid mixtures of ammonia and water. In principle, it is of interest to ensure that these variables are thermodynamically and mathematically consistent Furthermore, for system design purposes it is also important to ensure that the mathematical models providing the physical data are easily programmable for numerical execution in a computing platform, and also that they are timeefficient [ 5 ] in terms of execution time. The literature contains reports of ammoniawater mixture presented in terms of fundamental thermodynamic models, curvefitting approximations, and tabulated da ta. Of particular interest is the work of Tillner Roth et al. [ 6 ] who present a comprehensive set that includes most of the experimental data available. They include the data of Gillespie et al. [ 7 ], which has been used extensively by others to develop th ermodynamic modeling equations because the set includes a very wide range of temperatures and compositions. Macriss et al. [ 8 ] present data measured at high ammonia concentrations.

PAGE 27

27 Park et al. [ 9 ] present a thermodynamic model for the ammonia water binary system based on equations of state for the Helmholtz free energy. This approach is capable of predicting mixture properties up to 20MPa and 650K, a temperature and pressure range where experiments are particularly challenging to carry out. Another proposition based on the Helmholtz free energy is presented by Tillner Roth et al. [1 0 ], a work that includes a large set of experimental data [ 6 ]. These authors use density as an input variable rather tha n the more common choice of using pressure, hence avoidi ng the need to include in the model separate equations for the pure components and for the mixture. Very accurate density data, therefore, are necessary for thermodynamic consistency. In addition to the Helmholtz free energy, various equations of state hav e been developed. For example, thermodynamic models based on the Gibbs free energy are also frequently encountered. Bubble and dew point temperature equations and P T x equations, developed as polynomial functions of pressure and ammonia concentration, can be found in El Sayed et al. [1 1 ] and Ptek et al. [ 7 ]. Using such polynomial P T x equations, iteration processes can be removed to get equilibrium temperature. Even though thermodynamically inconsistent, a set of five equations provided by Ptek et al. i s easy to program for modeling a VARS, which operate in a pressure range from 0.2 to 2MPa. Tabulated P T x data are also available. Bogart [ 1 2 ] presents various mixture data and tabulated ammonia mass fraction and enthalpy data for saturated liquid and vapor according to temperature up to a pressure of approximately 2MPa. Using polynomial P T x equations rather than numerical iterations for phase equilibrium calculations by the fugacity or chemical potential equality method, is more convenient for determining

PAGE 28

28 mixture properties such as temperature, dryness, and physical state (superheated, subcooled, and equilibrium). Unlike any other tabulated data, Bogart provided temperature data for very high ammonia compositions. Xu et al. [ 1 3 ] and Tamm [ 1 4 ] combined El Sayeds P T x relations with the equations of state for the Gibbs free energy developed by Ibrahim et al. [ 1 5 ]. Ryu et al. [1 6 ] used Pteks bubble and dew point temperatures with Ibrahims enthalpy and density relations for their dynamic simulation of PoWER system, a highly efficient semi closed gas turbine system [ 3 ]. Process Models Lear and Sherif [ 3 ] patented a combined cooling and power cycle, which produces power, refrigeration, and water extraction, in addition to heat. Khan [ 1 7 ] studied several cases of operation modes and optimized the system. His steady state HPRTE model is used for this research. Adewusi et al. [ 1 8 ] compared a singlestage VARS with a doublestage VARS by second law based thermodynamic analysis. The entropy generation of each component and the total entropy generation of all the system components as well as the COPs of the VARSs are calculated from the thermodynamic properties of the working fluids at various operating conditions. A lumped parameter model of VARS was establis hed by Kim and Park [ 19]. The dynamic twophase thermal hydraulic characteristics are investigated during its start up operation period, which is most challenging to understand and model. In addition to their literature, MacArthur [2 0 ] gave clear method of discretized dynamic heat exchanger model for compressible twophase flow.

PAGE 29

29 He et al. [2 1 ] and Jensen et al. [22 ] modeled vapor compression refrigeration systems using moving boundary heat exchanger model to meet reasonable two phase physics and control app lication at once. They used Wedekind s [2 3 ] time invariant mean void fraction idea for twophase dynamics and Zivi s [2 4 ] void fraction model.

PAGE 30

30 CHAPTER 3 THERMODYNAMIC PROPERTIES OF AMMONIA WATER MIXTURES In this chapter, we show that the polynomial equations of El Sayed et al. and Ptek et al. and a nonmonotonic behavior at near pure compositions, an effect that becomes particularly pronounced at near pureammonia compositions and caused by the nature of polynomial equations (R unge s phenomenon) rather than as a consequence of the thermodynamics. In addition, frequently used P T x relationships and Bogarts P T x tabulated data are compared and used with Ibrahims equations, and the best method for developing controller using memory chips for VARSs and less mathematical error will be discussed. The data in Tillner Roth et al. [ 6 ] and Gillespie et al. [ 7 ] is used as a source of comparison in this work Gibbs Free Energy for Ammonia Water Mixture Schulz [ 26] made two separate equations for the liquid and t he vapor phases of the ammoniawater mixtures. His model is limited to a maximum pressure of 2.5MPa, which is extended by Ziegler et al. [ 2 6 ] to 5MPa and 500K. Ibrahim and Klein [15 ] modified the coefficients for the liquid Gibbs excess energy to include t he experimental data reported by Gillespie et al. [ 7 ]. By doing so, their correlations cover vapor liquid equilibrium pressures of 11MPa and temperature of 600K Gibbs Free Energy for Pure Components The equations of Gibbs free energy for the pure ammonia or pure water are as shown below for the liquid and vapor states Liquid phase:

PAGE 31

31 2233 3 2 ,,1, 22 3 12, 2 22 2 134 ,23 ln 2 2LLL rrorrorro rro rro r r rrrorrro ro rrrro rroB B GhTsBTTTTTT B T BTBTTTTTT T A AATATPPPP ( 3 1) Gas Phase: 2233 3 2 ,,1, 22 3 12, 1,2 334 ,, 3, 1111 ,23 ln 2 ln 43 1211GGG rrorrorro rro rro r r rrro rrro ro ro r rr r rro ro ro rr ro ro rr ro rroD D GhTsDTTTTTT D T DTDTTTTTT T P PPT TCPPC P P TTT P PT CP TTT 3 3 3 4 12 1111 12 ,,1211 3ro rr ro ro rro roP CP T P TTT ( 3 2) where the reduced thermodynamic properties are defined as follows: r BT T T ( 3 3) r B P P P ( 3 4) r BG G RT ( 3 5) The reference values for the reduced properties are R = 8.314 kJ/kmoleK TB = 100 K and PB = 10 bar Table 3. 1 reports the numerical values of the coefficients appearing in equations ( 3 1 ) and ( 3 2) for the cases of pure ammonia and pure water.

PAGE 32

32 The Gibbs f ree e nergy equations ( 3 1) ( 3 5) can in turn be used to calculate the molar specific enthalpy, entropy, and volume of the pure components in terms of reduced variables thr ough the following suite of equations 2rr Br rr PG hRTT TT ( 3 6) rr r PG sR T ( 3 7) rBr Br TRTG v PP ( 3 8) Vapor a n d Liquid Gibbs Free Energy for t he Binary Mixtures Vapor phase ammonia water mixtures are assumed to be ideal solutions. Hence, the following thermodynamic properties can be obtained through simple relationships involving only the gas composition of am m onia : 1GGG mawhyhyh ( 3 9) 1GG Gmix mawsysyss ( 3 10) ln1ln1mixsRxxxx ( 3 11) 1GGG mawvyvyv ( 3 12) For liquid mixtures, the Gibbs excess energy 2 12 312121E rGxFFxFx ( 3 13) must be included, where

PAGE 33

33 56 11234 2 r rr rrEE FEEPEEPT TT ( 3 14) 1112 278910 2 r rr rrEE FEEPEEPT TT ( 3 1 5 ) 1516 31314 2 r rrEE FEEP TT ( 3 1 6) Then, the liquidmixture thermodynamic properties can be postulated as follows: 1LLLE mawhxhxhh ( 3 17) 1LL LEmix mawsxsxsss ( 3 18) 1LLLE mawvxvxvv ( 3 19) Pressure Temperature Composition Properties of the Binary Mixtu res Pressure temperaturecomposition (P T x) relationships of a vapor liquid equilibrium state for the ammoniawater binary mixture can be obtained by solving a system of thermodynamic equations that imposes equality of pressures, temperatures and componen t fugacities across the two phases. Such approach requires an iterative numerical solution method, as carried out for example by Ibrahim et al. [1 5 ] and Park and Sonntag [ 9 ] The latter reference also discusses an approach for calculating equilibrium tempe ratures u sing flow charts. The calculation of thermodynamic equilibrium properties via numerical methods involves a significant computational cost. Hence, once the calculations are carried out, it is customary to report selected values in a table for futur e access, or alternatively, to approximate the results through a mathematical correlation often of polynomial form.

PAGE 34

34 The user then needs to compute the values of a polynomial expression or carry out an interpolation in a table to obtain specific thermodyna mic values, such as the dew point at a given pressure and mixture composition. Using a curvefitting approach, El Sayed and Tribus developed the following bubble ( Tb) and dew point ( Td) temperature polynomial correlations : 710 11lni j c bc iij ijP TTccx P ( 3 2 0) 64 11ln1.0001lni j c dc iij ijP TTaAy P ( 3 21) where 4 1 i ccwi iTTax ( 3 22) 8 1expi ccw i iPPbx ( 3 23) where P is the mixture pressure, x and y are ammonia mole fraction of saturated liqui d and vapor, respectively Tcw and Pcw are critical temperature and pressure for water, and Tc and Pc are critical temperature and pressure of ammoniawater mixtures. Ptek and Klomfar propose equations for describing the thermodynamic properties of the am monia water binary system with pressure ranges up to 2MPa. A particularly useful equation calculates the vapor composition as a function of the pressure and the liquid composition without requiring an iterative computational method. Even though the vapor enthalpy equation is not available for superheated region and bubble and dew

PAGE 35

35 point temperatures and vapor composition equations are not thermodynamically consistent, the equation set is in widespread use in industry [ 2 7 ] where rapid calculations are needed. T he bubble point and dew point temperature correlations proposed by these authors are the following equations : 14 11lni in m o boi iP TTax P ( 3 24) 17 /4 11lni in m o doi iP TTay P ( 3 25) where To and Po are reference temperature and pressure. Bogart [1 2 ] ta bulated calculated data. Unlike other published data, his data are available at very high ammonia composition as function of temperature at given pressures. The data is a compilation of various experimental results. Obtaining thermodynamic information from Bogarts tables reduces to an exercise in numerical interpolation, for example, using cubic polynomials, which entails a relatively low computational cost. Results and Discussion In this study, three different P T x relations with Ibrahims Gibbs free energy method are compared with most used experimental and smoothed data. Comparison with Gillespie s Data Figure s 3 1 and 3 2 show that Bogart tabulated data with cubic interpolation best follow the experimental data from Gillespie for saturated liquid while Pteks equations are generally better for saturated vapor. This trend is shown for a wide range of temperatures. However, the differences between three P T x relations and Gillespies

PAGE 36

36 experimental data are not significantly large except the El Sayeds at almost pure ammonia compositions. Ptek and Klomfar have their own enthalpy equations for saturated liquid and saturated vapor. Their saturated enthalpy equations with temperature equations show very good agreements at most regions. However, their saturat ed vapor model is not available for the superheated region. So, Ibrahims Gibbs free energy method will be used to get enthalpies of the mixtures for the comparison. Comparison with IGT Experimental Data Macriss et al. [ 8 ] provided both experimental and calculated data in their research bulletin for the Institute of Gas Technology (IGT). Their experimental data are compared with three P T x methods and explained, mainly while their smoothed data are used for figures to compare. Saturated liquid phase For lo w ammonia concentrations (0.0475 ~ 0.383 of ammonia mass fraction, IGT Table C 3 ~ C 7 [ 8 ]), three P T x methods show good agreem ent overall for liquid enthalpy (Figure 3 3 ) Bogarts and Pteks are somewhat better at lower pressure and El Sayeds relations are better at higher pressure (from about 1500kPa). Saturated vapor phase The behaviors of saturated vapor temperatures and enthalpies of three methods show the same trends at high pressure ranges (1378.95 (200), 1654.74 (240), and 2068.43 (300) kPa (ps ia), IGT Table A 1 and C 2 [ 8 ])) for a refrigeration unit and at high ammonia mass fraction regions. The percentage errors that Pteks and El Sayeds temperature equations make at those regions are always lower than those of Bogarts method. In particular Pteks temperatures best follow the IGT experimental data.

PAGE 37

37 Bogarts temperatures with cubic interpolation are about 5 K (approximately 1.5 % error) higher than the IGT temperature data, while El Sayeds temperature drops fast and fluctuates at near pur e ammonia compositions. With Ibrahims saturated vapor enthalpy relation, Pteks, El Sayeds, and Bogarts show good agreement at lower, middle, and higher ammonia compositions, respectively at high ammonia concentratio n The percentage errors are less than approximately 1 percent. Comparison with IGT Smoothed Data IGT smoothed (calculated) data for 344.74kPa, 1034.21kPa, and 1723.69kPa (IGT Table C 8, 9, 11, 13, and 14 [ 8 ]) were compared with three different P T x models. As shown at Figure 3 6 all of th ree models follow well the IGT data. Pteks temperature is somewhat more precise at higher pressure, and Bogarts temperature predictions are close to the experimental data at lower pressure. As a resolution of Figure 3 7 three P T x methods show almost identical enthalpy results. When near pure component regions are inspected in closer detail, two significant problems are exposed. At lower pressure (344.74kPa = 50psia), the vapor temperatures are lower than liquid temperature for both El Sayeds and Pteks correlations, even though it is difficult to visually observe the effect even at the high resolution of Figure 3 8 As the pressure increases, El Sayeds dew point temperature fluctuates and never meets the bubble point temperature in the limit of pur e ammonia. A temperature discrepancy between dew and bubble point temperatures of pure components is also observed in Pteks correlation. At very high ammonia concentration, dew point temperatures become lower than bubble point temperatures, which is phy sically impossible. Bogarts data with interpolation, however, do not show

PAGE 38

38 these problems. The oscillations appear at enthalpy composition F igures 3 9 and 310 also Temperature Discrepancies for Pure Components Figure 3 11 shows temperature differences b etween dew and bubble point temperatures at pure water and pure ammonia. Generally, the higher pressure the more discrepancy Although at very high ammonia concentrations (more than 99.9%), pure water and pure ammonia are not of high interest in v apor absorption refrigeration process es when continuous dynamic calculations and high accuracy iteration calculations are necessary, the small fluctuations at very high ammonia compositions and temperature discrepancies may compromise the validity of the resulting calculations. For example, consider the saturated vapor composition at a given temperature and liquid composition at the equilibrium state. If the P T x correlation of El Sayed and Tribus is used as shown in Figure 3 12 it is concluded that the saturated vapor composition will be higher than 100 percent when the mixture (pure ammonia) starts to boil. Because the dew point temperature fluctuates, three different vapor compositions exist for one liquid composition. To avoid these problems, it is necessary for the maximum vapor composition to be limited by the minimum dew point temperature. At 1700kPa, concentrations cannot be higher than 0.99965 of saturated vapor composition and 0.98938 of saturated liquid composition. When the P T x correlations b y Ptek and Klomfar is deployed at high pressures, it is important for the user to keep in mind that for a certain composition range the predicted dew point is lower than the bubble point. Therefore, saturated vapor composition by given liquid composition and temperature cannot be obtained beyond

PAGE 39

39 the temperature matching point as shown at Figure 3 13. El Sayeds also has same problem at lower pressure. Summary Various vapor liquid equilibrium methods were studied in this paper. To avoid timeconsuming iterative numerical calculation, three P T x methods were compared. Bubble and dew point temperature equations are easy to program and show very good agreement with experimental data. However, there are regions where the predicted thermodynamic variables exhibi t local oscillatory patterns characteristic (R unge s phenomenon) because of the nature of polynomial equations. This thermodynamically inconsistent behavior leads to erroneous predictions, including the observations that the predicted dew and bubble point temperatures for pure components do not match, that dew point temperatures are lower than bubble point temperatures, and that there are several different compositions for one given dew point temperature. Under these conditions, the use of polynomial correl ations of high order may lead to erroneous thermodynamic calculations. The alternative of using well developed tabulated data with interpolation can ensure that the problems discussed in the previous paragraph are avoided. Given the memory capabilities of modern computing platforms, it is possible to store and access very large tables at very little cost of execution time. However, to achieve this objective, more accurate wide range experimental data are necessary. Thermodynamic consistency test s might al so be useful and necessary.

PAGE 40

40 Table 3 1 Coefficients for equations (31) and (32) for pure water and ammonia Coefficient Ammonia Water A 1 3.971423e 2 2.748796e 2 A 2 1.790557e 5 1.016665e 5 A 3 1.308905e 2 4.452025e 3 A 4 3.752836e 3 8.389246e 4 B 1 1.634519e1 1.214557e1 B 2 6.508119 1.898065 B 3 1.448937 2.911966e 1 C 1 1.049377e 2 2.136131e 2 C 2 8.288224 3.169291e1 C 3 6.647257e2 4.634611e4 C 4 3.045352e3 0 D 1 3.673647 4.01917 D 2 9.989629e 2 5.17555e 2 D 3 3.617622e 2 1.951939e 2 h L r, o 4.878573 21.821141 h G r,o 26.468879 60.965058 s L r,o 1.644773 5.733498 s G r,o 8.339026 13.45343 T r,o 3.2252 5.0705 P r,o 2 3 Table 3 2. Coefficient for Gibbs excess energy function E 1 41.733398 E 5 63.608967 E 9 0.387983 E 13 3.553627 E 2 0.02414 0 E 6 62.490768 E 10 0.004772 E 14 0.000904 E 3 6.702285 E 7 1.761064 E 11 4.648107 E 15 24.361723 E 4 0.011475 E 8 0.008626 E 12 0.836376 E 16 20.736547 Table 33 Comparison of dew point temperatures P [kPa] x IGT Bogart % error El Say % error Ptek % error 153 0.6 0.9953 333.872 339.256 1.6126 331.804 0.6195 333.333 0.1615 1916.7 0.9953 339.706 345.104 1.5892 338.256 0.4267 338.877 0.2439 1551.3 0.9907 344.761 348.796 1.1703 343.806 0.2770 342.689 0.6012 1868.4 0.9907 348.761 353.607 1.3896 348.737 0.0070 347.232 0.4385 1634. 1 0.9824 355.039 360.086 1.4216 357.682 0.7444 354.332 0.1991 190 3.0 0.9824 360.206 364.322 1.1428 361.633 0.3962 358.141 0.5732 1544.4 0.9641 366.928 371.682 1.2956 371.013 1.1132 366.768 0.0434

PAGE 41

41 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 200 400 600 800 1000 1200 1400 1600 Ammonia Mass FractionPressure [kPa] Gillespie Bogart El-Sayed and Tribus Patek and Klomfar Figure 31. Comparison P T x relations with Gillespie s data at 313.15K 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Ammonia Mass FractionPressure [kPa] Gillespie Bogart El-Sayed and Tribus Patek and Klomfar Figure 32. Comparison P T x relations with Gillespie s data at 405.9K

PAGE 42

42 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 -100 0 100 200 300 400 500 600 700 800 900 Ammonia mass fractionEnthalpy [kJ/kg] IGT Bogart El-Sayed and Tribus Patek and Klomfar 1723.69 kPa 1034.21 kPa 344.74 kPa Figure 33. Comparison for saturated liquid enthalpy 0.965 0.97 0.975 0.98 0.985 0.99 0.995 1 310 320 330 340 350 360 370 380 Ammonia mass fractionTemperature [K]Saturated Vapor Temperature at 1654.74 kPa IGT Bogart El-Sayed and Tribus Patek and Klomfar Figure 3 4. Comparison with IGT data for saturated vapor temperature at 1654.74kPa

PAGE 43

43 0.965 0.97 0.975 0.98 0.985 0.99 0.995 1 1250 1300 1350 1400 1450 1500 Ammonia mass fractionEnthalpy [kJ/kg]Saturated Vapor Enthalpy at 1654.74 kPa IGT Bogart El-Sayed and Tribus Patek and Klomfar Figure 3 5. Comparison with IGT data for saturated vapor enthalpy at 1654. 74kPa 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 250 300 350 400 450 500 Ammonia mass fractionTemperature [K] IGT Bogart El-Sayed and Tribus Patek and Klomfar 1723.69 kPa 1034.21 kPa 344.74 kPa Figure 3 6. Comparison with IGT data for saturated vapor and liquid temperature

PAGE 44

44 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -500 0 500 1000 1500 2000 2500 3000 Ammonia mass fractionEnthalpy [kJ/kg] Bogart El-Sayed and Tribus Patek and Klomfar 1723.69 kPa 344.74 kPa 1034.21 kPa Figure 3 7. Comparison with IGT data for saturated vapor and liquid enthalpy 0.999 0.9991 0.9992 0.9993 0.9994 0.9995 0.9996 0.9997 0.9998 0.9999 1 265 270 275 280 285 290 295 Ammonia mass fractionTemperature [K] IGT Bogart El-Sayed and Tribus Patek and Klomfar Fig ure 3 8. Saturated temperatures for high ammonia mass fraction at 344.74kPa

PAGE 45

45 0.999 0.9991 0.9992 0.9993 0.9994 0.9995 0.9996 0.9997 0.9998 0.9999 1 298 300 302 304 306 308 310 312 314 Ammonia mass fractionTemperature [K] IGT Bogart El-Sayed and Tribus Patek and Klomfar Figure 3 9. Saturated temperatures for high ammonia mass fraction at 1034.21kPa 0.999 0.9991 0.9992 0.9993 0.9994 0.9995 0.9996 0.9997 0.9998 0.9999 1 314 316 318 320 322 324 326 328 Ammonia mass fractionTemperature [K] IGT Bogart El-Sayed and Tribus Patek and Klomfar Figure 3 10. Saturated temperatures for high ammonia mass fraction at 1723.69kPa

PAGE 46

46 0 200 400 600 800 1000 1200 1400 1600 1800 2000 -3 -2 -1 0 1 2 3 Pressure [kPa]Temperature Difference [K] El-Sayed and Tribus at Ammonia Patek and Klomfar at Ammonia El-Sayed and Tribus at Water Patek and Klomfar at Water Figure 3 11. Temperature discrepancies at pure components 0.999 0.9992 0.9994 0.9996 0.9998 1 1.0002 316 316.5 317 317.5 318 318.5 319 319.5 320 Ammonia mass fractionTemperature [K] x = 0.98938 at y = 0.99965 x = 0.98069 at y = 1 Figure 312. Discrepancy and o scillation of ElSayed and Tribus s P T x

PAGE 47

47 0.999 0.9991 0.9992 0.9993 0.9994 0.9995 0.9996 0.9997 0.9998 0.9999 1 313 314 315 316 317 318 319 320 321 322 323 Ammonia mass fractionTemperature [K] x = 0.9661 at y = 1 Figure 313. Discrepancy and o scillation of Ptek and Klomfar s P T x

PAGE 48

48 CHAPTER 4 SYSTEM DESIGN OF A R EFRIGERATION SYSTEM In this r esearch the focus is on a variant of the original PoWER system, one in which the VARS is designed to produce refrigeration at two temperatures. The primary evaporator would operate above freezing, cooling the gas path as well as a local air conditioning l oad. During the summer peak electrical demand in hot climates, the nighttime excess refrigerating capacity would be used to produce ice in a closed thermal storage subsystem. The ice would then be available during the afternoon demand peak to provide suppl emental air conditioning, thus providing a demandside management benefit. Second Law Analysis Using Constant Internal Temperatures For preliminary design of double expansion, single effect VARS using the Second Law approach, temperatures for main components need to be selected. Most evaporators show a single temperature for their refrigerant in the refrigeration side Therefore, constant internal temperatures at three evaporators have been used to obtain the Second Law efficiency. Thermodynamic Model Figur es 4 1 and 4 2 show the PoWER system, split for convenience into a gas turbine flow path (Fig ure 41) and the coupled vapor absorption refrigeration system ( Figure 42 ). The gas turbine operates on a semi closed cycle, in that a substantial fraction of the flow from the hot side recuperator exit is mixed with the fresh air leaving the low pressure compressor. The majority of the waste heat of the engine is removed from the mixture at an intermediate pressure (typically several atmospheres), allowing much sm aller heat exchangers than conventional ones at the exhaust. The captured

PAGE 49

49 waste heat powers the VARS in this implementation. The VARS cooling capacity is used in part to chill the gas at the entrance to the highpressure compressor, thereby significantly increasing the efficiency. The gas turbine flow path model has been reported by Lear and Laganelli [ 2 8 ]. The component models are conventional for such cycle calculations except that the extraction of water condensate from the evaporator is taken into account, along with the influence of water concentration on thermodynamic properties. Results from Khan et al. [ 29] were utilized in combination with a conceptual refrigeration system optimized for this application. Since the current analysis is in the conceptual design stage, it would be incorrect to specify the design of the VARS. Instead, we have chosen to calculate the ideal COP for a reversible refrigeration system, then parameterize the fraction of that ideal to be achieved by an actual system. Stated diff erently, this analysis utilizes a Second Law efficiency to allow calculation of the VARS performance. For the Second Law analysis, it is recognized that a control volume enclosing the two coolant temperature VARS is equivalent to one enclosing two conventi onal VARS systems, one for each low temperature reservoir as shown at Fig ure 4 3. For each conventional VARS unit, the First and Second Laws may be written as 3 ,, 1 in GeniEip iQQQW ( 4 1 ) 3 0, 1 out iin iQQQ ( 4 2 ) ,,, ,0GeniEioi Si GenEioQQQ P TTT ( 4 3 )

PAGE 50

50 Combining Eq uations ( 4 1 ) ( 4 2 ) and ( 4 3 ) with negligible pW ,, Eio oGen Ei Geni Ei GenTT TT QQ TT ( 4 4 ) For the reversible case, the entropy production is zero, so the equality in Eq uation ( 4 4 ) is appropriate, yielding 1,2 1,2 1,2() ()E EGenO Carnot GenGenOEQ TTT COP QTTT ( 4 5 ) ,3 3 ,3() ()E ICEGenO Carnot GenGenOICEQ TTT COP QTTT ( 4 6 ) where TE,i is TE, for E vaporator s 1 and 2, and TICE is the temperature for E vaporator 3. For each evaporator, the refrigeration efficiency is specified as a Second Law based efficiency parameter. Using the conventional COP definition, Ei i GeniQ COP Q ( 4 7 ) e nables solving for the COP values, given the Carnot COP values and refrigeration efficiency, from i Ri CarnotiCOP COP ( 4 8 )

PAGE 51

51 Solution Method A computer code described previously by Khan et al. [ 29 ] was written in FORTRAN to simulate the performance of the PoWER system, using traditional cycle analysis methods The code is capable of calculating engine performance (network output, water extraction flow rate and thermal efficiency) as a function of the input parameters (turbine inlet temperature, recuperator inlet temperature, generator temperature, evaporator exit temperature, low pressure compressor ratio, recirculation ratio, turbo machinery efficiencies, heat exchanger effectiveness, equivalence ratio, fuel type, and pressure drops). A wide range of VARS technologies exist, typically trading off between complexity and COP. The intent of the current paper is to quantify the potential for ice making and load leveling, not to analyze one particular design choice. Therefore, the VARS was studied by a Second Law analysis using as input the results of the HPRTE code. The refrigeration efficiency is taken as a parameter to represent the maturity of the VARS technology chosen. The temperatures used for determining the Carnot coefficient of performance, COPCarnot, in this research are tabulated in Table 4 1 all temperatures are either results from the HPRTE code or the specified environmental temperature. The variation of the outdoor temperature is shown in Fig ure 4 3, and the values of generator and evaporator temperatures are taken from Goswami et al. [ 3 0 ] TICE was chosen by referring to the ASHRAE H andbook [ 3 1 ]. Fig ure 4 4 shows the maximum air condition load at 4 PM, while Fig ure 45 shows the maximum outdoor air temperature at 3 PM [3 2 ] This phase lag is because of thermal storage of the building, the movement of heat sources such as people, and time dependent use of lights in the building [ 3 3 ]. Fig ure 4 6 shows the

PAGE 52

52 hourl y maximum COP s (or COPCarnot) calculated by using the variation of outdoor temperature, To as shown Fig ure 43. The cycle input and output parameters which will be used for a multistage VARS study are given in Table 4 1. The low pressure compressor pressure ratio was based on prior studies [ 3 4 ] and is chosen for maximum thermal efficiency and reasonable water extraction. The inlet temperature of the gaseous mixture entering the highpressure compressor is constant and is equal to 10C which is a practi cal limit taking into account the fact that Evaporator 1 temperature is fixed at 5C The turbo machinery polytrophic efficiencies were based on stateof the art values as reported by Mattingly [ 3 5 ] and Carcasci and Facchini [ 3 6 ]. The pressure drops in the recuperator and the combustor are similar to the values reported by Fiaschi et al. [ 3 7 ]. The equivalence ratio was set to 0.9 based on previous studies with the HPRTE as given in Boza et al. [ 3 8 ]. It was assumed to be equal to 0.9, because it was expected that at least 10% excess air was necessary to ensure complete combustion [ 3 4 ]. The gas entering the combustor is diluted with excess air and the recirculated combustion products, to lower the adiabatic flame temperature towards the specified turbine inlet temperature (This provides benefits in emissions as well as combustion efficiency). The effectiveness of the recuperator is assumed to be 0.85 and the recuperator inlet temperature (on the hot side) is assumed to be 80 0 C These are conservative values based on existing commercial equipment [ 39]. Turbine blade cooling is considered for higher turbine inlet temperatures, and the model for turbine blade cooling used is similar to that given by Massardo et al. [ 4 0 ]. This model gives a relation between the turbine inlet temperature and the percentage of main air flow used for turbine blade

PAGE 53

53 cooling. Note that the intent here is to model a future system with advanced cooling and recuperator technology. This is the conservative choice from the standpoint of cooling and icemaking provided by the VARS, since the high firing and recuperator temperatures lead to high gas tur bine efficiency, hence reduced waste heat for driving the VARS. That in turn limits the icemaking potential and the load leveling capability due to thermal energy storage. This conservative approach is chosen in order to balance the simplifications made in the thermal energy storage system, primarily the omission of parasitic power requirements and heat leaks. Results and Discussion The computer code of the HPRTE cycle and the second l aw approach of the multistage VARS were used to calculate hourly air conditioning and ice making capacities Given HPRTE data, which are calculated by Khan s model [ 30 ] together with the outdoor air temperature, air conditioning hourly required load, and refrigeration efficiency, the COPCarnot, COPs heat supplied to the gener ator, and heat removed from the surroundings for all three evaporators are calculated below and illuminated by various plots. In addition to those results the amount of ice produced daily and hourly, and the air conditioned floor area are shown. These val ues were determined for six cases, varying the ratio of the heat added from the air conditioner ( E vaporator 2) and ice maker (Evaporator 3) to the heat added at E vaporator 1 for two different refrigeration efficiencies, ,2,3 ,1 EE evp EQQ R Q ( 4 9)

PAGE 54

54 To produce additional heat to run Evaporators 2 and 3, the heat capacity of the generator needs to be higher than that of normal HPRTE system, which has only one evaporator. This higher generator capacity leads to making the required capacity of the cold gas cooler smaller to fix the properties at Point 9.1 and Point 3.2. Because there is a limitation, however, for the generator to be bigger, 0.7 is chosen as the maximum Revp in this study. Fig ure 4 6 shows the maximum COPs based on varying ambient temper atures, as well as constant maximum and minimum temperatures (generator and evaporator) Because the CGC ( E vaporator 1) and the air conditioner ( E vaporator 2) use the same temperatures, as shown in Fig ure 4 2, the Carnot COPs for the two evaporators are sa me. However, the ice maker needs colder refrigerant, and this reduces the Carnot COP of E vaporator 3 As expected, all Carnot COPs have minimum values at the hottest time of 3 PM. Fig ures 4 7 and 4 8 show the COP variations of the three evaporators obtained f rom Equation ( 4 8) The COPs have the same trend as the Carnot COPs since the refrigeration efficiency is held constant The total COP variations of the system are shown in Fig ures 4 9 and 4 10 for R of 0.5 and 0.7. Obviously, the higher refrigeration efficiency the higher COPsys. The minimum COPsyss occur at different time s, but they are in the day time between 12: 00 PM and 15 : 00PM For Revp= 0.3 Fig ures 4 11 through 4 14 represent the results when the ratio is 0.3 for R equal to 0.5 and 0.7. Heat removed from the surroundings at E vaporator 1 is fixed at all time s to maintain HPC inlet gas temperature. While E vaporator 1 cools the HPC inlet gas, E vaporator 2 is used as an air conditioner which maint ains the indoor

PAGE 55

55 temperature of 27C in a conditioned space of 671.1m2. As the air conditioning required load decreas es E vaporator 3 makes more ice. Because the COP for E vaporator 3 is lower than that of E vaporator 2, the allocated heat supply required from the generator for the ice maker is relatively higher even though the heat rejected from the surroundings is of the same order for E vaporator s 1 and 2. For Revp= 0.5: As the capacity of the heat rejected to E vaporator 1 is lower than that of the 0.3 Revp case, as shown in Fig ures 4 15 and 416 the heat supplied to E vaporator 1 is also somewhat less while the heat absorbed in E vaporator s 2 and 3 are somewhat more. Clearly when the refrigeration efficiency is higher as in Fig ures 4 17 and 4 18, the loads ar e smaller for E vaporator s 1 and 2, so the system can make more ice. For Revp= 0.7 Fig ures 4 19 through 4 22 show that the allocated heat supply to Evaporator 1 is reduced, while those of E vaporator s 2 and 3 are increased. Because the HPC inlet gas temperat ure must be maintained, more generator heat is needed for air conditioning and icemak ing Approximately 9.85% more heat is needed to increase Revp from 0.3 to 0.7. For this increase of generator heat capacit y almost double the floor area can be cooled, an d 28% and 19% more ice can be made for R of 0.5 and 0.7 respectively. Fig ures 4 23 and 4 24 and Table 4 2 show the comparison of results with various Revp values The amount of ice produced hourly is shown in Fig ures 4 9 and 4 18 As Revp is increased larger amounts of ice are produced, but the rate of increase is reduced The more the generator cools the gas in the gas turbine system, the less the required heat capacity of E vaporator 1, while the total amount of ice produced per day and conditioned area are increased.

PAGE 56

56 Summary A steady state gas path model and a s econd l aw VARS analysis were used for a thermodynamic performance study of a novel cooling and power cycle that combines a semi closed cycle gas turbine called the HPRTE with a VARS. As shown above, if the generator capacity is higher, E vaporator 1 can be downsized, and additional heat from generator is better used in different applications, such as air conditioning and ice making T he ice can be used to chill water to cool hot fluids inside the HPRTE/VARS the next day which can inc rease the total thermal efficiency of the system or for additional air conditioning The amount of ice available can ideally produce twice the total daily cooling capacity in air conditioning. This cooling capacity could be used to cool a second buildi ng of a size similar to the design building, or be used in a load leveling manner to downsize the entire system (thus reducing capital costs). I ce can also be used for other applications and if the air conditioning load is limited, much more ice can be produced. However, if the ratio of E vaporator 2 and 3 loads to E vaporator 1 load needs to change, the process of re sizing the heat exchanger s is nonlinear and needs to be taken into account Ther e is also a limitation o n the upsizing possible for the generator, thus requiring an optimization. Considering all the issues/points mentioned above it can be concluded that the combined HPRTE/VARS or PoWER, cycle seems to be a promising candidate for distributed power generation especially in hot climates where t he summer loadleveling benefits are important Se cond Law Analysis Using Variable External Temperatures As discussed above, most evaporators have a single temperature theoretically in the refrigeration side. However, the fluids being cooled down by evapor ator s or

PAGE 57

57 generators do not have to have a single value depending on heat exchanger type s. In this sub chapter, variable or constant external temperatures are selected to obtain Carnot COP. Thermodynamic Model The Second Law model for the VARS is based on t he system defined by the dashed line in F igure 4 25. The system receives high temperature heat from the gas turbine, rejects heat to the ambient, and receives low temperature heat from three reservoirs the gas turbine, building air conditioning, and the ice storage subsystem. Pump and fan power are neglected, though those parasitic losses may be incorporated into the gas turbine efficiency calculation. In utilizing a refrigeration efficiency, the first step is to calculate the performance of a reversible system operating among these thermal reservoirs. It may be shown that the reversible system indicated is equivalent to three traditional Carnot refrigerators (heat actuated, not vapor compression), each cooling one of the three loads. The Carnot COP of suc h systems is a well known function of the three temperatures. Since the cooling environments and temperatures differ for each of the three evaporators, each will have a unique Carnot COP. In this work, the Second Law refrigeration efficiencies of each of t he three refrigeration systems are assumed equal, corresponding to an equivalent maturity of the technology. The relationship between cooling rate and heating rate for each of the three refrigeration systems is set by the assumed refrigeration efficiency and the value of the Carnot COP determined from the reservoir temperatures. However, there are several reasonable approaches to calculating the Carnot COP, based on different choices of reservoir temperatures. One approach is to perform VARS cycle calculati ons for a particular configuration and to use the generator temperature for

PAGE 58

58 the maximum value, the ambient as the intermediate value, and the three evaporator temperatures of the refrigerant as the low temperature values. This is considered an internal Car not COP as in Lear et al. [41] and it does not include the irreversibilities in the heat exchangers. A second approach is to take the maximum and minimum te mperatures necessary for calculation of Carnot COP to be at the hot gas inlet (State 9.1 in Figure 4 1) and the cold gas exit (State 3.2), respectively. A variant of this approach is to take into account that the external reservoirs are not at fixed temperature, but rather change temperature as a result of the heat transfer. The COP for a reversible cyc le under these circumstances has been derived in the present work. The temperatures used for the Carnot COP calculations are given in Table 4 3 TH, TL1, TL2, and TL3 are the internal working fluid temperatures, which are constants at all times. The calcul ation involving variable reservoir temperatures require knowledge of both the inlet and the exit temperature, so these are also listed. However, the external temperatures of the air conditioner (Evaporator 2) and ice maker (Evaporator 3) are taken to be constants. In this paper, external temperatures are used for the generator and the evaporators. Figure 4 26 shows four cases using different external temperatures for the system : Constant generator and evaporator temperatures Variable generator temperature with constant evaporator temperature Constant generator temperature with variable evaporator temperature Variable generator and evaporator temperatures

PAGE 59

59 For Case (a), the higher temperature, TGi is used for the constant reservoir temperature while the lower temperature, TEo is used for the lower reservoir temperature (Figure 4 26 (a)). The First and Second laws for the case may be written as ,1,1,1,1,1,1 inHLpHLQQQWQQ ( 4 1 0 ) ,1,1,1 outoinQQQ ( 4 11) ,1,1,1 ,10HLo s GiEooQQQ P TTT ( 4 12) where ,1 inQ is energy input to the system, ,1 HQ is allocated heat input for the Evaporator 1 from the generator, ,1 LQ is heat input from the Evaporator 1, and ,1 pW is pump work for Evaporator 1, neglected here. ,1 outQ and ,1 oQ are the heat released outside of the system and ,1 sP is the entropy production for Subsystem 1. Combining E q uations ( 4 1 0 ) ( 4 11) and ( 4 12 ) we obtain ,1 ,1 Eoo oGi LH Eo GiTTTT QQ TT (4 13) For the reversible case, the entropy production is zero, so the equality in Eq uation ( 4 13) is appropriate, yielding the Carnot COP for Subsystem 1 and Case (a): ,1 1, ,1 L EoGio Ca HoEoGiQ TTT COP QTTT ( 4 14) For Case (b), in which the generator temperature changes from TGi to TGo as shown at Figure 4 27 (A), the local COP at Point x can be written as

PAGE 60

60 ,1 ,1 Lx dx HxdQ COP dQ ( 4 15) If mass flow rate and heat capacity are constants, the heat input s to the system are, 0 ,1 ,1 1()Gi GoxT H Hx pxpGiGo xTQdQmCdTmCTT ( 4 16) 00 ,1 ,1 ,1 11 x L Lx dxHx xQdQCOPdQ ( 4 17) When the Carnot COP is introduced, we get ,1Gi GoT xoEo Lp x T xoEoTTT QmC dT TTT ( 4 18) Combining Equations (416) and (418 ) with some algebra, we obtain 1,1lnEo o Gi Cb oEo GiGoGoTTT COP TTTTT ( 4 19) With simil ar analysis as shown in Figure 427 (B) the Carnot COP for Case (c) can be easily determined 0 ,1 ,1 1()Ei EoyT L Ly EpEyEpEEiEo yTQdQmCdTmCTT ( 4 20) 0 ,1 ,1 ,1 1Ei EoyT Ly H Hy yT dydQ QdQ COP ( 4 21) When the Carnot COP expression is again introduced, we obtain

PAGE 61

61 ,1 ,Ei EoT oy Gi H EpEy T GioyTT T Q mCdT TTT ( 4 22) Finally, we can get the Carnot COP for Case (c) as 1,ln1Gio Gi Cc o Ei EiEoEoTT T COP TT TTT ( 4 23) In the case that both the generator and evaporator temperatures vary, it is assumed that they are independent Then the Carnot COP for Case (d) is determined as 1,1ln ln1o Gi GiGoGo Cd o Ei EiEoEoTT TTT COP TT TTT ( 4 24) Because the external temperatures of Evaporators 2 and 3 are constant in this study, only cases (a) and (b) are considered. Therefore Equations ( 4 14) and ( 4 19), in which the lower temperature is replaced with the constant air conditioned room temperature and ice temperature, TAC and TICE, respectively, are also used for Evaporators 2 and 3. For each evaporator, the refrigeration efficiency is specified as a Second Law based e fficiency parameter. Using the conventional COP definition,

PAGE 62

62 Ei i HiQ COP Q ( 4 25) e nables solving for the COP values, given the Carnot COP values calculated from the specified temperatures, and refrigeration efficiency, from i Ri CiCOP COP ( 4 26) Solution Method A computer code described previously by Khan et al. [ 2 9] simulate s the performance of the PoWER system, using traditional cycle analysis methods The code is capable of calculating engine performance (network output, water extraction flow rate and thermal efficiency) as a function of the input parameters (turbine inlet temperature, recuperator inlet temperature, generator temperature, evaporator exit temperature, low pressure compressor ratio, recirculation ratio, turbomachinery efficiencies, heat exchanger effectiveness, equivalence ratio, fuel type, and pressure drops). The intent of the current paper is to quantify the potential for ice making and load leveling, not to analyze one particular design choice. Therefore, the VARS was simulated by a Second Law analysis as described above, using as input the results of the HPRTE code of Khan et al [ 2 9]. The refrigeration efficiency is taken as a parameter to represent the maturity of the VARS technology chosen. The temper atures used for determining the Carnot coefficient of performance, COPCarnot, in this paper are tabulated in Table 4 3 The variation of the air conditioning hourly load and outdoor temperature [ 32] are shown in Figures 4 4 and 4 5 and the values of gener ator and evaporator temperatures are taken from Goswami et al. [ 30 ]

PAGE 63

63 TICE was chosen by Huang et al. [ 42 ]. Fig ure 4 5 shows the maximum air conditioning load at 4 pm, while Figure 4 4 shows the maximum outdoor air temperature at 3 pm. This phase lag is bec ause of thermal storage of the building, the movement of heat sources such as people, and time dependent use of lights in the building [ 33]. The cycle input and output parameters which will be used for a multi temperature VARS study are given in Table 4 4 The low pressure compressor pressure ratio was based on prior studies [ 34 ] and is chosen for maximum thermal efficiency and reasonable water extraction. The inlet temperature of the gaseous mixture entering the highpressure compressor is constant an d is equal to 10C, which is a practical limit taking into account the fact that the internal temperature of Evaporator 1 is fixed at 5C. The turbomachinery polytrophic efficiencies were based on stateof the art values as reported by Mattingly [ 35] and C arcasci and Facchini [ 3 6 ]. The pressure drops in the recuperator and the combustor are similar to the values reported by Fiaschi et al. [ 3 7 ]. The equivalence ratio was set to 0.9 based on previous studies with the HPRTE as given in Boza et al. [ 3 8]. It w as assumed to be equal to 0.9, because it was expected that at least 10% excess air was necessary to ensure complete combustion [ 34 ]. The gas entering the combustor is diluted with excess air and the recirculated combustion products, to lower the adiabatic flame temperature towards the specified turbine inlet temperature ( t his provides benefits in emissions as well as combustion efficiency). The effectiveness of the recuperator is assumed to be 0.85 and its inlet temperature (on the hot side) is assumed to be 80 0 C. These are conservative values based on existing commercial equipment [ 39]. Turbine blade cooling is considered for

PAGE 64

64 higher turbine inlet temperatures, and the model for turbine blade cooling used is similar to that given by Massardo et al. [ 40]. T his model gives a relation between the turbine inlet temperature and the percentage of the main air flow used for turbine blade cooling. Note that the intent here is to model a future system with advanced cooling and recuperator technology. This is the conservative choice from the standpoint of cooling and icemaking provided by the VARS, since the high firing and recuperator temperatures lead to a high er gas turbine efficiency, hence reduced waste heat for driving the VARS. That in turn limits the ice maki ng potential and the load leveling capability due to thermal energy storage. This conservative approach is chosen in order to balance the simplifications made in the thermal energy storage system, primarily the omission of parasitic power requirements and heat leaks. Results and Discussion The computer code of the HPRTE cycle and the Second Law approach of the multistage VARS were used to calculate hourly air conditioning and ice making capacities Given HPRTE data, which are calculated by Khan s model [ 2 9 ] together with the outdoor air temperature, air conditioning hourly required load, and refrigeration efficiency, the COPCarnot, COPs, heat supplied to the generator, and heat removed from the surroundings for all three evaporators were calculated and the results plotted. In addition to those results the amount of ice produced daily and hourly and the air conditioned floor area are shown. These values were determined for six cases, varying the ratio of the heat added from the air conditioner ( E vaporator 2) and ice maker (Evaporator 3) to the heat added at E vaporator 1 for two different refrigeration efficiencies.

PAGE 65

65 Figures 4 2 8 to 4 3 0 show the hourly Carnot COP for the evaporators. All cases have same trend; when ambient temperature is high, the Carnot COP is low and vice versa. For all evaporators, the Carnot COPs using external temperatures are higher than those using internal temperatures. This is because the evaporator external temperatures are higher than their internal temperatures as shown at Table 4 3 The temperature differences between the external temperatures and ambient temperature are smaller before sunset. Therefore, the Carnot COP has a maximum at around 5am for Evaporators 2 and 3. Around the same time, the ambient temperature is less than the air conditioned room temperature. This results in the Carnot COP for Evaporator 2 being less than zero because when only the ambient and room temperatures are compared, no refrigeration effect is required. T o avoid this problem, the Carnot COP equations need certain conditions imposed for four cases (Equations ( 4 14), (4 19), (4 23 ), and ( 4 24 )); GioEoTTT ( 4 27) lnGiGo oEo Gi GoTT TT T T ( 4 28) lnEiEo Gio Ei EoTT TT T T ( 4 29) ln lnGiGo EiEo o Gi Ei Go EoTTTT T TT TT ( 4 30)

PAGE 66

66 For Evapor ator 2, conditions ( 4 27) and ( 4 28) above are used by replacing TEo with TAC. In this paper, rather than using negative Carnot COP, it is considered that Evaporator 2 would not function as an air conditioner when the ambient temperature is lower than the temperature produced in Evaporator 2. As a result, the air conditioning load from 1 to 7 a. m is zero. Figures 4 3 1 and 4 3 2 show the cooling effect of the eight different cases for Evaporators 2 and 3. It is obvious that the higher the refrigeration effic iency the higher the refrigerating effect. When a variable lower temperature, due to sensible heat addition from the gas path, for Evaporator 1 with a fixed generator temperature is considered, the heat gain from Evaporators 2 and 3 is always highest whi le it is lowest when a variable generator temperature and a fixed evaporator temperature are considered instead Ice produced per day is indicated in Table 4 5 based on the results of Figure 4 3 2 when no load and no thermal loss are considered during the process. Consider Case (a) when R = 0.5. When the ice storage is used to replace conventional air conditioning having 2.5 of COP without loss, 8279 kWh of electric energy can be saved a day. In a practical ice storage system, however, there must be a loss associated with mak ing ice and using ice for air conditioning or refrigeration but ice can be made and used simultaneously. Therefore, the total volume of a real ice storage does not have to be so large as shown in Table 4 5 The allocated heat input from the generator to each subsystem is shown in Figures 4 3 3 to 4 3 5 Because ,1 LQ is 234.6 kW continuously more heat is needed at lower refrigeration efficiencies to produce enough cooling. Cases (b) and (d) use a

PAGE 67

67 variable generator temperature, which means that the average value of the inlet and exit temperatures is lower than TGi, so, more heat is necessary than those of Cases (a) and (c), respectively. However, the allocated generator heat inputs for Evaporator 2 and 3 show the exa ct opposite trends. M ore heat is necessary when the efficiency is higher and Evaporator 1 is using a variable temperature. Figure 4 3 6 shows the efficiency ratio based on the efficiency of Case (a). To achieve the same COP for Case (a), about 45% higher ef ficiency is needed for Case (b). With lower efficiency, however, Cases (c) and (d) can achieve the same refrigerating effect. Summary A loadleveling implementation of the PoWER cycle has been analyzed in order to determine the potential for demand side ma nagement via cold thermal energy storage integrated with this novel distributed generation system. The analysis included an exploration of the appropriate methods of determinin g the performance of the refrigeration subsystem using Second Law efficiency as a parameter. The results indicated that the daily storage of thermal energy in the form of ice is roughly equivalent to the air conditioning supplied directly to a building, for typical hot climate conditions. The sizing of the engine relative to the buil ding depends on economic factors outside the scope of this analysis; however, it is clear that a very significant decrease in peak power requirement for air conditioning is achievable. Results presented in this paper also illustrate that internal temperatures are best used when calculating Carnot COP values, convenient for conceptual design of such systems. Often the external temperatures would be better defined, and therefore they would be attractive as a means of simplifying the overall analysis, combining

PAGE 68

68 irreversibilities of the heat exchangers with those of the refrigeration cycle. However, it is clear that nonphysical results can easily occur when the ambient temperature approaches the low temperature reservoir temperature, rendering that approach ine ffective. It is therefore recommended that heat exchanger irreversibilities be included separately from cycle irreversibilities.

PAGE 69

69 Table 41. Input/o utput parameters of the HPRTE system. Parameters Values Turbine i nlet t emperature [ C ] 1400 Recuperator i nlet t emperature [ C ] 800 High pressure compressor ratio 9.02 Low p ressure c ompressor r atio 2 Turbo machinery polytropic efficiencies [ % ] 90 Combustor e quivalence ratio, 0.9 Effectiveness of the r ecuperator 0.85 Effectiveness of the intercooler 0.85 Relative humidity of inlet air, 0.9 Combustor p ressure drop 0.05 Intercooler p ressure drop 0.03 Generator p ressure drop 0.03 Evaporator 1 pressure drop 0.03 Recup erator p ressure drop ( HPT exit side ) 0.06 Recuperator p ressure drop ( LP C exit side ) 0.02 Generator temperature, T G [ C ] 100 Evaporator 1, 2 temperature, T E1,2 [ C ] 5 Evaporator 3 temperature, T ICE [ C ] 25 Thermal efficiency, th [%] 40.4 Net power o f system [kW] 455 Mass flow rate of fuel [ kg /s ] 0.0243 Mass flow rate of water extracted [ kg /s ] 0.0368 Mass flow rate of air input [ kg /s ] 0.43 Modified recirculation ratio, R mod 2.45 Table 42. Comparison of results R evp 0.3 0.5 0.7 Generator heat c apacity, Q Gen [kW] 417.6 440.0 (5.36%) 458.7 3 (9.85%) Evaporator 1 heat capacity, E [kW] 257.0 234.6 ( 8.72%) 215.9 ( 16.0%) Area to be Air Conditioned [m 2 ] 671.1 1021.2 (52.2%) 1315.3 (96.0%) Ice Produced when R =0.5 [kg/day] 20235 23308 (13.2%) 25883 (24.2%) Ice Produced when R =0. 7 [kg/day] 40618 44782 (10.3%) 48273 (18.8%)

PAGE 70

70 Table 4 3 Input parameters of the VARS Parameters Values [ C ] T H Internal temperature of the generator 100 T Gi Inlet temperature of the generator 452.3 T Go Exit temperature of the generator 68.46 T L1 Internal temperature of the evaporator 1 5 T Ei Inlet temperature of the evaporator 1 29.25 T Eo Exit temperature of the evaporator 1 10 T L2 Internal temperature of the evaporator 2 5 T AC Air conditioned room temperature 25 T L3 Internal temperature of the evaporator 3 7 T ICE Ice temperature 0 Table 4 4 Input/output parameters of the HPRTE system. Parameters Values Turbine i nlet t emperature [ C ] 1400 Recuperator i nlet t emperature [ C ] 800 High pressure compresso r ratio 9.02 Low p ressure c ompressor r atio 2 Turbo machinery polytropic efficiencies [%] 90 Combustor e quivalence ratio, 0.9 Effectiveness of the r ecuperator 0.85 Effectiveness of the intercooler 0.85 Relative humidity of inlet air, 0.9 Combustor p ressure drop 0.05 Intercooler p ressure drop 0.03 Generator p ressure drop 0.03 Evaporator 1 p ressure drop 0.03 Recuperator p ressure drop ( HPT exit side ) 0.06 Recuperator p ressure drop ( LP C exit side ) 0.02 Thermal efficiency, th [%] 40.4 Net pow er of system [kW] 455 Mass flow rate of fuel [ kg /s ] 0.0243 Mass flow rate of water extracted [ kg /s ] 0.0368 Mass flow rate of air input [ kg /s ] 0.43 Modified recirculation ratio, R mod 2.45 Table 4 5 Ice produced a day [ Ton /day (L/day) ] R = 0.7 R = 0.5 Case (a) 75.354 (81907) 52.159 (56695) Case (b) 51.275 (55734) 34.960 (38000) Case (c) 78.964 (85830) 55.769 (60618) Case (d) 54.885 (59658) 38.570 (41924)

PAGE 71

71 Figure 4 1. Gas turbine flowpath of the PoWER system Figure 42. The multi stage v apor absorption refrigeration system

PAGE 72

72 Gas Turbine, A/C VARS Ice Maker VARS TO TICE TE TGEN 12 GenGenQQ 3 GenQ 1 EQ 2 EQ 3 EQ 03Q 0102QQ Figure 43. The multi stage simple vapor absorption refrigeration system 0 4 8 12 16 20 24 0 10 20 30 40 50 60 70 80 90 100 Time [hr]Air Condition Load [%] Fig ure 44. Air conditioner hourly load for typical summer day

PAGE 73

73 0 4 8 12 16 20 24 22 24 26 28 30 32 34 36 Time [hr]Temperature [oC] Fig ure 4 5. Typical variations of outdoor air temperature at 42 north latitude on July 1 0 4 8 12 16 20 24 0.5 1 1.5 2 2.5 3 3.5 Time [hr]COPCarnot Evap 1,2 Evap 3 Fig ure 46. Carnot COP variations for the three evaporators

PAGE 74

74 0 4 8 12 16 20 24 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 Time [hr]COP COP 1,2 COP 3 Fig ure 4 7. COP variations for three evaporators at R=0.5 0 4 8 12 16 20 24 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 Time [hr]COP COP 1,2 COP 3 Fig ure 48. COP variations for three evaporators at R=0.7

PAGE 75

75 0 4 8 12 16 20 24 0.75 0.8 0.85 0.9 0.95 1 Time [hr]COPsys Revp = 0.3 Revp = 0.5 Revp = 0.7 Fig ure 49. Total C OP variations of the system at R=0.5 0 4 8 12 16 20 24 0.9 0.95 1 1.05 1.1 1.15 1.2 Time [hr]COPsys Revp = 0.3 Revp = 0.5 Revp = 0.7 Fig ure 4 10. Total COP variations of the system at R=0.7

PAGE 76

76 0 4 8 12 16 20 24 0 50 100 150 200 250 300 Time [hr]Heat from surroundings [kW] Evap 1 Evap 2 Evap 3 Fig ure 4 11. Heat removed from surroundings at R=0.5, Revp=0.3 0 4 8 12 16 20 24 0 50 100 150 200 250 300 350 Time [hr]Heat supply from generator [kW] Qgen1 Qgen2 Qgen3 Fig ure 412. Allocated heat supply from generator at R=0.5, Revp=0.3

PAGE 77

77 0 4 8 12 16 20 24 0 50 100 150 200 250 300 Time [hr]Heat from surroundings [kW] Evap 1 Evap 2 Evap 3 Fig ure 413. Heat removed from surroundings at R=0.7, Revp=0.3 0 4 8 12 16 20 24 0 50 100 150 200 250 300 Time [hr]Heat supply from generator [kW] Qgen1 Qgen2 Qgen3 Fig ure 414. Allocated heat supply from generator at R=0.7, Revp=0.3

PAGE 78

78 0 4 8 12 16 20 24 0 50 100 150 200 250 Time [hr]Heat from surroundings [kW] Evap 1 Evap 2 Evap3 Fig ure 4 15. Heat removed from surroundings at R=0.5, Revp=0.5 0 4 8 12 16 20 24 0 50 100 150 200 250 Time [hr]Heat from surroundings [kW] Evap 1 Evap 2 Evap 3 Fig ure 4 16. Heat removed from surroundings at R=0.7, Revp=0.5

PAGE 79

79 0 4 8 12 16 20 24 0 50 100 150 200 250 300 Time [hr]Heat supply from generator [kW] Qgen1 Qgen2 Qgen3 Fig ure 4 17. Allocated heat supply from generator at R=0.5, Revp=0.5 0 4 8 12 16 20 24 0 50 100 150 200 250 300 350 Time [hr]Heat supply from generator [kW] Qgen1 Qgen2 Qgen3 Fig ure 4 18. Allocated heat supply from generator at R=0.7, Revp=0.5

PAGE 80

80 0 4 8 12 16 20 24 0 50 100 150 200 250 Time [hr]Heat from surroundings [kW] Evap 1 Evap 2 Evap 3 Fig ure 4 19. Heat removed from surroundings at R=0.5, Revp=0.7 0 4 8 12 16 20 24 0 50 100 150 200 250 300 Time [hr]Heat from surroundings [kW] Evap 1 Evap 2 Evap 3 Fig ure 420. Heat removed from surroundings at R=0.7, Revp=0.7

PAGE 81

81 0 4 8 12 16 20 24 0 50 100 150 200 250 300 350 Time [hr]Heat supply from generator [kW] Qgen1 Qgen2 Qgen3 Fig ure 421. Allocated heat supply from generator at R=0.5, Revp=0.7 0 4 8 12 16 20 24 0 50 100 150 200 250 300 350 400 Time [hr]Heat supply from generator [kW] Qgen1 Qgen2 Qgen3 Fig ure 4 22. Allocated heat supply from generator at R=0.7, Revp=0.7

PAGE 82

82 0 4 8 12 16 20 24 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Time [hr]Ice produced [kg/hr] Revp = 0.3 Revp = 0.5 Revp = 0.7 Fig ure 4 23. Ice produced hourly for various Revp at R=0.5 0 4 8 12 16 20 24 500 1000 1500 2000 2500 3000 Time [hr]Ice produced [kg/hr] Revp = 0.3 Revp = 0.5 Revp = 0.7 Fig ure 4 24. Ice produced hourly for various Revp at R=0.7

PAGE 83

83 Fig ure 4 2 5 The multi temperature simple vapor absorption refrigeration system Figure 42 6 Four cases of external temperatures A B Figure 4 2 7 COP for a small control volume when A) generator t emperature varies and B) E vaporator 1 temperature varies

PAGE 84

84 0 4 8 12 16 20 24 0 5 10 15 20 25 30 35 40 45 50 Time [hr]COPCarnot Internal Temperature External Case (a) External Case (b) External Case (c) External Case (d) Fig ure 4 28. Carnot COP for Evaporator 1 0 4 8 12 16 20 24 0 5 10 15 20 25 30 35 40 45 50 Time [hr]COPCarnot,2 Internal Temperature External Case (a) External Case (b) Fig ure 4 29. Carnot COP for Evaporator 2

PAGE 85

85 0 4 8 12 16 20 24 0 1 2 3 4 5 6 7 8 9 10 Time [hr]COPCarnot Internal Temperature External Case (a) External Case (b) Fig ure 4 3 0 Carnot COP for Evaporator 3 0 4 8 12 16 20 24 0 1000 2000 3000 4000 5000 6000 Time [hr]QL2 [kW] Case (a) R=0.7 Case (b) R=0.7 Case (c) R=0.7 Case (d) R=0.7 Case (a) R=0.5 Case (b) R=0.5 Case (c) R=0.5 Case (d) R=0.5 Fig ure 4 3 1 Heat removed at Evaporator 2

PAGE 86

86 0 4 8 12 16 20 24 0 500 1000 1500 2000 2500 Time [hr]QL3 [kW] Case (a) R=0.7 Case (b) R=0.7 Case (c) R=0.7 Case (d) R=0.7 Case (a) R=0.5 Case (b) R=0.5 Case (c) R=0.5 Case (d) R=0.5 Fig ure 4 3 2 Heat removed at Evaporator 3 0 4 8 12 16 20 24 0 20 40 60 80 100 120 Time [hr]QH1 [kW] Case (a) R=0.7 Case (b) R=0.7 Case (c) R=0.7 Case (d) R=0.7 Case (a) R=0.5 Case (b) R=0.5 Case (c) R=0.5 Case (d) R=0.5 Fig ure 4 3 3 Heat required for Subsystem 1

PAGE 87

87 0 4 8 12 16 20 24 0 50 100 150 200 250 300 350 400 450 Time [hr]QH2 [kW] Case (a) R=0.7 Case (b) R=0.7 Case (c) R=0.7 Case (d) R=0.7 Case (a) R=0.5 Case (b) R=0.5 Case (c) R=0.5 Case (d) R=0.5 Fig ure 4 3 4 Heat required for Sub system 2 0 4 8 12 16 20 24 0 50 100 150 200 250 300 350 400 450 Time [hr]QH3 [kW] Case (a) R=0.7 Case (b) R=0.7 Case (c) R=0.7 Case (d) R=0.7 Case (a) R=0.5 Case (b) R=0.5 Case (c) R=0.5 Case (d) R=0.5 Fig ure 4 3 5 Heat required for Sub system 3

PAGE 88

88 0 4 8 12 16 20 24 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 Time [hr]Efficiency Ratio R,b/R,a R,c/R,a R,d/R,a Fig ure 4 3 6 Efficiency ratio based on Case (a) efficiency

PAGE 89

89 CHAPTER 5 DYNAMIC MODELING WIT H CONSTANT AMMONIA MASS FRACTIO N APPROACH As discussed previous section the potential for the combined cycle PoWER plant in a configuration which allows ice making for the purpose of load leveling was explored. However, the refrigeration system model used was based on the s econd l aw efficiency, which was parametrically varied in order to explore the potential as a function of the technology maturity and complexity in the VARS. In this research the focus is on system performance using a specific design of a stateof the art VARS designed to produce refrigeration at two temperatures. The primary evaporator operates above freezing, cooling the gas path as well as a local air conditioning load. During the summer peak electrical demand in hot climates, the nighttime exces s refrigeration capacity is used to produce ice in a closed thermal storage subsystem. The ice would then be available during the afternoon demand peak to provide supplemental air conditioning, thus providing a demandside management benefit. Thermodynamic Model Schematic diagrams of the gas turbine and VARS portion of the PoWER system are shown in Fig ure 4 1 and Fig ure 5 1 Air enters the system at State Point 1 and is compressed by the low pressure compressor (LPC). It is adiabatically mixed with the rec irculated combustion products from the recuperator at State Point 3. The mixture is then allowed to enter the intercooler (warm gas cooler, WGC) before passing on to the evaporator of the VARS. In the evaporator (EVAP) 1, the mixture is cooled and water pr oduced as a product of combustion is extracted. The mixture then enters the highpressure core where it is compressed in the high pressure compressor (HPC), heated in the recuperator and combustor (CM), and then expanded in the high pressure turbine

PAGE 90

90 (HPT). After leaving the HPT, all of the combustion gases enter the recuperator where they lose heat to the gases entering the CM. At the exit of the recuperator, a portion of the gas recirculates to join the compressed air stream, and the remaining air is passe d through a low pressure turbine (LPT) before exiting to the atmosphere. The recirculating gas passes the generator (hot gas cooler, HGC) of the VARS. The VARS (Fig ure 5 1 ) in this study is an extended ammoniawater singleeffect system. Part of the hot gas from the recuperator passes through the heat recovery vapor generator (HRVG), commonly called the generator. Liquid ammoniawater strong solution (State Point 19) from the solution heat exchanger (SHX) is heated and flows to the rectifier (RECT). The lat ter produces nearly pure ammonia vapor (State Point 1) by using heat from the HRVG and cold strong solution (State Point 16). The almost pure ammonia vapor is cooled in the condenser (COND), which uses chilled water as a coolant, and in the refrigerant heat exchanger (RHX) 1. The liquid phase refrigerant passes through the thermal expansion valve (TXV) 1, then, it flashes into a two phase fluid at the intermediate pressure. The refrigerant is used both at EVAP1 to cool the HPC inlet and at EVAP2 as an air c onditioner during typical summer days. The twophase refrigerant leaving both EVAP1 and EVAP2 is cold enough to regeneratively cool the refrigerant leaving the condenser (State Point 2). Some refrigerant is further cooled through RHX2 and TXV2. This lowest temperature refrigerant is used to make ice. The return refrigerant, which is high quality or superheated vapor, mixes with strong solution at Absorber (ABS) 1 and ABS2 to produce week solution before entering the pump. When the air conditioning load is r educed, some refrigerant is further expanded and supplied to EVAP3, which operates as an ice maker. During nighttime, this

PAGE 91

91 evaporator makes ice for use as a coolant for the air conditioner at very high loads when EVAP2 could not meet the load on its own. A lmost saturated vapor, but still twophase refrigerant (State Points 6 and 83) leaves RHX1 and RHX2 to mix with refrigerant at State Points 24 and 114 from the SHX and ABS1, respectively. The saturated liquid strong solution (State Point 141) at the low pr essure of ABS2 is pumped and distributed to the rectifier (RECT) and the SHX. The design values of all state points are shown in Table 5 1 The following are the performance equations for each component considering mass, ammonia mass, and energy balances. Heat Recovery Vapor Generator 2019 HRVG HRVGQ mm q ( 5 1) 20201919 mcmc ( 5 2) 202019()HRVGQmhh ( 5 3) The HRVG is the only component that receives waste heat directly from the gas turbine system. Once HRVGQ [kW], the amount of heat input from the gas turbine system, is known, the mass flow rate of the solution can be calculated based on HRVGq the specific heat transfer. Rectifier 1221620mmmm ( 5 4) 11222216162020mcmcmcmc ( 5 5) 11222216162020mhmhmhmh ( 5 6)

PAGE 92

92 Condenser, Absorber, and Solution Heat Exchanger The COND, ABS1 and ABS2 use chilled water as a coolant. Even though the mechanical structure of typical absorbers is different from that of typical con densers, simple heat exchangers are applied for ABS1 and ABS2 by using mixing junctions before the components, yielding an equivalent thermodynamic process. The variations of the properties of State Points 18 and 22 are small, given the assumption that the operation of the VARS is well controlled at those points. So, these four components are considered as simple steady state heat exchangers. iomm ( 5 7) iioomcmc ( 5 8) 2,,,,()()jiiowipHOwowiQmhhmcTT ( 5 9) 181918222223()()SHXQmhhmhh ( 5 10) where the subscript i is 1, 13, 18, 22, 131, and the subscript o is 2, 14, 19, 23, 141 as input and output of the fluids, respectively. The subscript j indicates the components, COND, ABS1, and ABS2. The subscripts w,i and w,o are chilled water input and output, and these represent 7, 11, 111 as inputs, and 8, 12, 121 as outputs. The value of the specific heat of water is taken as Cp,H2O = 4.184 [kJ/kgK] Evaporator iomm ( 5 11) iioomcmc ( 5 12) ()jioiQmhh ( 5 13)

PAGE 93

93 where the subscript i is 41, 42, 63, and the subscript o is 51, 52, 73. The subscript j indicates evaporators, EVAP1, EVAP2, and EVAP3. Pump Because State Points 141 and 15 are always liquid, densities and specifi c volumes at those points can be considered constant. This gives the pump work as 14114115141()P PmvPP W ( 5 14) 14115141()PWmhh ( 5 15) where p is the assumed efficiency of the pump. Thermal Expansion Valve There are four TXVs to decrease pressure of the fluids in VARS. For an ideal expansion valve the following holds: iomm ( 5 16) iocc ( 5 17) iohh ( 5 18) The heading above shows that if you have a subheading of a certain level, you must have more than one. The rationale is that you cannot have a list of only one item. Refrigerant Heat Exchangers One of the assumptions in thi s study is that the system is well controlled. Therefore, steady state analysis can be reasonably used for most components. However, dynamic analysis is also needed for RHX1 and RHX2 because as the mass flow rate at State Points 5 and 73 changes based on t he air conditioning load, the

PAGE 94

94 temperatures of the refrigerant leaving the RHXs will change. While a quasi steady analysis may serve here for some purposes, we choose to lay the foundation for control algorithm development in future studies. General conserv ation equations [ 4 2 ] may be written as ()0 u t ( 5 19) () u uuf t ( 5 20) where f and are b ody force vector and stress tensor respectively. 2 2 uu e uu uh qfuuQ t ( 5 21) where q is the heat flux vector To simplify modeling, some reasonable assumptions are applied. The first well used approximation is the one dimensional approach. This leads the velocity vector to be a scalar quantity. Spatial variations of pressure relating to body forces and viscous dissipation will be neglected. So, there is no need to consider the momentum equation in the analysis. Sin ce convection effect is much greater than conduction, axial conduction will be neglected. Considering a simple shell and tube heat exchanger (Fig ure 5 2 ) with the above assumptions, and multiplying by the cross sectional areas, we obtain () 0 iiAm tx ( 5 22)

PAGE 95

95 ()() ()0ii iiiwAhmh HDTT tx ( 5 23) where i is t for tube or s for shell. Ht and Hs are the heat transfer coefficient of the working fluids in the tube and shell, respectively. As a first approximation, spatial temperature gradient of th e wall is ignored as shown in Fig ure 5 3 ()w pwT AC t ()()tttwssswHDTTHDTT ( 5 24) 22() 4w w stV ADD x ( 5 25) For simplicity, a lumped heat exchanger model is used with the assumption that the working fluid is well mixed. Th erefore, the properties of the working fluid in a control volume are the same as those at its outlet. Because analytical methods can be used only in special cases, a numerical approach is illustrated in this research [ 4 3 ]. A numerical approach always requi res that either x t or both x and t be discretized. Here both x and t are discretized to apply a trial and error approach, yielding the following for RHX1 11 32 33{(0)}tAL mm t ( 5 26) 2233 13 111 w ttmhmh TT HDL 13333 11{(0)(0)}t ttAhh HDt ( 5 27) 1131 61 11()ttw w ssHDTT TT HD 1 11 11() {(0)}Pw ww ssAC TT HDt ( 5 28)

PAGE 96

96 11 65 66{(0)}sAL mm t ( 5 29) 5566 61 111 w ssmhmh TT HDL 16666 11{(0)(0)}s ssAhh HDt ( 5 30) For RHX2, the subscripts 2, 3, 5, 6, t1, s1, and w1 can be simply replaced to 43, 53, 73, 83, t2, s2, and w2 from above equations and their control volumes are shown at Fig ure 5 4 Ice Maker and Storage Heat gain to the ice storage unit is considered to be a part of the cooling load in this work. When the air conditioning load is less than or equal to the maximum cooling capacity of EVAP2, the ice maker produces ice while the accumulated ice melts when the air conditioning load exceeds the maximum cooling capacity of EVAP2. And the rate of producing ice is 3, EVAPACice ice latent QQ dm dtq ( 5 31) where mice is the mass of ice [kg] and qlatent is latent heat of fusion of water [kJ/kg]. 3 EVAPQ is heat transfer rate of EVAP3 and ACiceQ is additional air conditioning load, which exceeds the maximum cooling capacity of EVAP2 during daytime. Solution Method A computational model described previously by Khan et al. [ 30] was written in FORTRAN to simulate the performance of the PoWER system using traditional cycle analysis methods. The model is capable of calculating eng ine performance (net work output, water extraction flow rate, and thermal efficiency) as a function of the input

PAGE 97

97 parameters (turbine inlet temperature, recuperator inlet temperature, generator temperature, evaporator exit temperature, low pressure compress or ratio, turbomachinery efficiencies, heat exchanger effectiveness, equivalence ratio, fuel type, and pressure drops). The input data to the model were appropriate for a small to medium size gas turbine system. To fix the operating point in the gas path, the combined cooling capacity of the HGC ( HRVG), the WGC ( intercooler), and the CGC ( EVAP1) has been held constant, even though the individual cooling capacity of each component is allowed to change. For the VARS and air conditioner, a newly developed transient computational model was written in Simulink as shown in Figure 5 5 This model uses typical summer air conditioning hourly load data as the main input (Fig ure 5 6 ) [ 33 ] and assigns the required cooling quantities to EVAP2 and EVAP3. It also calculates the mass of ice storage at each time step. To determine the properties of every state point of the VARS, the equations presented in the previous section were used. From the discretized dynamic equation, two T6s for RHX1, and two T83s for RHX2 are closely matched by using a trial and error approach. The values of the mass flow rate, temperature, pressure, concentration, enthalpy, entropy, specific volume, and quality at each state point were then found as a function of time. During the simulation, some pro perties (temperature, pressure, concentration, and enthalpy) of the ammoniawater mixture were calculated by using the method described by Ptek and Klomfar [ 6 ]. Also, the method described by Ibrahim and Klein [ 16 ] was used for calculating entropy and spec ific volume (or density).

PAGE 98

98 Results and Discussion Two VARS modes were considered, the higher air conditioning load mode (maximum EVAP2 cooling capacity), and the lower air conditioning load mode (maximum ice making mode). During daytime a nd early night (11:00AM ~ 9:00 PM ), the remaining refrigerant (50% of mass at State Point 41) flows through the valve at State Point 42 to EVAP2 while the valve at State Point 43 is fully closed. At the same time, the stored ice helps EVAP2 to meet the required air conditioning load. The data at each state point during this mode are shown at Table 5 2 Saturated liquid refrigerant goes to both EVAP1 and EVAP2, twophase refrigerant cools down the fluid from COND at RHX1. Finally, this refrigerant, whose temperature is still l ower than that of fluid from COND (State Point 2), becomes high quality twophase flow at the end of the RHX1 (State Point 6). During this mode, the ice maker stops making ice and the ice is melted to produce additional coolant to accommodate the additional air conditioning load (see Fig ure 5 7 ). When the air conditioning load is minimum (3:00AM ~ 6:00AM ), the values of each state point are shown at Table 5 3 At this time, the mass flow rate of refrigerant to EVAP2 is smaller than that of EVAP3. For this r eason, the cooling effect at RHX1 is reduced, and the temperature of State Point 3 is higher than that of during daytime. Even though the VARS cannot produce fully saturated liquid after the TXV, low quality two phase refrigerant has almost the same temper ature in both EVAP1 and EVAP2 because the state of the refrigerant at the exit of the two evaporators still has a low quality. This two phase refrigerant has enough cooling capacity at RHX1 to operate the VARS appropriately, and at State Point 6 this refri gerant is either saturated vapor or

PAGE 99

99 superheated vapor. Notice that the temperature of State Point 6 is lower than the temperature of State Point 2. Similarly, saturated liquid or subcooled liquid refrigerant leaving RHX2 is in the two phase or saturated liquid regime before EVAP3. This fluid leaves as a high quality two phase fluid at the exit of EVAP3 (State Point 73) and as saturated vapor at the exit of RHX2 (State Point 83). The COP of the lower air conditioning mode is slightly higher than that of th e other mode. Similarly, the differences of all other quantities except those of EVAP2, EVAP3, and ABS1, ABS2 for both modes are negligible. It is noticed that the sum of HRVGQ 1 EVAPQ 2 EVAPQ and 3 EVAPQ is approximately equal to the sum of 1 ABSQ 2 ABSQ and CONDQ for both modes with under 1% error. This result shows the significant role that the VARS coolant plays. As shown in Fig ure 5 8 12.99 kg/s of chilled water is supplied to the COND, ABS1, and ABS2. If ambient temperature air or water were used as the coolant at the condenser and absorbers instead, high capacity compressors or pumps would have been ne eded, and it would not have been appropriate to neglect this power loss. In Fig ure 5 9 and Fig ure 5 10 the heat gained from the surroundings at the three evaporators and their COPs as functions of time are shown. Both figures have similar trends as expect ed. Because the effect of the outdoor temperature on the PoWER system could be ignored due to the temperature control provided by the VARS, the EVAP1 cooling capacity and COP were unchanged during any one day. As mentioned above, during the daytime, all re frigerant flows to EVAP1 and EVAP2, so the COP of EVAP2 has the highest values at that time. Except for that time, the air conditioning load

PAGE 100

100 changes with time. Therefore, the heat gain at EVAP2 and EVAP3 and their COPs change as functions of time. These variations make the properties of refrigerant before each evaporator change. Figure s 5 11 and 5 12 show the temperature differences at RHX1 and RHX2. Because the VARS in this paper is assumed to be well controlled, the temperature of the fluid leaving the CO ND is always fixed as shown at Fig ure 5 11. During late nighttime, T3 is higher than that of the daytime because the mass flow rate at State Point 5 is decreased. However, the state of the fluid at State Point 5 before the RHX1 is saturated and not very hi gh quality, so, T5 is almost constant all day. Unlike T5, T6 changes because the ammoniawater mixture at State Point 6 has high quality even though it is saturated or because it is vapor. RHX2 behaves in a similar manner. The temperature difference between T43 and T73 (the input temperatures of RHX2) is smaller than that of RHX1. Some daytime data are meaningless because EVAP3 is not used at that time as shown in Fig ure 5 12. EVAP3 turns on around 10:00 PM and because there is no flow at State Point 73 at that moment, T53 increases instantaneously and has a higher value than the design value. Consistency of First Law and Second Law Models The results from the cycle analysis in this chapter are expected to agree exactly with the results from the Second Law analysis in Chapter 4. To verify this, three cases are considered at 4AM, 14PM, and 24A M. Based on three evaporator temperatures HRVG temperature (Table 51), and ambient temperature (Figure 4 5), Carnot COPs are calculated for each evaporator Then the s econ d law efficiencies, R, are determined by Equation (48). When these numbers are used as inputs for the Second Law model

PAGE 101

101 in Chapter 4, we can determine the evaporator heat inputs, EiQ and the refrigeration ratios Revp, (Equation 4 9) for three differ ent cases. Compared to the constant refrigeration ratio, 0.5, for the cycle analysis in this chapter, calculated refrigeration ratios show the similar results with less than 1% error as shown in Table 54. Therefore the Second Law analysis is determined to represent a valid and useful approach for preliminary design, when the details of the VARS have not yet been specified. Summary A steady state gas path model and a dynamic VARS model were used to simulate the thermodynamic performance of the PoWER cooling and power cycle, which combines a semi closed cycle gas turbine with a VARS in a novel way. Ice storage has been quantified over a typical daily variation in ambient conditions and building air conditioning load in order to show the effect on load level ing. As seen, the potential for reduction of the summer peak capacity requirement is significant. Several design tradeoffs are possible in optimizing the PoWER system for particular applications. If the HRVG capacity is higher, EVAP1 can be downsized, and the additional heat from the HRVG can be better utilized in other applications, such as air conditioning and/or ice making. The ice produced was used to meet the excessive air conditioning load at EVAP2 during the daytime. By doing so, part load operation of the total system can be eliminated, potentially providing economic and reliability advantages. The amount of ice available is ideally capable of producing as much as the total daily cooling capacity of EVAP2. This cooling capacity could be used to cool a second building of a size similar to the design building, or be used to downsize the entire system (thus reducing capital costs). If the air conditioning load is limited such as

PAGE 102

102 during a cool day, much more ice can be produced, and it can also be used for other applications. Considering the combination of attributes mentioned above, it can be concluded that the PoWER system seems to be a promising candidate for distributed power generation, especially in hot climates where the summer loadleveling benefit s are important. I t is also shown that the S econd Law analysis is the most convenient approach and at the same time is effective for preliminary design, accomplished by comparing the results between the conceptual Second Law based model in Chapter 4 and the specific cycle model in this chapter.

PAGE 103

103 Table 5 1. Design values or states at each state point Parameters Values and states State Point 1 temperature [K] 326.254 State Point 1 mass fraction of ammonia [%] 99.8 State Point 1 state Sat. vapor Mass fra ction; m 1 /m 20 0.285 State Point 2 state Sat. liquid State Point 4 pressure [kPa] 521 EVAP1 specific heat gained [kJ/kg] 1100 EVAP1 temperature [K] 278.150 EVAP2 specific heat gained [kJ/kg] 1100 EVAP2 temperature [K] 278.150 EVAP3 specific heat gain ed [kJ/kg] 1110 EVAP3 temperature [K] 268.150 Mass fraction; ( m 42 + m 43 ) / m 41 0.5 State Point 63 pressure [kPa] 268.107 State Point 24 pressure [kPa] 500.368 State Point 114 pressure [kPa] 342.823 State Point 141 temperature [K] 303.256 State Point 15 pressure [kPa] 1792.637 Mass fraction; m 16 / m 18 0.782 State Point 19 temperature [K] 365.070 HRVG specific heat gained [kJ/kg] 440 HRVG temperature [K] 373.15 Supplied chilled water temperature (7, 11, 111) [K] 290 State Point 8 temperature [K] 310 State Point 12, 121 temperature [K] 300 Pump efficiency 0.5 Pressure drop at each heat exchanger [%] 2 Pressure drop at EVAP 3 [%] 1

PAGE 104

104 Table 5 2. Results for higher air conditioning load State Point Mass [kg/s] Temperature [K] Pressure [kPa] NH 3 mass fraction S tate 1 0.1888 326.2542 1723.6893 0.998 Sat. vap. 2 0.1888 316.0315 1689.2155 0.998 Sat. liq. 3 0.1888 298.8792 1655.4312 0.998 Liquid 4 0.1888 278.1543 521 0.998 Sat. liq. 5 0.1888 279.0330 510.5800 0.998 2 phase 6 0.1888 292.90 11 500.3684 0.998 2 phase 13 1.1788 337.9796 500.3684 0.4381 2 phase 14 1.1788 314.9600 490.3610 0.4381 Sat. liq. 15 1.1788 304.0197 1792.6368 0.4381 Liquid 16 0.5171 304.0197 1792.6368 0.4381 Liquid 18 0.6617 304.0197 1792.6368 0.4381 Liquid 19 0.66 17 365.0700 1756.7841 0.4381 Liquid 20 0.6617 408.0452 1721.6484 0.4381 2 phase 22 0.9900 376.2403 1723.6893 0.3313 Liquid 23 0.9900 335.3104 1689.2155 0.3313 Liquid 24 0.9900 335.3104 500.3684 0.3313 Liquid 41 0.1259 278.1543 521 0.998 Sat. liq. 42 0.0629 278.1543 521 0.998 Sat. liq. 43 0 51 0.1259 279.0330 510.5800 0.998 2 phase 52 0.0629 279.0330 510.5800 0.998 2 phase 53 0 63 0 73 0 83 0 114 1.1788 311.2367 343.8232 0.4381 2 phase 131 1.1788 311.477 0 347.3316 0.4381 2 phase 141 1.1788 303.2563 340.3850 0.4381 Sat. liq. HRVG [kW] 440 COND [kW] 215.6486 ABS1 [kW] 373.6421 EVAP1 [kW] 138.4798 EVAP2 [kW] 69.2399 EVAP3 [kW] 0 W P [kW] 4.0499 SHX [kW] 190.6864 ABS2 [kW] 62.4789 COP 1 0. 3147 COP 2 0.1574 COP 3 0 COP total 0.4721

PAGE 105

105 Table 5 3. Results for lower air conditioning load State Point Mass [kg/s] Temperature [K] Pressure [kPa] NH 3 mass fraction S tate 1 0.1888 326.2542 1723.6893 0.998 Sat. vap. 2 0.1888 316.0315 1689.21 55 0.998 Sat. liq. 3 0.1888 303.8135 1655.4312 0.998 Liquid 4 0.1888 278.1558 521 0.998 2 phase 5 0.1888 280.0389 510.5800 0.998 2 phase 6 0.1399 299.5383 500.3684 0.998 Sat. vap. 13 1.1299 340.8979 500.3684 0.3981 2 phase 14 1.1299 314.9600 490.3610 0.3981 Liquid 15 1.1788 304.0169 1792.6368 0.4230 Liquid 16 0.5171 304.0169 1792.6368 0.4230 Liquid 18 0.6617 304.0169 1792.6368 0.4230 Liquid 19 0.6617 365.0700 1756.7841 0.4230 Liquid 20 0.6617 409.9966 1721.6484 0.4230 2 phase 22 0.9900 375.3206 1723.6893 0.3313 Liquid 23 0.9900 334.1742 1689.2155 0.3313 Liquid 24 0.9900 334.1742 500.3684 0.3313 Liquid 41 0.1259 278.1558 521 0.998 2 phase 42 0.0140 278.1558 521 0.998 2 phase 43 0.0489 278.1558 521 0.998 2 phase 51 0.1259 280.0389 510.5800 0. 998 2 phase 52 0.0140 280.0389 510.5800 0.998 2 phase 53 0.0489 277.5856 510.5800 0.998 Sat. liq. 63 0.0489 268.0984 358 0.998 2 phase 73 0.0489 268.4098 354.42 0.998 2 phase 83 0.0489 272.5600 347.3316 0.998 Sat. vap. 114 1.1299 314.9600 343.8232 0. 3981 Liquid 131 1.1788 316.4753 347.3316 0.4230 2 phase 141 1.1788 303.2563 340.3850 0.4230 Liquid Q HRVG [kW] 440 COND [kW] 215.6486 ABS1 [kW] 305.5540 EVAP1 [kW] 138.4798 EVAP2 [kW] 15.4157 EVAP3 [kW] 54.3135 W P [kW] 4.0250 SHX [kW] 189.8940 ABS2 [kW] 131.0223 COP 1 0.3147 COP 2 0.035 COP 3 0.1234 COP total 0.4732

PAGE 106

106 Table 54. Cons istency of the results between Chapters 4 and 5. Time 4AM 14PM 24PM Evaporator 1 2 3 1 2 3 1 2 3 COP 0.315 0.035 0.123 0.315 0.157 0 0.315 0.070 0.088 COP carnot 3.073 3.073 1.925 1.643 1.643 1.180 2.778 2.778 1.788 R 0.102 0.011 0.064 0.192 0.096 0 0. 113 0.025 0.049 Heat input 138.5 15.42 54.31 138.5 69.24 0 138.5 30.83 38.76 R evp 0.503 0.500 0.502

PAGE 107

107 Fig ure 5 1. Extended vapor absorption refrigeration system of the PoWER system Fig ure 52. Simple shell and tube heat exchanger

PAGE 108

108 Fig ure 5 3 Lumped wall model A B Fig ure 54. Control volumes for A) RHX1 and B) RHX2 2 Qev2AC 1 QIce 14 114 TXV3 23 24 TXV2 53 63 TXV1 3 4 TXV 15 16 18 Solution Distributor 22 18 19 23 SHX 4 ACLoad 41 42 43 Refrigerant Distributor 43 73 53 83 RHX2 2 5 3 6 RHX1 20 16 1 22 RECT Qhrvg_gp 141 15 Pump 19 Qhrvg_gp 20 HRVG 63 73 QIce EVAP3 (ICE) 42 52 Qev2AC EVAP2 (A/C) 41 51 EVAP1 (GP) Tcws Chilled Water SupplyTemperature[K] mcw Chilled Water Return 7 1 2 8 COND 111 131 141 121 ABS2 11 13 12 14 ABS1 114 83 131 83 + 114 -->131 51 52 5 51+52 --> 5 24 6 13 24 + 6 --> 13 1 EV2_Load Fig ure 55. Extended VARS Simulink model

PAGE 109

109 0 4 8 12 16 20 24 0 10 20 30 40 50 60 70 80 90 100 Time [hr]A/C Load [%] Fig ure 5 6. Air conditioning hourly load 0 4 8 12 16 20 24 0 1000 2000 3000 4000 5000 6000 Time [hr]Mass [kg] Fig ure 57. Mass change of the stored ice [kg]

PAGE 110

110 0 4 8 12 16 20 24 0 2 4 6 8 10 12 14 Time [hr]Mass [kg/s] COND ABS1 ABS2 Total Fig ure 5 8 Chilled water u se [kg/s] 0 4 8 12 16 20 24 0 20 40 60 80 100 120 140 160 Time [hr]Qevp 1,2,3 [kW] EVAP1 EVAP2 EVAP3 Fig ure 59. Heat gained from surroundings at three evaporators

PAGE 111

111 0 4 8 12 16 20 24 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 Time [hr]COP EVAP1 EVAP2 EVAP3 Total Fig ure 510. Coefficient of performance ( COP ) 0 4 8 12 16 20 24 275 280 285 290 295 300 305 310 315 320 Time [hr]Temperature [K] T2 T3 T5 T6 Fig ure 511. Temperature of RHX1

PAGE 112

112 0 4 8 12 16 20 24 266 268 270 272 274 276 278 280 Time [hr]Temperature [K] T43 T53 T73 T83 Fig ure 512. Temperature of RHX2

PAGE 113

113 CHAPTER 6 DYNAMIC MODELING FOR TWO COMPONENT FLUIDS A dy namic model of a special case, constant ammonia mass fraction, has been presented in the previous chapter Even though the fluids in the VARS dynamic model are two com ponent (ammoniawater), they were treated as single component fluids in Chapter 5. Further, the time discretized dynamic model has the limitation that it cannot be applied to conventional control theory. In this chapter, a dynamic model is derived to satisfy both approximate two phase twocomponent fluid physics and the requirements of the int ended control application. Introduction to Dynamic Heat Exchanger Model ing There are various heat exchanger mathematical models in the literature. Among them, only two phase dynamic models are discussed and then adapted to meet the goal of this research. M oving Boundary Model for Single Component Fluid To model tw o phase flow in heat exchangers such as evaporator s and condenser s, the conservation equations are expressed as partial differential equations (PDEs). Complex PDEs sometimes cannot be solved analyt ically and are usually difficult to be transformed into ordinary differential equations (ODEs). F or heat exchanger applications, PDEs can be spatially discretized, but the discretized models are of high order and do not lend themselves to further model reduction. Therefore, models of this type are not well suited for control design. A moving boundary model is one of the better alternatives for two phase flow modeling in heat exchangers For system level modeli ng, moving boundary models represent the physics of two phase flow well and satisf y the simplicity requirement

PAGE 114

114 Although various mathematical model variants are necessary based on boundary conditions, the resulting model s are generally fast enough to provide real time results compared to discretized models and are robust relative to sudden changes in the boundary conditions. Based on the standard physical laws of compressible fluid mechanics, mass, momentum, and energy balance equations are derived as shown in Chapter 5 (Equations 519 to 521) To simpl ify these equations, several assumptions are applied. The heat exchanger is idealized as a long, thin, horizontal tube; the flow is considered onedimensional ; axial heat conduction is neglected. Compared with the total absolute pressure, the pressure drop along the heat exchanger tube due to momentum change in the refrigerant and viscous friction is typically small and is neglected. Therefore, refrigerant pressure can be assumed uniform along the heat exchanger and the momentum equation is no longer necess ary. The resulting set of PDEs is as follows: Mass balance: 0 u tz (6 1) Energy balance: 4iwr ihPuh TT tzD (6 2 ) In the two phase section, the refrigerant temperature is at its saturated value with no spatial variation. F or the evaporator shown schematically in Figure 61, the temperature of the refrigerant in the superheated section increases as it travels toward

PAGE 115

115 the evaporator outlet due to the singlephase region. F or this situation including temperature variation, we l ump the evaporator dynamics into two nodes: the twophase node and the superheated node. The key feature of the conventional moving boundary twophase model is an assumption of a timeinvariant mean void fraction. Wedekind et al. [ 23] showed that the mean void fraction remains relatively invariant in the twophase region of a heat exchanger during most regimes of operation. This implies that under different inflow conditions, the liquid dry out point in an evaporator may change its location along the evapor ator; however, the distribution of the liquid and vapor remains similar at all time s. T o prepare for control algorithm development, it is necessary to create ODEs from the onedimensional PDEs (61) and (62 ) for both control volumes in Figure 61. Following the principle of timeinvariant mean void fraction, Equations (61) and (62 ) can be integrated from z= 0 to l1( t ) for two phase refrigerant. (see Appendix C for derivation in detail). Mass balance for twophase region: 11 11 iintddl AlAmm dtdt (6 3 ) Energy balance for twophase region: 11 1 1 111 1111 iiintgiiwrdhdl dP AlAhAlmhmhDlTT dtdtdt (6 4 ) where A is the crosssectional area of heat exchanger tube, l1 i s the length of two phase region, im and intm are the mass flow rat e at heat exchanger inlet and interface,

PAGE 116

116 respectively. hg is the enthalpy at the interface and t his value is same as that of saturated vapor of refrigerant. He et al. [ 21 ] used absolute speed for the mass flow rate at the interface. However, the interface between twophase and vapor fluids moves back and forth, so the mass flow rate of the refrigerant at interface, intm is always related to the relative speed (Equation (C 4)). The Equations (63) and (64) are corrected equations fro m He et al. [ 21 ]. The mass and energy balances for superheated vapor refrigerant can be obtained by integration from z = l1( t ) to L (see Appendix C for derivation). Mass balance for superheated vapor region: 21 22 intoddl AlAmm dtdt (6 5 ) Energy balanc e for superheated vapor region: 22 1 2 222 2222 intgooiiwrdhdl dP AlAhAlmhmhDlTT dtdtdt (6 6 ) where l2( t ) is the length of superheated vapor region, om is the mass flow rate at heat exchanger exit ho is refrigerant exit enthalpy. The energy equations for the tube wall of control volumes 1 and 2 are 1 111 1 w p iirwooaw wdT CADTTDTT dt (6 7 ) 212 1 222 1 2 www p iirwooaw w dTTT dl CA DTTDTT dtldt (6 8 )

PAGE 117

117 The mean values of density and enthalpy in twophase region are functions of pressure only For superheated vapor refrigerant, however, t hey are functions of pressure and temperature or the density and the temperature are functions of pressure and enthalpy It is assumed that the bulk enthalpy of superheated vapor region is half of enthalpies at interface and exit. After cancelling out intm (Appendix C) we obtain the final form of conservation equations as matrix form: -1xDfx,u (6 9 ) where 111213 2122 313233 44 51 5500 000 00 0000 000 ddd dd ddd d dd D 1 2 3 4 5f f f f f fx,u 1 2 1 2 r w wl P T T T x and i i om h m u x is the vector of state variables and u is the vector of inputs ( The coefficients of matrices D and f are indicated in Appendix C ). Two C ompone nt F luids In this research ammonia water mixtures are used as a working fluid in vapor absorption refrigeration system rather than a pure component as in a simple vapor compression refrigeration system. T herefore, some extensions are necessary to model tw o component twophase flow in a heat exchanger. First a new state variable, concentration, is necessary Because almost pure ammonia is used as a refrigerant, ammonia mass fraction is equivalent to a concentration. Unlike a pure component flow heat exchanger the temperature decrease s for a condenser and increases for an evaporator in the two phase region, even at uniform

PAGE 118

118 pressure. Figure 62 shows the two phase flow temperature increase in an evaporator for several different pressures. At the same time, t he ammonia mass fraction in the vapor and liquid drops because ammonia is more volatile than water. (Figure 63 and 64) Densities and enthalpies for saturated vapor and liquid are functions of pressure, temperature, and ammonia mass fraction, so saturatio n properties are not uniform along the heat exchanger The m ost important change is that the mean void fraction is no longer uniform because ammonia concentration is involved as well. When vapor quality is fixed in two phase flow, the density ratio determi nes mean void fraction. Density ratio can be considered as a constant in the pressure range of inter e st However, it is affected by ammonia concentration. Two Phase Model Options For control oriented dynamic modeling of a VARS, the conventional moving boundary two phase model is extended to include two component fluids. Species balances are added for the conservation equations. Thermodynamic properties in the two phase region are no longer uniform even at constant pr essure along the heat exchanger To model this complex heat exchanger using lumped parameters, three approaches were considered: 1. Uniform mass fraction model; 2. Thermodynamic equilibrium model; 3. S patial mean value model. For the first two models, both vapor and liquid in the two phase region are considered to be saturated. It is assumed that bulk ammonia mass fractions of saturated vapor and liquid are the same for uniform mass fraction model while temperature is uniform along the entire heat exchanger for the thermodynamic equilibrium model. At steady state, saturated vapor and saturated liquid have different

PAGE 119

119 bulk temperatures when the uniform mass fraction model is applied. For the thermodynamic equilibrium model, however, saturated vapor and liquid have different ammonia mass fractions. Fig ure 65 shows two very limited cases of the flow state on enthal py versus ammonia mass fraction (h x) diagram at fixed pressure. The h x diagram is based on Bogart s T P x tables [ 12] and Ibrahim s enthalpy relations [ 15] at 1,000 kPa. The upper and lower curves are saturated vapor and saturated liquid enthalpy lines, respectively. The dotted lines between the two curves indicate isothermal line in the two phase region. The solid vertical line is based on the uniform mass frac tion model and the inclined sol id line is based on thermodynamic equilibrium model. When average enthalpy, havg, and bulk ammonia mass fraction, xb, are the same for the two models shown in the F igure 6 5 the saturated vapor temperature, Tsg, is always higher than the thermodynamic equilibrium temperature, Te, and the saturated liquid temperature Tsl, is lower than that for uniform mass fraction model. Ammonia mass fraction in saturated vapor xeg, is higher than that of saturated liquid, xel, for thermodynamic equilibrium model. Because of those behaviors, the two models show different vapor qualities and void fractions except for the pure component region. This difference makes it difficult to apply the void fraction model to both models. To avoid this problem, using spatial ly averaged values of thermodynamic properties in the two phase region is the most appropriate method of the three. The spatial mean values are determined based on the assumptions, which are: pressure and ammonia mass fraction are uniform along the heat exchanger a nd each local infinitesimal control volume is in thermodynamic equilibrium.

PAGE 120

120 With these assumptions, temperature distribution information is available in the two phase region (Figure 6 2) and densities and enthalpies can be expressed spatially at fixed pre ssure and ammonia mass fraction. These spatial distributions of thermodynamic properties can be tabulated or can be curvefit as functions of pressure, ammonia concentration, initial conditions and boundary conditions. However, fitted equations may encounter R unge s phenomen on, which is discussed in Chapter 3. Therefore, tables for thermodynamic property data in the two phase regime are used in this research. Thermodynamic M odels of VARS for T wo C omponent F luids In this research, a spatial ly mean value model is used with tabulated twophase thermodynamic property data to model and solve the conservation equations. This approach represents twocomponent twophase flow physics relatively well and meets control theory requirements at the same time. With this approach, some modifications are necessary related to conventional moving boundary models [21, 22] For the system model in this research, pressure is determined from the system model Therefore, pressure is an input for all heat exchangers in the VARS. Eithe r inlet or outlet mass flow rate is an output. From the schematic diagram of the VARS shown in Figure 51, there are six different heat exchangers; condenser (COND), absorber (ABS), generator (HRVG), evaporator (EVAP), refrigerant heat exchanger (RHX), and solution heat exchanger (S HX). Among them RHX has similar fluid condition to that of the EVAP and RHX has not been modeled separately.

PAGE 121

121 Condenser For general refrigeration applications high pressure superheated vapor is condensed to subcooled liquid in condensers for normal operation modes. Therefore, three different phases can exist in condensers the superheated, twophase and subcooled regions. However, in this research, only twophase and subcooled regions are considered because theoretically only saturated vapor exits from the top of rectifier when there is no pressure drop in connecting ducts b etween rectifier and condenser and modeling of the single phase region is straightforward. Because saturated vapor enters the condenser, the full two phase region always exists in that device As a secondary fluid (coolant), sub cooled liquid is assumed, though this choice does not directly enter the analysis A schematic of a condenser is shown in Figure 66 illustrating the twophase region (on the left) and the subcooled region (on the right). General conservation equations are derived next based on two control volumes. B asic assumptions used for conventional moving boundary models are applied, so the equations (61) and (6 2 ) are used for mass balance and en ergy balance, respectively. H eat input, which is the right hand side of E quation (6 2 ) is expressed as heat flux, Q taken here as approximately constant. For a two component fluid, a species balance for ammonia concentration is needed as an additional conservation equation: Species balance: 0 xux tz (6 10) where x is ammonia mass fraction.

PAGE 122

122 To obtain ODEs of the conservation equations, the above equations are integrated along the control volume length. Af ter applying Leibniz rule, the follow ing six equations are obtained: Mass balance f or refrigerant in control volume 1: 11 1 1 12 cc cc cc cicddl AlAmm dtdt (6 11) Species balance for refrigerant in control volume 1: 11 1 1 11 1212 cc c cc ccc ciciccdxdl AlAxmxmx dt dt (6 12) Energ y balance for refrigerant in control volume 1: 11 1 1 11 1 1212 111 1 cc c c cc ccc cc cicicc ccacr cdhdldP AlAhAlmhmhUDlTT dt dtdt (6 13) Mass balance f or refrigerant in control volume 2: 21 2 2 12 cc cc cc ccoddl AlAmm dtdt (6 14) Species balance for refrigerant in control volume 2: 22 1 2 22 1212 cc c cc ccc cccocodx dl AlAxmxmx dt dt (6 15) Energy balance for refrigerant in control volume 2:

PAGE 123

123 22 1 2 22 2 1212 222 2 cc c c cc ccc cc cccoco ccacr cdh dldP AlAhAlmhmhUDlTT dt dtdt (6 16) w here (UD)c1 and (UD)c2 are universal heat transfer coefficient per length for control volume 1 and 2, respectively. Six base conservation equations ( Equat ions 6 11 to 6 16) are used for the ammoniawater mixture (refrigerant) side in the condenser and additional conservation equations are necessary for the coolant ( secondary fluid). It is assumed that the coolant is i ncompressible and has a constant heat ca pacity. Therefore, mass balances for two control volumes are trivial because densities are the constant. Then, only the energy balance remains: Energy balance f or coolant in control volume 1: 1 1 21 111 1 ca pc capcacacao ccrca c cadT AClmCTTUDlTT dt (6 17) Energy balance for coolant in control volume 2; 2 2 21 222 2 ca pc capcacaica ccrca c cadT AClmCTTUDlTT dt (6 18) For the refrigerant, mass balances can be substituted for species and energy balances and the outlet mass flow rate is d etermined by the fluid con dition in the condenser at the exit Then we have four remaining equations: Species balance for refrigerant in control volume 1: 11 1 1 1 12 11112 12 cc c c cc c ccccc cicicdxd dl AlxAxxmxx dtdt dt (6 19)

PAGE 124

124 Energy balance for refrigerant in control volume 1: 11 1 1 1 12 11112 12 111 1 cc cc c cc c ccccc cicic ccacr cdhddP dl Alh Ahh dtdtdt dt mhhUDlTT (6 20) Species balance f or refrigerant in control volum e 2: 1 22 2 112 2 1 112 222 12 c cc c cccco cc co c ccccocccco ciccoddxd AlxxAl x dt dtdt dl Axxxxmxx dt (6 21) Energy balance for refrigerant in control volume 2: 1 22 2 112 2 1 112 222 12 222 2 c cc cc cccco cc co c ccccocccco cicco ccacr cddhddP AlhhAl h dt dtdtdt dl AhhhhmhhUDlTT dt (6 22) Because ammonia mass fraction is assumed uniform, over bars on xc1 and xc2 are not necessary and similarly for the bulk tem perature of control volume 2 Tcr2. Now all three independent variables, pressure, temperature, and concentration are uniform rather than spatially distributed in control volume 2 so its densit y and enthalpy do not need over bars. For the coolant, it is assumed that both pressure a nd temperature are uniform also although the only fundamental requirement is that it provides nearly constant heat flux. At the interface between twophase and subcooled liquid, the ammonia mass fraction is assumed to be the s ame as twophase bulk ammonia mass fraction:

PAGE 125

125 121 ccxx (6 23) Then, E quation (619) is rewritten as 1 11 12 c ccc cicicdx Almxx dt (6 24) Mean values of density and enthalpy in the two phase region are functions of pressure and ammonia mass fraction : 1111 1 ccccc ccddPdx dtPdtxdt (6 25) 1111 111 1 cccccccc ccdhhdPhdx dtPdtxdt (6 26) Sub cooled liquid enthalpy and density are functions not only of pressure and ammonia concentration but also of temperature: 222222 22 cccccrcc c cr cddPdTdx dtPdtTdtxdt (6 27) 2222 222222 22 ccccccccrccc c cr cdhhdPhdThdx dtPdtTdtxdt (6 28) Then, the species equation for subcooled liquid is

PAGE 126

126 1 11 112 2 2 112 1 22 22 222 22 22 12 12 112 22 c cc ccccoccoc cccco c cc ccr ccccco cccco c cr c cc ciccocccco cccco cc dl dx AxxxxAlxx dt xdt dx dT Alxx Alxx xdt Tdt dP mxxAlxxAlxx P Pdt (6 29) Energy equations for both phases are rewritten after substitution of E quations (620) to (6 22). 1 11 11 11112 1 12 11 11 1 12 1111 12 11c cc cc ccccc cc c cc cc cc cicic ccacrcc c c ccdl h dx AhhAlh dtxxdt h dP mhhUDlTTAlh PPdt (6 3 0 ) 1 11 112 2 2 112 1 222 222 22 2 22 2 22 22 12 222 2 1 112 c cc ccccoccoc cccco c c cc c ccr cccco c cccco c c c cr cr cicco ccacr c c ccccodl dx AhhhhAlhh dt xdt hdx hdT Alhh Alhh xxdt TTdt mhhUDlTT Alhh 22 22 21 ccc cccco c c cchdP Alhh P PPdt (6 31) E quations (617, 18, 2 4 29 30 and 31) can be expressed in matri x form: ccccCxFxu (6 32)

PAGE 127

127 w here 12 21222324 3132 41424344 55 6600000 00 0000 00 00000 00000 c cccc cc cccc c c C 1 1 2 2 1 2 c c c cr ca cal x x T T T cx 1 2 3 4 5 6 c c c c c cf f f f f f cF and c c ci ci ci cai caP P x h m T m cu cx is the vector of state variables and cu is the vector of inputs for the condenser. The vector cF and t he six by six matrix C are comp osed of coefficients shown in Appendix C The derivative of the state variable vector cx is determined as -1 ccccxCFxu (6 33) Outputs are 1 T ccocococaolxhmT cy consisting of the condenser mass flow rate, a mmonia mass fraction and enthalpy at the exit of the refrigerant side outlet temperature at coolant side, and the length of the two phase region. Outlet mass flow rate is determined from the derivative of state variables and equations (611 ) and (6 14): 1 111 121 1 22222 2 22 c cccc cociccc cc cc ccccccr cc c c crdl dPdx mmA Al dtPdtxdt dPdxdT Al PdtxdtTdt (6 34) For single phase twocomponent fluid, exit ammonia mass fraction is same as bulk ammonia mass fraction of control volume 2. The bulk enthalpy is assumed to be the

PAGE 128

128 average of the control volume inlet and outlet enthalpies. Therefore, outlet ammonia mass fraction and enthalpy are 2 coc xx (6 35) 2122cocc hhh (6 36) Because coolant in the condenser is assumed to have constant density, temperature differences are in proportion to enthal py differences. S o, the bulk temperature of the coolant is also assumed to be the average of the inlet and outlet temperatures 1 211 2ca cacaoTTT (6 37) 2 211 2ca caicaTTT (6 38) 2122cacacaiTTT (6 39) 1212caocacaTTT (6 40) Absorber A conventional vapor compression refrigeration system uses a compressor to pressurize superheated refrigerant. Typical compressors need significant power to run the refri gerant system. However, a VARS uses an absorber and an energy efficient pump, which pressurize s liquid. Therefore, the absorber must produce a liquid absorbent refrigerant fluid before the pump. Superheat ed or two phase refrigerant is absorbed in a dilute solut ion in this device For an ammoniawater VARS, alm ost pure ammonia vapor or twophase refrigerant is absorbed in a weak solution. To maximize

PAGE 129

129 the interface area of vapor and liquid, the weak liquid solution is sprayed into the vapor or vapor is injected into the liquid solution. In this research, however, the absorber is treated as an adiabatic mixing step, followed by a simple tube and shell heat exchanger. Twophase nearly pure ammonia and a weak liquid solut ion m ix before the absorber and their mixture, a strong solution, is cooled by the coolant A s chematic of an absorber is shown in Figure 6 7 The absorber analysis is very similar to that of the condenser in this research. The o nly difference is that the inlet condition of the refrigerant side is not saturated vapor but two phase. Therefore, the conservation equations for the absorber are also similar to those of the condenser (Equations ( 6 24) ( 6 29) ( 6 30) and ( 6 31) ) : Species balance for refrigerant in control volume 1: 1 11 12 a aaa aiaiadx Almxx dt (6 41) Energy balance for refrigerant in c ontrol volume 1: 1 11 11 11112 1 12 11 11 1 12 1111 12 11a aa aa aaaaa aa a aa aa aa aiaia aaaaraa a a aadl h dx AhhAlh dtxxdt h dP mhhUDlTTAl h PPdt (6 42) Species balance for refrigerant in control volume 2:

PAGE 130

130 1 11 112 2 2 112 1 22 22 222 22 22 12 12 112 22 a aa aaaaoaaoa aaaao a aa aar aaaaao aaaao a ar a aa aiaaoaaaao aaaao aa dl dx AxxxxAlxx dt xdt dx dT Alxx Alxx xdt Tdt dP mxxAlxxAlxx P Pdt (6 43) Energy balance for refrigerant in control volume 2: 1 11 112 2 2 112 1 222 222 22 2 22 2 22 22 12 222 2 1 112 a aa aaaaoaaoa aaaao a a aa a aar aaaao a aaaao a a a ar ar aiaao aaaar a a aaaaodl dx AhhhhAlhh dt xdt hdx hdT Alhh Alhh xxdt TTdt mhhUDlTT Alhh 22 22 21 aaa aaaao a a aahdP Alhh P PPdt (6 44) To determine the mea n values of the twophase properties, a reference ( imaginary ) saturated li quid location needs to be found first. This imaginary moving boundary location is determined by inlet boundary conditions, pressure, and bulk ammonia mass fraction in the two phase f low If heat flux is uniform along the heat exchanger length, the specific enthalpy distribution is linear along the length. Based on pressure and bulk ammonia mass fraction, enthalpies for saturated vapor and liquid can be found. Then the specified inlet enthalpy determines the ratio of the dimensionless two phase region length to that based on imaginary boundary The distributions of temperature, density and enthalpy are determined based on pressure and bulk ammonia mass fraction. Their mean values are calculated from properties between the inlet and the moving boundary inside the absorber for the

PAGE 131

131 refrigerant side. For the coolant and subcooled refrigerant in the absorber the analysis is similar to th at of the condenser So, refrigerant mass fl ow rate is determined similarly : Energy balance f or coolant in control volume 1: 1 1 21 111 1 aa pa aapaaaaaao aaraa a aadT AClmCTTUDlTT dt (6 45) Energy balance for coolant in control volume 2; 2 2 21 222 2 aa pa aapaaaaiaa aaraa a aadT AClmCTTUDlTT dt (6 46) The matrix form of the conservation equations is aaaaAxFxu (6 47) w here 12 21222324 3132 41424344 55 6600000 00 0000 00 00000 00000 a aaaa aa aaaa a a A 1 1 2 2 1 2 a a a ar aa aal x x T T T ax 1 2 3 4 5 6 a a a a a af f f f f f aF and a a ai ai ai aai aaP P x h m T m au ax is the vector of state variables and au is the vector of inputs for the absorber. The vector aF and the six by six matrix A are composed of coefficients shown in Appendix C. The differentiation of state variables, x yields 1,aaaaxAFxu (6 48)

PAGE 132

132 Outputs are 1 T aaoaoaoaaolxhmT ay considering of the absorber mass flow rate, ammonia mass fraction and enthalpy at the exit of re frigerant side, outlet temperature at coolant side, and the length of twophase region. Th e o utlet mass flow rate is again determined by derivatives of state variables and mass balances: 1 111 121 1 22222 2 22 a aaaa aoaiaaa aa aa aaaaaar aa a a ardl dPdx mmA Al dtPdtxdt dPdxdT Al PdtxdtTdt (6 49) For single phase twocomponent fluid, exit ammonia mass fraction is the same as the bulk ammonia mass fraction of control volume 2. The bulk enthalpy is assumed to be the average of the control volume inlet and outlet enthalpies. Therefore, outlet ammonia mass fraction and enthalpy are 2 aoaxx (6 50) 2122aoaahhh (6 51) Because the coo lan t in the absorber is assumed to have constant density and heat capacity temperature differences are in proportion to enthalpy differences. So, the bulk temperature of coolant is the average of the inlet and outlet temperatures. 212 2 aaaaaai TTT (6 52) 1212aaoaaaaTTT (6 53)

PAGE 133

133 Generator A VARS may use various heat sources as an energy input such as power cycle waste heat as in this study This heat generates refrigerant vapor from a two component liquid in a (so called) generat or to provide the highquality energy input needed for a VARS. Unlike the first two heat exchangers, subcooled liquid enters and twophase fluid exits ; in the actual system, the phases would be subsequently separated. Figure 6 8 shows a schematic of a generator heat exchange process The equations for the generator are also very similar to those of the first two heat exchangers. For the generator, mean values are used for control volume 2 instead of control volume 1. The result ing equations are : Species balance for refrigerant in control volume 1: 1 11 1112 1112 1 11 11112 1 1 12 1112 g ggr gggg gggg gr gg ggggg g gg gigiggggg gdl dT AxxAlxx dt Tdt dx Alxx xdt dP mxxAlxx Pdt (6 54) Species balance for refrigerant in control volume 2: 1 11 112 112 1 11 2 112 22 1 1 12 112 g ggr ggggo ggggo gr gg g ggggo ggg g g a giggoggggo adl dT AxxAlxx dt Tdt dx dx Alxx Al xdt dt dP mxxAlxx Pdt (6 55)

PAGE 134

134 Energy balance for refrigerant in control volume 1: 1 111 1112 1112 1 11 111 1112 1 11 11 12 111 1112 1 11g g ggr gggggggg g gr gr ggg gggg g gg ggg gigig ggagrgggg g g ggdl hdT AhhAlhh dt TTdt hdx Alhh xxdt hdP mhhUDlTTAlhh PPd t (6 56) Energy balance for refrigerant in control volume 2: 1 11 112 222 112 1 11 22 22 112 2 1 22 12 222 2 1 22 112 2 g ggr ggggoggggo ggggo gr gg gg gg ggggo gg go g gg giggo ggagr g g gg ggggo gg gdl dT AhhhhAlhh dt Tdt dx h dx Alhh Al h xdt xxdt mhhUDlTT h AlhhAl PP 21 gg go ggdP h Pdt (6 57) Energy balance f or coolant control volume 1: 1 1 21 111 1 ga pg gapgagagao ggrga g ga dT AClmCTTUDlTT dt (6 58) Energy balance for coolant control volume 2: 2 2 21 222 2 ga pg gapgagaiga ggrga g ga dT AClmCTTUDlTT dt (6 59) The matrix form of the conservation equations is

PAGE 135

135 ggggGxFxu (6 60) w here 13 21222324 313233 41424344 55 6600000 00 000 00 00000 00000 g gggg ggg gggg g g G 1 1 1 2 1 2 g gr g g ga gal T x x T T gx 1 2 3 4 5 6 g g g g g gf f f f f f gF and g g gi gi gi gai gaP P x h m T m gu gx is the vector of state variables and gu is the vector of inputs for the generator. The vector gF and the 6 by 6 matrix G are composed of coefficients shown at Appendix C. The differentiati on of state variables, gx yields -1 ggggxGFxu (6 6 1) Outputs are 1 T ggolgoggolgoggolgogaaolxxhhmmT gy consisting of the generator mass flow rates, ammonia mass fractions and enthalpies in both saturated vapor and liquid a t the exit of re frigerant side outlet temperature at coolant side, and the length of singlephase region. To find the refrigerant outputs, the state of the refrigerant at the exit must be known. The saturated vapor position is after the outlet as shown i n Figure 610. There are two different ways to determine the exit One is to find the refrigerant mean temperature, and the other is to use the imaginary moving boundary position directly. For a generator, the first method is used whereas the second one i s used for an evaporator

PAGE 136

136 Finding the refrigerant mean temperature is similar to the method used in absorber. For a given instant in time, the two phase mean temperature is calculated from heat source thermodynamic properties and the universal heat transfer coefficient: 222 2 22gapga grgagagai g gmC TTTT UDl (6 62) Based on the average refrigerant temperature in control volume 2, the i maginary saturated vapor location, gL, can be found. Once gL is calculated the pr ofiles of e nthalpies and ammonia mass fractions in both saturate d vapor and liquid, as well as vapor quality (or dryness), dg( z ) in two phase region are determined with the assumptions that heat flux along the generator is constant, and thermodynamic equ ilibrium exists within each infinitesimal control volume. Outlet total mass flow rate is determined from derivatives of state v ariables and mass balances : 1 222 122 2 11111 1 11 g gggg gogiggg gg gg ggggggr gg g g grdl dPdx mmA Al dtPdtxdt dPdxdT Al PdtxdtTdt (6 63) From the total mass flow rate and quality distribution, vapor and liquid mass flow rates are determined at the exit: gogggomdm (6 64) 1gol ggomdm (6 65)

PAGE 137

137 Because the fluid heat source in the generator is assumed to have constant density and heat capacity temperature differences ar e in proportion to enthalpy differences. S o, the bulk temperature of the coolant is the average of the inlet and outlet temperatures 2122gagagaiTTT (6 66) 1212gaogagaTTT (6 67) Evaporator For most refrigeration systems using single component refrigerant, twophase, but almost saturated liquid or liquid is fully vaporized in evaporator to maximize cooling effect and to protect compressor from liquid droplets a fter evaporator. For an ammoniawater VARS, however, almost pure amm onia refrigerant always contains small amount of water because rectifier cannot make pure ammonia. Saturation temperature is not constant any more even with very small amount of water as shown in Figure 62. At very high ammonia mass fraction, 0.998, satur ation temperature can be treated as a constant for some region, but temperature rises up dramatically near saturated vapor area. To use lowest possible evaporator temperature, refrigerant is not fully vaporized intentionally in an evaporator for a VARS. Th erefore, fluid conditions of inlet and outlet of evaporator are two phase during normal evaporator operation. A schematic of an evaporator is shown in Figure 69 For the evaporator imaginary moving boundary position is directly used to get inlet and outl et conditions of the evaporator. Consider two control volumes, one is from z= 0 to eL and the other is from z=Le to eL eL is the location of imaginary moving boundary.

PAGE 138

138 Note that control volume 1 includes control volume 2. Conservations equations for the evaporator are determined as follows (see Appen dix C for detailed derivations). Mass balance for refrigerant : ** 12 12 ee e eee ee eee eieodLd d A ALALLmm dtdt dt (6 68) Species balance for refrigerant : ** 12 e eeeeee eieieodx ALLLmxx dt (6 69) Energy balance for refrigerant : 111222 ** 11 1 22 2 ** 11 1 221e eeeeeoeeoee ee e ee ee ee eo eee eo ee ee eieieo eaer e ee e ee ee eo eee eo ee edL Ahhhh dt h h dx ALhALL h xx xxdt mhhUATT hh ALh ALL h PP P 21ee edP Pdt (6 70) Energy balance f or the secondary fluid: ea pe eapeaeaieao erea e eadT ACLmCTTUATT dt (6 71) The matrix form of the conservation equations for the evaporator is eeeeExFxu (6 72)

PAGE 139

139 w here 12 2122 3300 0 00 e ee e E e e eaL x T ex 1 2 3 e e ef f f eF and e e ei ei ei eai eaP P x h m T m eu ex is the vector of state variables and eu is the vector of inputs for the evaporator. The vector eF and the 3 by 3 matrix E are composed of coefficients shown at Appendix C. The differentiation of state variables, ex yields -1 eeeexEFxu (6 73) Outputs are T eeoeoeoeaoLxhmT ey consisting of the evaporator mass flow rate, ammonia mass fraction, and enthalpy at the exit of refrigerant side, outlet temperature at coolant side and the imaginary boundary l ocation. Outlet mass flow rate is determined by derivatives of state variables and equations (668 ): ** 12 12 ** 12 ee ee eoeieee ee eee ee e ee ee eee eedL dx mmA ALALL dtx xdt dP ALALL P Pdt (6 74) Because the secondary fluid in evaporator is assumed to have constant density, temperature differences are in proport ion to enthalpy differences. S o, the bulk temperature of the secondary fluid is the average of the inlet and outlet temperatures

PAGE 140

140 2eaoeaeaiTTT (6 75) Solution H eat E xchanger The fluid from the bottom of rectifier is hot enough to warm up the fluid from pump to generator in SHX Therefore, heat energy required to run a VARS from generator can be reduced. The fluids through SHX are always subcooled liquid without phase change in the heat exchanger. Because there is no moving interface in SH X, only two control volumes are remained to be modeled. A schematic of SHX is shown Figure 6 10. Even though the fluid is subcooled liquid, species balance is necessary because the variance of ammonia mass fraction cannot be ignored. Mass balance for col der fluid: sc scs sciscod ALmm dt (6 76) Species balance for colder fluid: scsc sc sc scsscsco scsscsco sc sc sc scs scisciscoscsscsco sdT dx ALxx ALxx Tdt xdt dP mxxALxx Pdt (6 77) Energy balance for colder fluid: 1sc scsc sc scsc scsscsco sc scsscsco sc sc sc sc sc sc sc s sciscisco shscscsscsco sc s sshdT hdx ALhh ALhh TTdt xxdt hdP mhhUATTALhh PPdt (6 78)

PAGE 141

141 Mass balance for hotter fluid: sh shs shishod ALmm dt (6 79) Species balance for hotter fluid: shsh sh sh shsshshi shsshshi sh sh sh shs shoshishoshsshshi sdT dx ALxx ALxx Tdt xdt dP mxxALxx Pdt (6 80) Energy balance for hotter fluid: 1sh shsh sh shsh shsshschi sh shsshshi sh sh sh sh sh sh sh s shoshisho scshshsshshi sh s sshdT hdx ALhh ALhh TTdt xxdt hdP mhhUATTALhh PPdt (6 81) The matrix form of the conservation equations is ssSxF (6 82) w here 1112 2122 3334 4344 00 00 00 00 ss ss ss ss S sc sc sh shT x T x sx 1 2 3 4 s s s sf f f f sF The differentiation of state variables, sx are determined as: -1 ssxSF (6 83)

PAGE 142

142 Outputs of SHX are mass flow rate, ammonia mass fraction and enthalpy at the exit of colder refrigerant side and inlet mass flow rate and outlet ammonia mass fraction and enthalpy at hotter side. Outlet mass flow rate of colder refrigerant is determined by derivatives of state variables and mass balance: scsscscscsc scosciscs s sc scdPdTdx mmAL PdtTdtxdt (6 84) scoscxx (6 85) 2scoscscihhh (6 86) Inlet mass flow rate of hotter refrigerant is determined by derivatives of state variables and mass balance : shsshshshsh xhisoshs s sh shdPdTdx mmAL PdtTdtxdt (6 87) shosh xx (6 88) 2shoshshihhh (6 89) Results and Discussion C omputer codes implementing the model for the he at exchangers as described above are now used to calculate exit conditions of those components based on step inputs. Ordinary differential equations for each component constitute the governing equations to be solved simultaneously with constitutive relations. They were solved numerically by Gear s method, which uses backward differentiation and is used when a problem is stiff. The values of temperature (or enthalpy), mass flow rate, and ammonia mass fraction of each component were input as a function of time. T he t ime increment

PAGE 143

143 in numerical integration was automatically adjusted to keep the largest acceptable solver error to be 1 105 during each time step. For general expression s of outputs non dimensional variables were used for time and output variables Non dimensional tim e and variables are expressed as j jt t (6 90) final initialfinalxx x xx (6 91) where characteristic time, j j jim m jm and jim are total mass in the control volume and inlet mass flow rate of the refrigerant side in the heat exchanger, j x is a output variable, xi nitial and xfinal are initial and final values of x Condenser Figures 61 1 to 6 1 3 show nondimensional response to three step inputs Inlet ammonia mass fraction, inlet mass flow rate, and secondary fluid inlet mass flow rate are input perturbations; amm onia mass fraction, enthalpy and secoundary fluid temperature at the exit and the interface position are shown as outputs Inputs, initial conditions, and geometric parameters of the condenser are given in Table 6 1 When inlet ammonia mass fraction is su ddenly increased, saturated vapor enthalpy is decreased. As a result, the twophase region is shorter, and the sub cooled liquid is further cooled so that the exit enthalpy is decreased. W hile outputs of ammonia mass fraction and secondary fluid temperatur e at the exit transit ion to the final values monotonically the interface position and enthalpy at the exit show undershoots (Figure

PAGE 144

144 6 11). However, if a higher flow rate of saturated vapor is coupled with constant coolant flow (Figure 6 1 2 ), the two phase /single phase interfac e position is close to the exit, therefore, exit enthalpy is increased. On the other hand, higher flow rate of coolant (higher heat flux) reduces the two phase region (Figure 613). When mass flow rates are perturbed, the time constants are shorter than for the case when inlet ammonia mass fraction is changed. Absorber Figures 61 4 to 6 1 7 show nondimensional response to four step inputs for the absorber Inlet ammonia mass fraction, enthalpy and mass flow rate and secondary fluid mas s flow rate are input perturbations; ammonia mass fraction and enthalpy at the exit and the interface position are shown as outputs Inputs, initial conditions, and geometric parameters of the absorber are given in Table 6 2 Unlike the condenser, absorber inlet enthalpy is an independent variable. Therefore, inlet ammonia mass fraction and pressure change do not affect inlet enthalpy directly because the absorber inlet condition is two phase as shown in Figure 67. The effect of an increase of inlet ammonia mass fraction and secondary fluid mass flow rate is an increased cooling effect to the ammoniawater mixture. Therefore, the length of the two phase region is shorter as exit enthalpy is decreased. W e can see the opposite behavior when inlet enthalpy and mass flow rate are increased. There is unexpected inaccuracy in Figure 614. Exit temperature of secondary flow is expected to decrease, but the temperature increase s in early stages. This artifact is because of the assumption that the control volume tem perature is the average of inlet and outlet temperature for the secondary fluid. Single phase fluids in moving boundary models traditionally have the potential for this nonphysical effect [22]. As shown at

PAGE 145

145 condenser, the response time is longer for the perturbation of inlet ammonia mass fraction for absorber. Generator Figures 618 and 619 show nondimensional res ponse to two step inputs for the generator. Inlet ammonia mass fraction and inlet mass flow rate are the input perturbations, while exit enthalpies and ammonia mass fractions in both vapor and liquid are the outputs shown, which are important input variables for the rectifier Inputs, initial conditions, and geometric parameters of the generator are given in Table 6 3 For the generator, ammoniaw ater mixtures are heated rather than cooled as in the condenser and absorber. Therefore, nondimensional enthalpies for both vapor and liquid tend to increase. If inlet ammonia mass fraction and mass flow rate are increased, vapor and liquid enthalpies wil l be decreased while the ammonia mass fraction at the exit will be increased. Although the final results are as expected, there is an unexpected jump in the opposite direction when ammonia mass fraction is suddenly increased, as in Figure 618. The reason is the same as for the absorber. Because the enthalpy of single phase flow is assumed to be the average of inlet and outlet enthalpies, if initial enthalpy is suddenly changed, the outlet value is immediately affected. This means that the model is inaccura te early in step changer transients, as He et al. [ 21] mentioned in his research. Evaporator Figures 62 0 to 6 2 2 show nondimensional res ponse to three step inputs for the evaporator. Inlet ammonia mass fraction, enthalpy, and mass flow rate are input per turbations; exit ammonia mass fraction, enthalpy, imaginary boundary position and

PAGE 146

146 secondary fluid temperature are shown as outputs Inputs, initial conditions, and geometric parameters of the evaporator are given in Table 6 4 The refrigerant in the evaporator is heated by the hot secondary fluid. Therefore, lower inlet enthalpy increases the cooling effect. For almost saturated vapor, when ammonia mass fraction is decreased, the enthalpy is also decreased, unlike the case for nearly saturated ammonia l iqu id. Both cases increase the cooling effect of the evaporator and the results decreased length of the imaginary moving boundary and lower exit enthalpies T he cases for perturbed enthalpy and mass flow rate show changed show relatively shorter response tim e than that of ammonia mass flow change, just like the abovementioned two heat exchangers. Solution Heat Exchanger Figures 62 3 and 624 show nondimensional res ponse to two step inputs for the solution heat exchanger. Because both fluids are ammoniawate r mixture s and are liquid phase, only the colder fluid results are considered here Inlet ammonia mass fraction and mass flow rate are input perturbations; exit ammonia mass fractions and enthalpies for both fluids are shown as outputs Inputs, initial con ditions, and geometric parameters of the SHX are given in Table 65. For s ubcooled liquid of ammoniawater mixtures, the mixture enthalpy is a function of not only pressure and ammonia mass fraction but also temperature. When only inlet ammonia mass frac tion is increased at the specified fixed pressure and inlet enthalpy, inlet temperature can be increased or decreased because mixture temperature is not monotonic on an enthalpy composition plane. For the case represented in Table 6 5, colder flow temperat ure is increased. When mass flow rate of

PAGE 147

147 the colder flow is increase d the overall SHX is cooled more, resulting in lower enthalpies for both fluids. Summary A dynamic modeling method appropriate to heat exchangers for control applications has been developed in this c hapter. To satisfy both twophase flow physics and the control design requirements, the conventional moving boundary model ha s been adapted. This model has been used in the literature to model singlecomponent two phase fluid heat exchangers fo r control applications The conventional moving boundary model has been extended in this work to include species balance for twocomponent twophase fluids. Among three modeling options considered, the spatial mean value model was employed, along with tabulated steady state property data for the two phase region. Use of t abulated data significantly reduce s the number of equations and numerical complexity of determining equilibrium state. Also, the data can be easily applied to imaginary moving boundaries, w hich are reference states outside of the physical heat exchangers. Because thermodynamic properties at the boundaries are known, twophase inlet and outlet conditions are easily obtained. However, this newly developed m oving boundary model was modified fro m the widespread conventional moving boundary model, so inherent problems of the conventional model still exist

PAGE 148

148 Table 6 1 Parameters of the condenser Parameters Values Inlet ammonia mass fraction, x ci 0.998 0 Condenser pressure, P c [kPa] 1791.9 Inle t mass flow rate, m ci [kg/s] 0.0618 Secondary flow mass flow rate, m ca [kg/s] 0.6442 Condenser length, L c [m] 1.3335 Condenser tube area, A c [m 2 ] 0.0041 Universal heat transfer coefficient, ( UD ) c1 [kW/m K] 3.097 0 Universal heat transfer coefficient, ( UD ) c2 [kW/m K] 1.910 0 Step input: Inlet ammonia mass fraction, x cif 0.999 0 Step input: Inlet mass flow rate, m cif [kg/s] 0.06 80 Step input: Secondary flow mass flow rate, m caf [kg/s] 0. 7087 Table 62. Parameters of the absorber Parameters Values Inlet ammoni a mass fraction, x a i 0.62 00 Absorber pressure, P a [kPa] 833.96 Inlet mass flow rate, m a i [kg/s] 0.2166 Inlet enthalpy, h a i [kJ/kg] 330.69 Secondary flow mass flow rate, m aa [kg/s] 1.6584 Absorber length, L a [m] 1.3208 Absorber tube area, A a [m 2 ] 0.0079 Universal heat transfer coefficient, ( UD ) a 1 [kW/m K] 3.090 0 Universal heat transfer coefficient, ( UD ) a2 [kW/m K] 4.269 0 Step input: Inlet ammonia mass fraction, x a i f 0.626 2 Step input: Inlet enthalpy, h a i f [kJ/kg] 333.99 Step input: Inle t mass flow rate, m a i f [kg/s] 0. 2188 Step input: Secondary flow mass flow rate, m aaf [kg/s] 1.6749 Table 63. Parameters of the generator Parameters Values Inlet ammonia mass fraction, x g i 0.62 00 Generator pressure, P g [kPa] 1791.9 Inlet mass flow rate, m g i [kg/s ] 0.1860 Secondary flow mass flow rate, m ga [kg/s] 1.1529 Generator length, L g [m] 0.6096 Generator tube area, A g [m 2 ] 0.0411 Universal heat transfer coefficient, ( UD ) g 1 [kW/m K] 0.9657 Universal heat transfer coefficient, ( UD ) g2 [kW/m K] 0.8539 Step input: Inlet ammonia mass fraction, x g i f 0.626 2 Step input: Inlet mass flow rate, m g i f [kg/s] 0.1879

PAGE 149

149 Table 64. Parameters of the evaporator Parameters Values Inlet ammonia mass fraction, x e i 0.998 0 Evaporator pressure, P e [kPa] 833.96 Inl et mass flow rate, m e i [kg/s] 0.0618 Inlet enthalpy, h e i [kJ/kg] 125.27 Secondary flow mass flow rate, m ea [kg/s] 1.1529 Evaporator length, L e [m] 0.9144 Evaporator tube area, A e [m 2 ] 0.0592 Universal heat transfer coefficient, ( U A ) e [kW/K] 2.7 84 0 Step input: Inlet ammonia mass fraction, x e i f 0.999 0 Step input: Inlet enthalpy, h e i f [kJ/kg] 137.80 Step input: Inlet mass flow rate, m e i f [kg/s] 0.0680 Table 65. Parameters of SHX Parameters Values Inlet ammonia mass fraction for colder si de x sci 0.62 00 Inlet ammonia mass fraction for hotter side, x shi 0.469 0 SHX pressure, P s [kPa] 1791.9 Inlet mass flow rate for colder side, m sci [kg/s] 0.1860 Outlet mass flow rate for hotter side, m sho [kg/s] 0.1548 SHX length, L s [m] 2.032 0 SHX tube area, A s [m 2 ] 0.0027 Universal heat transfer coefficient, ( U A ) s [kW/m K] 1.7958 Step input: Inlet ammonia mass fraction, x scif 0.682 0 Step input: Inlet mass flow rate, m sc i f [kg/s] 0.2046

PAGE 150

150 Figure 61. A schematic of an evaporator m odel for single component flow 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 280 285 290 295 300 305 310 315 320 325 330 Dimensionless lengthTemperature [K] 600kPa 800kPa 1000kPa Figure 62. Two phase flow t emperature distribution in an evaporator

PAGE 151

151 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.4 0.5 0.6 0.7 0.8 0.9 1 Dimensionless lengthAmmonia mass fraction in liquid 600kPa 800kPa 1000kPa Figure 63. Saturated vapor a mmonia mass fraction in an evaporator 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.994 0.995 0.996 0.997 0.998 0.999 1 1.001 Dimensionless lengthAmmonia mass fraction in vapor 600kPa 800kPa 1000kPa Figure 64. Saturated liquid ammonia mass fraction in an evaporator

PAGE 152

152 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -500 0 500 1000 1500 2000 2500 3000 NH3 mass fractionEnthalpy [kJ/kg] Tsg Te Tsl xel xeg xb havg Figure 65 Enthalpy Ammonia mass fraction ( h x) diagram Figure 66 A s chematic of a condenser

PAGE 153

153 Figure 67 A s chematic of an absorber Figure 68 A schematic of a generator

PAGE 154

154 Figure 69 A s chematic of an evaporator Figure 61 0 A schemati c of SHX

PAGE 155

155 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 -5 -4 -3 -2 -1 0 1 Dimensionless Time, t* cDimensionless Outputs lc1 hco Tcao xco Figure 611. Inlet ammonia mass fraction step increase for the condenser 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 Dimensionless Time, t* cDimensionless Outputs lc1 hco Tcao Figure 612. Inlet mass flow rate step increase for the condenser

PAGE 156

156 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 Dimensionless Time, t* cDimensionless Outputs lc1 hco Tcao Figure 613. Secondary fluid inlet mass flow rate step increase for the condenser 0 0.5 1 1.5 2 2.5 3 3.5 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 Dimensionless Time, t* aDimensionless Outputs la1 hao Taao xao Figure 61 4. Inlet ammonia mass fraction step increase for the absorber

PAGE 157

157 0 0.5 1 1.5 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 Dimensionless Time, t* aDimensionless Outputs la1 hao Taao Figure 615. Inlet enthalpy step increase for the absorber 0 0.5 1 1.5 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 Dimensionless Time, t* aDimensionless Outputs la1 hao Taao Figure 616. Inlet mass flow rate step increase for the absorber

PAGE 158

158 0 0.5 1 1.5 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 Dimensionless Time, t* aDimensionless Outputs la1 hao Taao Figure 6 17. Secondary fluid inlet mass flow rate step increase for the absorber 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 Dimensionless Time, t* gDimensionless Outputs hgov hgol xgov xgol xgo Figure 6 18. Inlet ammonia mass fraction step increase for the generator

PAGE 159

159 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 Dimensionless Time, t* gDimensionless Outputs hgov hgol xgov xgol Figure 6 19. Inlet mass flow rate step increase for the generator 0 1 2 3 4 5 6 7 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Dimensionless Time, t* eDimensionless Outputs L* e heo Teao xeo Figure 6 20. Inlet ammonia mass fraction step increase for the evaporator

PAGE 160

160 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Dimensionless Time, t* eDimensionless Outputs L* e heo Teao Figure 6 21. Inlet enthalpy step increase for the evaporator 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Dimensionless Time, t* eDimensionless Outputs L* e heo Teao Figure 6 22. Inlet mass flow rate step increase for the evaporator

PAGE 161

161 0 1 2 3 4 5 6 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Dimensionless Time, t* sDimensionless Outputs xsco hsco hsho Figure 6 23. Colder side inlet ammonia mass fraction step increase for SHX 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Dimensionless Time, t* sDimensionless Outputs hsco hsho Figure 6 24. Colder side inlet mass flow rate step increase for SHX

PAGE 162

162 CHAPTER 7 CONCLUSIONS The PoWER system has been described, including its system attributes and dynamic challenges as motivation for this work towards development of physics based controls In th e current research, we focused primarily on the stea dy state and dynamic modeling of the VARS appropriate for development of overall system controls algorithms F urthermore, for accurate modeling, thermodynamic properties of ammoniawater mixtures have been studied and critiqued, and a recommended approach has been implemented for overcoming the limitations of current curvefit based models. Thermodynamic Properties of Ammonia Water Mixtures Several commonly used P T x relations have been studied to avoid timeconsuming iterative numerical calculation. Two high order fitted relationships and tabulated data with interpolation have been compared with experimental data. Highorder fitted P T x equations are: easy to program and their execution time is very short. They show very good agreement with experimental data. Local oscillatory patterns ( R unge s phenomenon) exist, near the pure component limits Bubble and dew point temperature lines do not meet at the pure component limits The se problems can lead to instability in iterative system solutions especially in systems with nearly pure ammonia such as the VARS They may be most useful for rapid calculations in so me steady state analyse s. However, alternative methods should be considered for programmed iterative calculations. Tabulated data are used in this rese arch with good results T he use of t abulated data and interpolation method show:

PAGE 163

163 Single values for both bubble and dew point temperature at pure component regime. Elimination of R unge s phenomenon. Only slightly slower execution time than fitted equations. Given the memory capabilities of modern computing platforms, it is possible to store and access very large tables with very little cost in execution time. However, to achieve this objective, accurate experimental data over the range on interest in the ca lculation are necessary. Preliminary Design of VARS A conceptual VARS model with an icemaking capability has been designed and analyzed. From this research, it is observed that waste heat from the HPRTE can be utilized effectivel y with a VARS having three evaporators Based on a typical daily load demand for a building, cooling capacity has been calculated and distributed to air conditioning and to an ice maker for thermal storage. Ice produced primarily at night can be used as a refrigeration source to su pplement the direct air conditioning from the VARS during the heat of the day. Initial modeling of the VARS used the Second Law efficiency, which produced the following conclusions: When the generator capacity is increased the gas turbine subsystem evapor ator can be downsized, as additional heat from generator is better used in air conditioning and ice making. The ice can be produced when air conditioning load is low and it can be used to chill water to cool the PoWER system thereby increasing efficiency, or for additional air conditioning. The energy storage in daily ice production is ideally approximately twice the total daily air conditioning capacity. This additional cooling capacity could be used to air condition a second building of similar size to the design building, or used in a load leveling manner to downsize the PoWER system.

PAGE 164

164 Considering the combination of attributes listed above, it can be concluded that the PoWER system seems to be a promising candidate for distributed power generation, espec ially in hot climates where the summer loadleveling benefits are important. A discretized dynamic VARS model has been developed under the assumption of uniform ammonia mass fraction, though initially applied to a quasi steady problem Heat exchangers have been spatially and temporally discretized and the governing differential equations thereby converted to algebraic form. From this research, a detailed multi stage VARS model has been developed and the results have been compared with the Second Law analysi s model. Comparing these two steady state analyses, the following conclusions may be reached: The S econd Law analysis is more convenient approach and is effective for preliminary design. The analysis is useful for comparing this concept to competing options and for conceptual design when the details of the VARS have necessaryily not yet been specified. Dynamic Modeling of Heat Exchangers To model twocomponent two phase dynamic heat exchanger s, the conventional moving boundary model has been extended. Som e modifications are needed for twocomponent fluids to allow two component fluids, removing the singlecomponent limitation of the conventional model. It has been clearly shown that the moving boundary model satisfies both twophase flow physics and the requirements of control design. A dynamic modeling method of heat exchangers for control applications has been discussed as presented in the literature. To satisfy both twophase flow physics and the

PAGE 165

165 control design requirements, the conventional movi ng boundary model has been found to have the following characteristics : It is a onedimensional two phase dynamic heat exchanger model. The interface between twophase and singlephase moves in response to timevarying boundary conditions Void fraction is always assumed constant in the two phase region. The model is capable of only singlecomponent flow and it is limited to the condition that the moving boundary is inside of the heat exchanger To extend the conventional moving boundary model for an ammoniawater VARS, species balances are added to the governing equation set From the three modeling options considered the spatial mean value model is employed and steady state calculation data, functions of pressure and ammonia mass fraction, are used in tabulated form for the two phase region. A reference, or imaginary moving boundary, which may fall upstream or downstream of the hea t exchanger, has been successfully introduced as a means of handling moving boundary problems with twophase flow at the inlet or exi t Tabulated data was found to significantly reduce complexity of the c onservation equation s et However, this newly developed moving boundary model was modified from the widespread conventional moving boundary model, so inherent problems of the convention al model still exist. Future Work The ultimate goal of this research is to enable development of advanced control of the PoWER system. For system control linear dynamic models of the system and more

PAGE 166

166 accurate thermodynamic properties of the working fluid are necessary. To improve the model, following research should be done: Thermodynamic consistency test with existing and newly found thermodynamic properties of ammonia water mixtures. Non equilibrium modeling for two phase ammoniawater mixtures during tr ansient state. Improved moving boundary models for singlephase flow to reduce or eliminate the i mmediate time response at the singlephase exit. Complete cycle model ing for a VARS including rectifier, storage, pump, and expansion valves. Once the aboveme ntioned research is finished, the nonlinear dynamic model for a VARS can be linearized to develop a robust control algorithm.

PAGE 167

167 APPENDIX A C OMPUTER CODE FOR THE MODYNAMIC PROPERTIES OF AMMONIA WATER MIXTURES The results shown in Chapter 3 are based on following computer code s. The necessary files to obtain th e results are shown at Table A 1 with descriptions Table A 1. Matlab files for thermodynamic properties of ammoniawater mixtures File name Descriptions bogart_si.mat Bogart P T x tables in SI unit The file consists of 5 tables P_bog: 24 pressure ranges Tb_bog: bubble point temperature table Td_bog: dew point temperature table x_bog: ammonia mass fraction for Tb_bog y_bog: ammonia mass fraction for Td_bog Tb_Bogart.m Bubble point temperature from p ressure and liquid ammonia mass fraction Td_Bogart.m Dew point temperature from pressure and vapor ammonia mass fraction y_Bogart.m Vapor ammonia mass fraction from pressure and liquid ammonia mass fraction liqprop_ibrahim.m Specific enthalpy, entropy, and volume from temperature, pressure, and liquid ammonia mass fraction vapprop_ibrahim.m Specific enthalpy, entropy, and volume from temperature, pressure, and vapor ammonia mass fraction T he files, Tb_Bogart.m Td_Bogart.m, and y_Bogart.m use Bogart t ables with twodimensional linear interpolation method for P T x of ammoniawater mixtures. For specific enthalpy, entropy, and volume, Ibrahim s equations [15] are used for liqprop_ibrahim.m and vapprop_ibrahim.m. Bubble and dew point temperatures are fundtions of pressure and ammonia concentration. Scalar values of pressure and ammonia mass fraction are used as inputs for both Tb_Bogart.m and Td_Bogart.m. C onsider P, x and y as pressure and ammonia mass fraction in saturated liquid and vapor Their temp eratures are easily obtained by

PAGE 168

168 typing Tb_Bogart(P,x) or Td_Bogart(P,y) on Matlab command window if P, x and y are memorized in Matlab Workspace. Pressure is in kPa and concentration is in kg/kg and resulting temperature is in Kelvin. To get ammonia mass fraction in saturated vapor, type y_Bogart( P,x). Ibrahim s relations use T (temperature), P, and x or y as inputs. To use the m files, liqprop_ibrahim.m and vapprop_ibrahim.m, typing order for inputs is important. Temperature is always the first followed by P and x. Output order is specific enthalpy, entropy, and volume. F or example, if saturated vapor enthalpy is interested, simply type vapprop_ibrahim(Td_Bogart(P,y),P,y) on the command window. Matlab code for each file is shown below: Tb_Bogart.m % Tb_Bogart.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% function [Tb flag1] = Tb_Bogart(P,x) load bogart_si.mat if P > max(P_bog) disp('pressure is too high, the max pressure is 2068.4271kPa = 300psia') Tb = []; flag1 = 0; % To stop simulation elseif P < min(P_bog) disp('pressure is too low, the min pressure is 41.368542kPa = 6psia') Tb = []; flag1 = 0; elseif x > 1 disp('Ammonia concentration cannot be higher than 1') Tb = []; flag1 = 0; elseif x < 0 disp('Ammonia concentration cannot be lower than 0') Tb = []; flag1 = 0; else Tb = interp2(x_bog,P_bog,Tb_bog,x,P,'linear'); flag1 = 1; end % End of Tb_Bogart.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

PAGE 169

169 Td_Bogart.m % Td_Bogart.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% function [Td flag1]= Td _Bogart(P,y ) load bogart_si.mat if P > max(P_bog) disp('pressure is too high, the max pressure is 2068.4271kPa = 300psia') T d = []; flag1 = 0; % To stop simulation elseif P < min(P_bog) disp('pressure is too low, the min pressure is 41.368542kPa = 6psia') T d = []; flag1 = 0; elseif y > 1 disp('Ammonia concentration cannot be higher than 1') T d = []; flag1 = 0; elseif y < 0 disp('Ammonia concentration cannot be lower than 0') T d = []; flag1 = 0; else T d = interp2(y _bog,P_bog,Td _bog,y ,P,'linear'); flag1 = 1; end % End of Td_Bogart.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% y_Bogart.m % y_Bogart.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% function [y flag1] = y_Bogart(P,x) load bogart_si.mat if P > max(P_bog) disp('pressure is too high, the max pressure is 2068.4271kPa = 300psia') y = []; flag1 = 0; % To stop simulation elseif P < min(P_bog) disp('pressure is too low, the min pressure is 41.368542kPa = 6psia') y = []; flag1 = 0; elseif x > 1 disp('Ammonia concentration cannot be higher than 1') y = []; flag1 = 0; elseif x < 0 disp('Ammonia concentration cannot be lower than 0') y = []; flag1 = 0; else y = interp2(x_bog,P_bog,y_bog,x,P,'linear');

PAGE 170

170 end % End of y_Bogart.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% liqprop_ibrahim.m % liqprop_ibrahim.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% function [hl sl vl] = liqprop_ibrahim(T,P,x) % Ibrahim thermodynamic property of the liquid mixture % % Input: T[K], Temperature (scalar) % P[kPa], Pressure (scalar) % x[kg/kg], NH3 mass fraction (scalar btw 0 & 1) % Output: hl[kJ/kg], Specific enthalpy % sl[kJ/kg/K], Specific entropy % vl[m3/kg], Specific volume % % Ver. 0, 20080205 % Ver. 1, 20080221 enthalpy eliminated (to use Patek's enthalpy) % Ver. 2, 20090604 enthalpy reuse, x in row vector % Ver, 3, 20090615 x and results in a scalar (no need to be a vector) % made by Choon Jae Ryu: choonjae.ryu@gmail.com % Energy and Gas Dynamics Lab. % Mechanical and Aerospace Engineering, University of Florida % % Data and equations from O.M. Ibrahim, S.A. Klein, % "Thermodynamic Properties of AmmoniaWater Mixtures", % ASHRAE Transactions: Symposia, Vol.99, pp.14951502, 1993 cmol = 18.015268*x./(0.985008*x+17.03026); %[kmol/kmol] mole fraction mwmix = cmol*17.03026 + (1cmol)*18.015268; %[kg/kmol] mix. mole. weight pbar = P/100; %[bar] tb = 100; %[K] reference temperature pb = 10; %[bar] reference pressure R = 8.314; %[kJ/kmol/K] gas constant coliq = [3.971423e2 2.748796e2; 1.790557e5 1.016665e5; 1.308905e2 4.452025e3; 3.752836e3 8.389246e4; 1.634519e1 1.214557e1; 6.508119 1.898065; 1.448937 2.911966e1; 4.878573 21.821141; 1.644773 5.733498; 3.2252 5.0705; 2 3 ]; exc = [41.733398 0.02414 6.702285 0.011475 ... 63.608967 62.490768 1.761064 0.008626 ... 0.387983 0.004772 4.648107 0.836376 ... 3.553627 0.000904 24.361723 20.736547];

PAGE 171

171 pr = pbar/pb; %dimensionless pressure tr = T/tb; %dimensionless temperature hal = R tb (coliq(8,1) + coliq(5,1) (coliq(10,1) tr) ... + coliq(6,1) / 2 (coliq(10,1)^2 tr^2) ... + coliq(7,1) / 3 (coliq(10,1)^3 tr^3) ... + (coliq(4,1) tr^2 coliq(1,1)) (pr coliq(11,1)) ... coliq(2,1) / 2 (pr^2 coliq(11,1)^2)); % [kJ/kmole] NH3 molar enthalpy sal = R (coliq(9,1) + coliq(5,1) log(tr / coliq(10,1)) ... + coliq(6,1) (tr coliq(10,1)) ... + coliq(7,1) / 2 (tr^2 coliq(10,1)^2) ... + (coliq(11,1) pr) (coliq(3,1) + 2 coliq(4,1) tr)); % [kJ/kmole/K] NH3 molar entropy val = R tb / pb / 100 (coliq(1,1) + coliq(2,1) pr ... + coliq(3,1) tr + coliq(4,1) tr^2); % [m3/kmole], NH3 molar volume hwl = R tb (coliq(8,2) + coliq(5,2) (coliq(10,2) tr) ... + coliq(6,2) / 2 (coliq(10,2)^2 tr^2) ... + coliq(7,2) / 3 (coliq(10,2)^3 tr^3) ... + (coliq(4,2) tr^2 coliq(1,2)) (pr coliq(11,2)) ... coliq(2,2) / 2 (pr^2 coliq(11,2)^2)); % [kJ/kmole] H2O molar enthalpy swl = R (coliq(9,2) + coliq(5,2) log(tr / coliq(10,2)) ... + coliq(6,2) (tr coliq(10,2)) ... + coliq(7,2) / 2 (tr^2 coliq(10,2)^2) ... + (coliq(11,2) pr) (coliq(3,2) + 2 coliq(4,2) tr)); % [kJ/kmole/K] H2O molar entropy vwl = R tb / pb / 100 (coliq(1,2) + coliq(2,2) pr ... + coliq(3,2) tr + coliq(4,2) tr^2); % [m3/kmole], H2O molar volume % Excess properties for liquid only if (cmol == 0) | (cmol == 1) he = 0; se = 0; ve = 0; slmix = 0; else he = R tb (1 cmol) .* (exc(1) + exc(2) pr + 2 exc(5) / tr ... + 3 exc(6) / tr^2 + (2 cmol 1) .* (exc(7) + exc(8) pr ... + 2 exc(11) / tr + 3 exc(12) / tr^2) + (2 cmol 1).^2 ... (exc(13) + exc(14) pr + 2 exc(15) / tr + 3 exc(16) / tr^2)); se = R (1 cmol) .* (exc(3) + exc(4) pr exc(5) / tr^2 ... 2 exc(6) / tr^3 + (2 cmol 1) .* (exc(9) + exc(10) pr ... exc(11) / tr^2 2 exc(12)/tr^3 + (2 cmol 1).^2 ... (exc(15) / tr^2 2 exc(16) / tr^3))); ve = R tb / pb / 100 (1 cmol) .* (exc(2) + exc(4) tr ... + (2 cmol 1) .* (exc(8) + exc(10) tr) ... + exc(14) (2 cmol 1).^2);

PAGE 172

172 slmix = R (cmol .* log(cmol) + (1 cmol) .* log(1 cmol)); end hml = cmol hal + (1 cmol) hwl + cmol .* he; %[kJ/kmol] mix mol enthalpy sml = cmol sal + (1 cmol) swl + cmol .* se + slmix; vml = cmol val + (1 cmol) vwl + cmol .* ve; hl = hml ./ mwmix; %[kJ/kg] mixture enthalpy sl = sml ./ mwmix; %[kJ/kg/K] vl = vml ./ mwmix; %[m3/kg] % End of liqprop_ibrahim.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% vapprop_ibrahim.m % vapprop_ibrahim.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% function [hv sv vv] = vapprop_ibrahim(T,P,y) % Ibrahim thermodynamic property of the vapor mixture % % Input: T[K], Temperature (scalar) % P[kPa], Pressure (scalar) % y[kg/kg], NH3 mass fraction (scalar between 0 & 1) % Output: hv[kJ/kg], Specific enthalpy % sv[kJ/kg/K], Specific entropy % vv[m3/kg], Specific volume % % Ver. 0, 20080205 % Ver. 1, 20080221 enthalpy eliminated (to use Patek's enthalpy) % Ver. 2, 20090605 enthalpy reuse, y in row vector % Ver, 3, 20090615 y and results in a scalar (no need to be a vector) % made by Choon Jae Ryu: choonjae.ryu@gmail.com % Energy and Gas Dynamics Lab. % Mechanical and Aerospace Engineering, University of Florida % % Data and equations from O.M. Ibrahim, S.A. Klein, % "Thermodynamic Properties of AmmoniaWater Mixtures", % ASHRAE Transactions: Symposia, Vol.99, pp.14951502, 1993 cmol = 18.015268*y ./ (0.985008*y+17.03026); %[kmol/kmol] mole fraction mwmix = cmol*17.03026 + (1cmol)*18.015268; %[kg/kmol] mix mole weight pbar=P/100; %[bar] tb = 100; %[K] reference temperature pb = 10; %[bar] reference pressure R = 8.314; %[kJ/kmol/K] gas constant covap = [1.049377e2 2.136131e2; 8.288224 3.169291e1; 6.647257e2 4.634611e4; 3.045352e3 0; 3.673647 4.019170; 9.989629e2 5.175550e2; 3.617622e2 1.951939e2; 26.468879 60.965058;

PAGE 173

173 8.339026 13.453430; 3.2252 5.0705; 2 3 ]; pr = pbar/pb; %dimensionless pressure tr = T/tb; %dimensionless temperature hav = R tb (covap(8,1) + covap(5,1) (covap(10,1) tr) ... + covap(6,1) / 2 (covap(10,1)^2 tr^2) ... + covap(7,1) / 3 (covap(10,1)^3 tr^3) ... covap(1,1) (pr covap(11,1)) ... + 4 covap(2,1) (covap(11,1) / covap(10,1)^3 pr / tr^3) ... + 12 covap(3,1) (covap(11,1) / covap(10,1)^11 pr / tr^11) ... + 4 covap(4,1) (covap(11,1)^3 / covap(10,1)^11 pr^3 / tr^11)); %[kJ/kmol] NH3 enthalpy sav = R (covap(9,1) + 11 covap(3,1) ... (covap(11,1) / covap(10,1)^12 pr / tr^12) ... + 11 / 3 covap(4,1) (covap(11,1)^3 / covap(10,1)^12 pr^3 / tr^12) ... + 3 covap(2,1) (covap(11,1) / covap(10,1)^4 pr / tr^4) ... + covap(6,1) (covap(10,1) tr) + covap(7,1) / 2 ... (covap(10,1)^2 tr^2) + log(pr / covap(11,1)) ... covap(5,1) log(tr / covap(10,1))); %[KJ/kmol/K] vav = R tb / pb / 100 (tr / pr + covap(1,1) ... + covap(2,1) / tr^3 + covap(3,1) / tr^11 ... + covap(4,1) pr^2 / tr^11); %[m3/kmol] hwv = R tb (covap(8,2) + covap(5,2) (covap(10,2) tr) ... + covap(6,2) / 2 (covap(10,2)^2 tr^2) ... + covap(7,2) / 3 (covap(10,2)^3 tr^3) ... covap(1,2) (pr covap(11,2)) ... + 4 covap(2,2) (covap(11,2) / covap(10,2)^3 pr / tr^3) ... + 12 covap(3,2) (covap(11,2) / covap(10,2)^11 pr / tr^11) ... + 4 covap(4,2) (covap(11,2)^3 / covap(10,2)^11 pr^3 / tr^11)); %[kJ/kmol] H2O enthalpy swv = R (covap(9,2) + 11 covap(3,2) ... (covap(11,2) / covap(10,2)^12 pr / tr^12) ... + 11 / 3 covap(4,2) (covap(11,2)^3 / covap(10,2)^12 pr^3 / tr^12) ... + 3 covap(2,2) (covap(11,2) / covap(10,2)^4 pr / tr^4) ... + covap(6,2) (covap(10,2) tr) + covap(7,2) / 2 ... (covap(10,2)^2 tr^2) + log(pr / covap(11,2)) ... covap(5,2) log(tr / covap(10,2))); %[KJ/kmol/K] vwv = R tb / pb / 100 (tr / pr + covap(1,2) ... + covap(2,2) / tr^3 + covap(3,2) / tr^11 ... + covap(4,2) pr^2 / tr^11); %[m3/kmol] if (cmol == 0) | (cmol == 1) svmix = 0; else svmix = R (cmol .* log(cmol) + (1 cmol) .* log(1 cmol));

PAGE 174

174 end hmv = cmol*hav + (1cmol)*hwv; %[kJ/kmol] mixture molecular enthalpy smv = cmol*sav + (1cmol)*swv + svmix; %[kJ/kmol/K] mix mol entropy vmv = cmol*vav + (1cmol)*vwv; %[m3/kmol] mix mol specific volume hv = hmv ./ mwmix; %[kJ/kg] gas mixture enthalpy sv = smv ./ mwmix; %[kJ/kg/K] gas mixture entropy vv = vmv ./ mwmix; %[m3/kg] gas mixture specific volume % End of vapprop_ibrahim.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

PAGE 175

175 APPENDIX B COMPUTER CODE FOR DYNAMIC MODELING WIT H CONSTANT AMMONIA M ASS FRACTION APPROACH For dynamic modeling of VARS with constant ammonia mass fraction approach, the softwares, Simulink and Matlab, are used. To run the simulation, three files are needed as shown in Table B 1. Before running main file, AC_input_md.mdl, vars_es08_03.m needs to be run for system parameters. Main input for the Simulink file is air conditioning load. For the results in Chapter 5, ac_load.mat has been used for air conditioning load. All files are developed using Matlab version 2007. It is confirmed that these files do not work on Matlab version 2009 because embedded Matlab function box do not allow an additional function to be placed in Matlab co de. Table B 1. Matlab and Simulink files for dynamic modeling with constant ammonia mass fraction approach. File name Descriptions ac_load.mat Air conditioning load for two days with time vars_es08_03.m System parameters to run AC_input_md.mdl AC_input _md.mdl Main code for dynamic simulation vars_es08_03.m % vars_es08_03.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % Steady state modeling of VARS for ES08 (Energy Sustainability 2008) % Triad ThermoCharger % Temperature, and Pressure data are from Aug. 10th 2008 % Data for each point are average data between 3309.3091, and 3948.228 % seconds, and solution pump pressure is 260psi = 1792.64kPa, which is % relatively higher pressure than normal. (Useally, under 250psi) % % Version 0, Feb/09/2008 % % made by Choon Jae Ryu (choonjae.ryu@gmail.com, choonjae.ryu@ufl.edu) % Energy and Gas Dynamics Lab. (Prof. William E. Lear) % Mechanical and Aerospace Engineering, University of Florida % time gap, design value dt_ds = 60000; %[sec], 100 minutes to quasisteady state dt = 600; %[sec], 10 minutes dp_test = 0.98; % 2% of pressure drop based on test data

PAGE 176

176 % Pressures Pmax = 1792.63682; %[kPa], 260psi design value Pmdl = 521; %[kPa], to make 5oC at Evap1, and Evap2 Pmin = 358; %[kPa], to make 5oC at Evap3, not available %during maximum A/C (Evap2) operation mode % Chilled water supply temperature Tcws = 290; %[K], design value Tcw8 = 310; %[K], COND heated coolant water temp. Tcwabs = 300; %[K], ABS1, ABS2 heated coolant water temp. % HRVG: Heat Recover Vapor Generator % Important: This block is the basis of the whole model % Input Qhrvg_gp = 440; %[kW] = 15.99[TR] from gas path to solution qhrvg = 665; %[kJ/kg], design value m_sol = Qhrvg_gp/qhrvg; %[kg/s], solution mass, design value % RECT: Rectifier % Input % Note: the state of the point 1 must always be saturated vapor Pds1 = 1723.6893; %[kPa] cds1 = 0.998; %[kg/kg] m1 = 0.2854*m_sol; %[kg/s], this value is always fixed. % EVAP1: Evaporator 1, linked with Gas Turbine System % Input qevap1 = 1100; %[kJ/kg], design value % EVAP2: Evaporator 2, Air Conditioner % Input qevap2 = 1100; %[kJ/kg], design value m2max = m1/3; %Revp = 0.5 Qev2_max = qevap2*m2max;%[kW], design value % EVAP3: Evaporator 3, Ice Maker % Input qevap3 = 1110; %[kJ/kg], design value % Pump (temporary), In this model, there is no pump. Only fixed data are % used at state point 15. Peff = 0.5; %Pump efficiency % It should be determined to use the power of Pump, or not for net energy % balance. % SHX: Solution heat exchanger T19 = 365.07; %[K], design value % ABS1: Absorber 1 T14 = 314.96; %[K], design value % ABS2: Absorber 2 T141 = 303.2563; %[K], design value % End of vars_es08_03.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

PAGE 177

177 AC_input_md.mdl AC_input_md: Gas turbine system for air conditioner made by Choon Jae Ryu (choonjae.ryu@gmail.com, choonjae.ryu@ufl.edu) Energy and Gas Dynamics Lab. (Prof. William E. Lear) Mechanical and Aerospace Engineering, University of Florida Version 0, Mar/03/2008 This model operates with "vars_es08_03.m". Before run this model, start the m file, first. Qev2,ice QacICE QacICE_load QIce_in QacICE Ice Maker EV2_Load QIce_out Qev2AC_out Gas Turbin with VARS Qev2_max/0.45 1/0.45 1/100 ac_load.mat From File Figure B 1. Simulink model for Chapter 5.

PAGE 178

178 APPENDIX C GOVERNING EQUATIONS AND COMPUTER CODE FO R MOVING BOUNDARY MODEL Governing Equations for Moving Boundary Model Governing equations for each component in a vapor absorption refrigeration system are derived here in detail. Coefficie nts in matrix form of the conservation equations are also indicated. Leibniz s rule is often used for differentiation of integrals: 11 22 21 21 ,,, zt zt zt ztfzt dzt dzt d dzfztdzfzttfztt tdt dt dt ( C 1 ) Derivations and Coefficients of Single Component Evaporator Model For mass balance of c ontrol volume 1, integrate Equation (61): 11000llAu Adz dz tz (C 2 ) where A is the cross sectional area of the evaporator and a constant. Applying Leibniz s rule to the first term of Equation ( C 2 ) gives: 11 1 00l int int idl d Adz AuAu dt dt (C 3 ) Thi s equation is rewritten 1 11 i intdl d AlAuAu dt dt (C 4 ) where 11 111 00 1 1 0 ll ldzdz l dz

PAGE 179

179 Because the interface between twophase and vapor fluids moves back and forth, the mass flow rate of the refrigerant at interface, intm is related to the relative speed: 1int intdl mAu dt (C 5 ) Therefore, the mass balance for control volume 1 is 11 11 iintddl AlAmm dtdt (C 6 ) S ame approach is used for ene rgy balance of control volume 1. Integrate Equation (6 2 ): 111000 lll iiwrhP Auh A dz dzDTTdz tz (C 7 ) Apply Leibniz s rule to the first term of Equation (C 7 ): 111 11 1111 00 ll intint int iiiwr int idl d AhdzPdzhPAuhAuhDlTT dt dt (C 8) This equation is rewritten 11 111 1111 i intiiwr i intdl dl d AhPlAPAuhAuhDlTT dt dt dt (C 9 ) where 11 111 11 00 11 1 0 ll lhdzhdz h l dz

PAGE 180

180 From Equation (C 5 ), energy balance for control volume 1: 11 1 1 111 1111 iiintintiiwrdhdl dP AlAhAlmhmhDlTT dtdtdt (C 10) Note that 1111hh even though they can be close to each other. For control volume 2, mass balance and energy balance after integrating z=l1( t ) to L : 110LL llAu Adz dz tz (C 11) 111LLL iiwr lllhP Auh A dz dzDTTdz tz (C 12) Applying Leibniz s rule to Equations (C 11) and (C 12) gives: 11 20L int o int ldl d Adz AuAu dt dt (C 13) 111 22 2222 LL intint o intiiwr o int lldl d AhdzPdzhPAuhAuhDlTT dt dt (C 14) where L = l1 + l2. These equations are rewritten after applying Equation (C 5 ) 1 22 o intdl d AlAu Au dt dt (C 15) 11 222 2222 int oiiwr o intdl dl d AhPlAPAuhAuhDlTT dt dt dt (C 16)

PAGE 181

181 where 11 1 22 2 2 LL ll L ldzdz l dz and 11 122 22 22 2 LL ll L lhdzhdz h l dz Therefore, mass balance and energy balance for control volume 2 are: 21 22 intoddl AlAmm dtdt (C 1 7 ) 22 1 2 222 2222 intintooiiwrdhdl dP AlAhAlmhmhDlTT dtdtdt (C 1 8 ) Energy equations for the tube walls can be derived similarly. It is assumed that the interface wall temperature, Tw12, is same as the control volume 1 wall temperature, Tw1. Then, the result equations are: 1 111 1 w p iirwooaw wdT CADTTDTT dt ( C 19) 212 1 222 1 2 www p iirwooaw wdTTT dl CA DTTDTT dtldt ( C 20) To cancel out interface mass flow rate, intm Equations (C 6 ) and (C 17) are combined. Equation (C 6 ) is substitute to Equation (C 10) and Equation (C 17) is substitute to Equation (C 18): 121 12 12 iodd dl AlAlA mm dtdt dt (C 21) 11 1 1 1 111 1111 int int iiintiiwrdhd dl dP Alh AhhmhhDlTT dtdtdt dt (C 22) 22 2 1 2 222 2222 int int ointoiiwrdhd dl dP Alh AhhmhhDlTT dtdtdt dt (C 23)

PAGE 182

182 Density and enthalpy are functions of only pressure in twophase region and are functions of pressure and temperature in vapor region for singlecomponent evaporator. 11d dP dtPdt (C 24) 1111dhh dP dtPdt (C 25) 2222 2 r rd dT dP dtPdtTdt (C 26) 2222 222 2 r rdhhhdTdP dtPdtTdt (C 27) Above equations are substitute to Equations (C 21) to (C 23): 112 22 1212 2 2 r io rdld dT dP A All Al mm dtdPPdtTdt (C 28) 1 11 1 111 1 11111int int iiintiiwrdlh dP AhhAlh mhhDlTT dtPPdt (C 29) 1 22 2 22 22 222 2 2 22 22221r int int int rr ointoiiwrdlh h dT dP dP AhhAl h Alh dtPdtPdtTTdt mhhDlTT (C 30) Total conservation equations for a singlecomponent evaporator are expressed as matrix form as shown at Equation (613 ). Their coefficients are: 1112dA 12 121 2d dAll dPP

PAGE 183

183 2 132 2 rdAl T 21111 intdAhh 11 1 2211inth dAlh PP 31 222 intdAhh 22 2 3221inth dP dAl h PdtP 22 2 332 22 int rrh dAlh TT 44 p wdCA 12 51 2 wwTT d l 55 p wdCA 1 iofmm 2 1111 iiintiiwrfmhhDlTT 3 2222 ointoiiwrfmhhDlTT 4111 1 iirwooawfDTTDTT 5222 1 iirwooawfDTTDTT Coefficients of Two Compone nt Condenser Model The coefficients of m atrix C and vector cF in Equation (632 ) are shown below: 1211 ccccAl

PAGE 184

184 21 112 2 2 ccccoccoccAxxxx 1 22112 1 c cccco ccAlxx x 2 23222 2 c ccccco ccAlxx x 2 2422 2 c cccco crcAlxx T 31 11112 ccccccAhh 11 1 321 12 11 cc c cc c cch cAlh xx 41 112 2 2 ccccoccoccAhhhh 1 42112 1 c cccco ccAlhh x 22 4322 2 22 cc cccco c cc h cAlhh xx 22 4422 2 22 cc cccco c cr crh cAlhh TT 55 1 pc cacACl 66 2 pc cacACl 1 12 ccicicfmxx 12 2 12 112 22 c cc cciccocccco cccco ccdP fmxxAlxxAlxx P Pdt

PAGE 185

185 11 1 3 12 1111 12 11cc cc ccicic ccacrcc c c cch dP fmhhUDlTTAlh PPdt 4 12 222 2 1 22 112 22 21ccicco ccacr c c ccc cccco cccco c c ccfmhhUDlTT hdP AlhhAlhh P PPdt 5 21 111 1 ccapcacacao ccrca cfmCTTUDlTT 6 21 222 2 ccapcacaica ccrca cfmCTTUDlTT Coefficients of Two Compone nt Absorber Model The coefficients of matrix A and vector aF in Equation (647 ) are shown below: 1211 aaaaAl 21 112 2 2 aaaaoaaoaaAxxxx 1 22112 1 a aaaao aaAlxx x 2 23222 2 a aaaaao aaAlxx x 2 2422 2 a aaaao araAlxx T 31 11112 aaaaaaAhh 11 1 321 12 11 aa a aa a aah aAlh xx 41 112 2 2 aaaaoaaoaaAhhhh 1 42112 1 a aaaao aaAlhh x

PAGE 186

186 22 4322 2 22 aa aaaao a aah aAlhh xx 22 4422 2 22 aa aaaao a ar arh aAlhh TT 55 1 pa aaaACl 66 2 pa aaaACl 1 12 aaiaiafmxx 12 2 12 112 22 a aa aaiaaoaaaao aaaao aadP fmxxAlxxAlxx P Pdt 11 1 3 12 1111 12 11aa aa aaiaia aaaaraa a a aah dP fmhhUDlTTAl h PPdt 4 12 222 2 1 22 112 22 21aaiaao aaaar a a aaa aaaao aaaao a a aafmhhUDlTT hdP AlhhAlhh P PPdt 5 21 111 1 aaapaaaaaao aaraa afmCTTUDlTT 6 21 222 2 aaapaaaaiaa aaraa afmCTTUDlTT Coefficients of Two Compone nt Generator Model The coefficients of matrix G and vector gF in Equation (660 ) are shown below: 1311 ggggAl 21 112 ggggogAxx 1 22112 1 g ggggo grgAlxx T

PAGE 187

187 1 23112 1 g ggggo ggAlxx x 2422 ggggAl 31 1112 gggggAhh 11 321112 1 11 gg gggg g gr grh gAlhh TT 11 331112 1 11 gg gggg g ggh gAlhh xx 41 112 222 ggggoggggogAhhhh 1 42112 1 g ggggo grgAlhh T 1 43112 1 g ggggo ggAlhh x 22 2 442 22 gg g gg go ggh gAl h xx 55 1 pg gagACl 66 2 pg gagACl 1 12 ggigigfmxx 1 2 12 112 gg ggiggoggggo gdP fmxxAlxx Pdt 11 3 12 111 1112 1 11ggg ggigig ggagrgggg g g gghdP fmhhUDlTTAlhh PPdt

PAGE 188

188 4 12 222 2 1 22 2 112 21ggiggo ggagr g g gg g g ggggo gg go gggfmhhUDlTT h dP AlhhAl h P PPdt 5 21 111 1 ggapgagagao ggrga gfmCTTUDlTT 6 21 222 2 ggapgagaiga ggrga gfmCTTUDlTT Derivations and Coefficients of Two Compone nt Evaporator Model For control volume 1, mass balance that L eibniz s rule is applied from (61): ** 1 0eL e eee ei edL d Adz AuAu dt dt ( C 31) Left hand side is rewritten using mean de nsity: ** 1 e eeee ei edL d AL AuAu dt dt ( C 32) A nd we can obtain ** 1 1 ee ee ee eiedLd AALmm dtdt ( C 33) where ei eimAu and **ee ee edL mAuA dt Just like control volume 1, mass balance for control volume 2 is ** 2 2 ee ee eee eoedL d AALLmm dt dt ( C 34)

PAGE 189

189 We are interested in the fluid inside the evaporator, so, sub tract E quation ( C 34) from ( C 33) and the m ass balance for refrigerant : ** 12 12 ee e eee ee eee eieodLd d A ALALLmm dtdt dt ( C 35) The coefficients of matrix E and vector eF in Equation (672 ) are shown below: ** 12 1 2 eeeeeeeALLL 21 111 222 eeeeeoeeeeoeAhhhh ** 11 1 22 2 22 ee e ee e ee eo eee eo ee eehh eALhALL h xxxx 33 pe eaeACL 1 eeieieofmxx 2 ** 11 1 22 2 11 eeieieo eaer e ee e ee e e ee eo ee eo ee eefmhhUATT h h dP ALhLLh PP PPdt 3 eeapeaeaieao erea efmCTTUATT Coefficients of Two Compone nt SHX Model The coefficients of matrix S and vector sF in Equation (682) are shown below: 11 sc scsscsco scsALxx T 12 sc scsscsco sc scsALxx x

PAGE 190

190 21 sc sc scsscsco sc sc sch sALhh TT 22 sc sc scsscsco sc sc sch sALhh xx 33 sh shsshshi shsALxx T 34 sh shsshshi sh shsALxx x 43 sh sh shsshshi sh sh shh sALhh TT 44 sh sh shsshshi sh sh shh sALhh xx 1 scs sscisciscoscsscsco sdP fmxxALxx Pdt 21sc sc s ssciscisco shscscsscsco sc s sshdP fmhhUATTALhh PPdt 3 shs sshoshishoshsshshi sdP fmxxALxx Pdt 41sh sh s sshoshisho scshshsshshi sh s sshdP fmhhUATTALhh PPdt Computer Code for Moving Boundary Model For dynamic modeling of twocomponent twophase flow heat exchangers, Simulink and Matlab are used. Level 2 S function in Simulink is employed for each heat exchanger to calculate continues dynamic problem. Bogart tables and Matlab codes introduced in Appendix A are applied to heat exchanger model s. Each heat exchanger

PAGE 191

191 model has a simulation file, **_Sim.mdl and **_Dynamics.m for its level 2 m file. To run these files, the driver file, **_Driver.m is needed. This file has inputs, initial values, and geometric parameters for **_Sim.mdl and **_Dynamics.m. To obtain twophase thermodynamic properties of ammoniawater mixture, **_2ph.m and its data file, **_pxz.mat files are necessary. For example, c ondenser simulation model, C OND_Sim.mdl file, is shown in Figure C1 and other models have similar Simulink file s 2 u 1 y Scope Inputs Inputs for Condenser COND_Dynamics Condenser S-func (L2) Figure C 1. Condenser Simulink model Before running COND_Driver.m, open COND_Sim.mdl and Configuration Parameters. We can select a solver, simulation time, relative tolerance, etc in the window. After tuning input parameters, step inputs in COND_Driver.m, push F5 key. This allows runing COND_Sim.mdl. The results are shown at workspace. Source codes are shown below: Condenser COND_Driver.m % COND_Driver.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % This is a driver file to run a dynamic simulation of a condenser. % Written by ChoonJae Ryu, Energy and Gas Dynamics Lab (Prof. William E. % Lear). % Department of Mechanical and Aerospace Engineering, University of Florida

PAGE 192

192 % This format is modified from the original written by Prof. Oscar D. % Crisalle (Chemical Egineering, % University of Florida). clear all %% Initial values %% % Initial inputs % Pc = 1.791850466e+3; xci = 0.998; [xcil Tci] = y2x_Bogart(Pc,xci); hci = vapprop_ibrahim(Tci,Pc,xci); mdci = 0.06182711; Tcai = 2.887111e+2; mdca = 0.644232220068264; % Geometry and Parameters% Lc = 1.3335; Ac = 0.004141911097410; UDc1 = 3.097209791054786; UDc2 = 1.909558741495225; % Initial values of state variables% lc1 = 1.313597926931102; xc1 = 0.998; xc2 = xc1; Tcr2 = 3.167085762441687e+2; Tca1 = 3.020517268311589e+2; Tca2 = 2.889072768311589e+2; %% Step Inputs %% stime = 0; % step time or start time % Pressure Pcf = Pc; Pcslope = 0; % step input for dPcdt Pcst = stime; % sec. start time % Ammonia mass fraction and enthalpy xcif = xci; % final value xcist = stime; % step time [xcil Tci] = y2x_Bogart(Pcf,xcif); hcif = vapprop_ibrahim(Tci,Pcf,xcif); % Mass flow rate refrigerant mdcif = mdci; % final value mdcist = stime; % step time % Mass flow rate secondary fluid mdcaf = mdca; % final value mdcast = stime; % step time % Inlet temperature secondary fluid Tcaif = Tcai; % final value Tcaist = stime; % step time

PAGE 193

193 %% Run Simulink Model %% sim('COND_Sim'); % End of COND_Driver.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% COND_Dynamics.m % COND_Dynamics.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% function COND_Dynamics(block) % Dynamic modeling of a condenser % Primary fluid is twophase and twocomponent (twophase in, subcooled % liquid out) and secondary fluid is incompressible subcooled liquid. % % This SFunction is utilized in the SIMULINK file named COND_Sim.mdl % and initialized by COND_Driver.m % Written by ChoonJae Ryu, Energy and Gas Dynamics Lab (Prof. William E. % Lear). % Department of Mechanical and Aerospace Engineering, University of Florida %% %% The setup method is used to setup the basic attributes of the %% S function such as ports, parameters, etc. Do not add any other %% calls to the main body of the function. %% setup(block); %endfunction %% Function: setup =================================================== %% Abstract: %% Set up the Sfunction block's basic characteristics such as: %% Input ports %% Output ports %% Dialog parameters %% Options %% %% Required : Yes %% CMex counterpart: mdlInitializeSizes %% function setup(block) % Register parameters block.NumDialogPrms = 15; % Register number of ports block.NumInputPorts = 1; block.NumOutputPorts = 1; % Setup port properties to be inherited or dynamic block.SetPreCompInpPortInfoToDynamic; block.SetPreCompOutPortInfoToDynamic; % Override input port properties block.InputPort(1).Dimensions = 7; % size_inputs %block.InputPort(1).DatatypeID = 0; % double

PAGE 194

194 %block.InputPort(1).Complexity = 'Real'; block.InputPort(1).DirectFeedthrough = 0; % size_feedthrough % Override output port properties block.OutputPort(1).Dimensions = 9; % size_outputs %block.OutputPort(1).DatatypeID = 0; % double %block.OutputPort(1).Complexity = 'Real'; % Register sample times % [0 offset] : Continuous sample time % [positive_num offset] : Discrete sample time % % [1, 0] : Inherited sample time % [2, 0] : Variable sample time block.SampleTimes = [0 0]; % States block.NumContStates = 6; % Specify the block simStateComliance. The allowed values are: % 'UnknownSimState', < The default setting; warn and assume DefaultSimState % 'DefaultSimState', < Same sim state as a builtin block % 'HasNoSimState', < No sim state % 'CustomSimState', < Has GetSimState and SetSimState methods % 'DisallowSimState' < Error out when saving or restoring the model sim state block.SimStateCompliance = 'DefaultSimState'; %% ----------------------------------------------------------------%% The Mfile Sfunction uses an internal registry for all %% block methods. You should register all relevant methods %% (optional and required) as illustrated below. You may choose %% any suitable name for the methods and implement these methods %% as local functions within the same file. See comments %% provided for each function for more information. %% ----------------------------------------------------------------%block.RegBlockMethod('PostPropagationSetup', @DoPostPropSetup); block.RegBlockMethod('InitializeConditions', @InitializeConditions); %block.RegBlockMethod('Start', @Start); block.RegBlockMethod('Outputs', @Outputs); % Required %block.RegBlockMethod('Update', @Update); block.RegBlockMethod('Derivatives', @Derivatives); %block.RegBlockMethod('Terminate', @Terminate); % Required %end setup %% %% PostPropagationSetup: %% Functionality : Setup work areas and state variables. Can %% also register runtime methods here %% Required : No %% CMex counterpart: mdlSetWorkWidths %%

PAGE 195

195 %function DoPostPropSetup(block) %block.NumDworks = 1; % % block.Dwork(1).Name = 'x1'; % block.Dwork(1).Dimensions = 1; % block.Dwork(1).DatatypeID = 0; % double % block.Dwork(1).Complexity = 'Real'; % real % block.Dwork(1).UsedAsDiscState = true; % % block.Dwork(2).Name = 'x2'; % block.Dwork(2).Dimensions = 1; % block.Dwork(2).DatatypeID = 0; % double % block.Dwork(2).Complexity = 'Real'; % real % block.Dwork(2).UsedAsDiscState = true; % %% %% InitializeConditions: %% Functionality : Called at the start of simulation and if it is %% present in an enabled subsystem configured to reset %% states, it will be called when the enabled subsystem %% restarts execution to reset the states. %% Required : No %% CMEX counterpart: mdlInitializeConditions %% function InitializeConditions(block) lc1_0 = block.DialogPrm(1).Data; xc1_0 = block.DialogPrm(2).Data; xc2_0 = block.DialogPrm(3).Data; Tcr2_0 = block.DialogPrm(4).Data; Tca1_0 = block.DialogPrm(5).Data; Tca2_0 = block.DialogPrm(6).Data; block.ContStates.Data = [lc1_0; xc1_0; xc2_0; Tcr2_0; Tca1_0; Tca2_0]; %end InitializeConditions %% %% Start: %% Functionality : Called once at start of model execution. If you %% have states that should be initialized once, this %% is the place to do it. %% Required : No %% CMEX counterpart: mdlStart %% %function Start(block) %endfunction %% %% Outputs: %% Functionality : Called to generate block outputs in %% simulation step %% Required : Yes

PAGE 196

196 %% CMEX counterpart: mdlOutputs %% function Outputs(block) % parameters and constants Lc = block.DialogPrm(7).Data; % m Ac = block.DialogPrm(8).Data; % m2 UDc1 = block.DialogPrm(9).Data; % kW/mK UDc2 = block.DialogPrm(10).Data; % kW/mK % inputs [Pc dPcdt xci mdci mdca Tcai] Pc = block.InputPort(1).Data(1); if Pc == 0 % only for initial output (t=0) Pc = block.DialogPrm(11).Data; dPcdt = 0; xci = block.DialogPrm(12).Data; hci = block.DialogPrm(13).Data; mdci = block.DialogPrm(14).Data; Tcai = block.DialogPrm(15).Data; else dPcdt = block.InputPort(1).Data(2); xci = block.InputPort(1).Data(3); hci = block.InputPort(1).Data(4); mdci = block.InputPort(1).Data(5); Tcai = block.InputPort(1).Data(7); end % state variables, xc = []' lc1 = block.ContStates.Data(1); xc1 = block.ContStates.Data(2); if xc1>1 xc1=1; end xc2 = block.ContStates.Data(3); if xc2>1 xc2=1; end Tcr2 = block.ContStates.Data(4); Tca1 = block.ContStates.Data(5); Tca2 = block.ContStates.Data(6); lc2 = Lc lc1; % Thermodyanmic properties [Tcr1 rhoc1 rhoc1hc1 drhoc1dPc drhoc1hc1dPc drhoc1dxc1 drhoc1hc1dxc1 hc12] ... = COND_2ph(Pc,xc1); xc12 = xc1; hc1 = (hci + hc12)/2; [rhoc2 hc2 drhoc2dPc dhc2dPc drhoc2dxc2 dhc2dxc2 ... drhoc2dTcr2 dhc2dTcr2] = liqp(Tcr2,Pc,xc2); xco = xc2; hco = 2*hc2 hc12; Tca21 = 2*Tca2 Tcai; Tcao = 2*Tca1 Tca21; c12 = Ac*lc1*rhoc1; c21 = Ac*(rhoc1*(xc12 xco) + rhoc2*(xco xc2)); c22 = Ac*lc1*(xc12 xco)*drhoc1dxc1;

PAGE 197

197 c23 = Ac*lc2*((xc2 xco)*drhoc2dxc2 + rhoc2); c24 = Ac*lc2*(xc2 xco)*drhoc2dTcr2; c31 = Ac*(rhoc1hc1 rhoc1*hc12); c32 = Ac*lc1*(drhoc1hc1dxc1 hc12*drhoc1dxc1); c41 = Ac*(rhoc1*(hc12 hco) + rhoc2*(hco hc2)); c42 = Ac*lc1*(hc12 hco)*drhoc1dxc1; c43 = Ac*lc2*((hc2 hco)*drhoc2dxc2 + rhoc2*dhc2dxc2); c44 = Ac*lc2*((hc2 hco)*drhoc2dTcr2 + rhoc2*dhc2dTcr2); fc1 = mdci*(xci xc12); fc2 = mdci*(xc12 xco) (Ac*lc1*(xc12 xco)*drhoc1dPc ... + Ac*lc2*(xc2 xco)*drhoc2dPc)*dPcdt; fc3 = mdci*(hci hc12) + UDc1*lc1*(Tca1 Tcr1) ... Ac*lc1*(drhoc1hc1dPc hc12*drhoc1dPc 1)*dPcdt; fc4 = mdci*(hc12 hco) + UDc2*lc2*(Tca2 Tcr2) ... (Ac*lc1*(hc12 hco)*drhoc1dPc ... + Ac*lc2*((hc2 hco)*drhoc2dPc + rhoc2*dhc2dPc 1))*dPcdt; xdc1 = fc1/c12; ldc1 = fc3/c31 c32/c31*xdc1; xdc2 = (c44*(fc2 c21*ldc1 c22*xdc1) c24*(fc4 c41*ldc1 ... c42*xdc1))/(c23*c44 c24*c43); Tdcr2 = (c43*(fc2 c 21*ldc1 c22*xdc1) + c23*(fc4 c41*ldc1 ... c42*xdc1))/(c23*c44 c24*c43); mdco = mdci Ac*(rhoc1 rhoc2)*ldc1 ... Ac*lc1*(drhoc1dPc*dPcdt + drhoc1dxc1*xdc1) ... Ac*lc2*(drhoc2dPc*dPcdt + drhoc2dxc2*xdc2 + drhoc2dTcr2*Tdcr2); block.OutputPort(1).Data = [mdco; xco; hco; Tcao; lc1; xc1; xc2; hc12; hc2]; %block.OutputPort(1).Data = block.Dwork(1).Data + block.InputPort(1).Data; %end Outputs %% %% Update: %% Functionality : Called to update discrete states %% during simulation step %% Required : No %% CMEX counterpart: mdlUpdate %% %function Update(block) %block.Dwork(1).Data = block.InputPort(1).Data; %end Update %% %% Derivatives: %% Functionality : Called to update derivatives of %% continuous states during simulation step %% Required : No %% CMEX counterpart: mdlDerivatives

PAGE 198

198 %% function Derivatives(block) % parameters or constants Lc = block.DialogPrm(7).Data; % m Ac = block.DialogPrm(8).Data; % m2 UDc1 = block.DialogPrm(9).Data; % kW/mK UDc2 = block.DialogPrm(10).Data; % kW/mK rhoca = 998; % kg/m3 Cpca = 4.18; % kJ/kgK Aca = Ac; % m2 % inputs [Pc dPcdt xci mdci mdca Tcai] Pc = block.InputPort(1).Data(1); dPcdt = block.InputPort(1).Data(2); xci = block.InputPort(1).Data(3); hci = block.InputPort(1).Data(4); mdci = block.InputPort(1).Data(5); mdca = block.InputPort(1).Data(6); Tcai = block.InputPort(1).Data(7); % state variables, xc = []' lc1 = block.ContStates.Data(1); xc1 = block.ContStates.Data(2); if xc1>1 xc1=1; end xc2 = block.ContStates.Data(3); if xc2>1 xc2=1; end Tcr2 = block.ContStates.Data(4); Tca1 = block.ContStates.Data(5); Tca2 = block.ContStates.Data(6); lc2 = Lc lc1; % Thermodyanmic properties [Tcr1 rhoc1 rhoc1hc1 drhoc1dPc drhoc1hc1dPc drhoc1dxc1 drhoc1hc1dxc1 hc12] ... = COND_2ph(Pc,xc1); xc12 = xc1; [rhoc2 hc2 drhoc2dPc dhc2dPc drhoc2dxc2 dhc2dxc2 ... drhoc2dTcr2 dhc2dTcr2] = liqp(Tcr2,Pc,xc2); xco = xc2; hco = 2*hc2 hc12; Tca21 = 2*Tca2 Tcai; Tcao = 2*Tca1 Tca21; c12 = Ac*lc1*rhoc1; c21 = Ac*(rhoc1*(xc12 xco) + rhoc2*(xco xc2)); c22 = Ac*lc1*(xc12 xco)*drhoc1dxc1; c23 = Ac*lc2*((xc2 xco)*drhoc2dxc2 + rhoc2); c24 = Ac*lc2*(xc2 xco)*drhoc2dTcr2; c31 = Ac*(rhoc1hc1 rhoc1*hc12); c32 = Ac*lc1*(drhoc1hc1dxc1 hc12*drhoc1dxc1); c41 = Ac*(rhoc1*(hc12 hco) + rhoc2*(hco hc2));

PAGE 199

199 c42 = Ac*lc1*(hc12 hco)*drhoc1dxc1; c43 = Ac*lc2*((hc2 hco)*drhoc2dxc2 + rhoc2*dhc2dxc2); c44 = Ac*lc2*((hc2 hco)*drhoc2dTcr2 + rhoc2*dhc2dTcr2); c55 = rhoca*Aca*Cpca*lc1; c66 = rhoca*Aca*Cpca*lc2; fc1 = mdci*(xci xc12); fc2 = mdci*(xc12 xco) (Ac*lc1*(xc12 xco)*drhoc1dPc ... + Ac*lc2*(xc2 xco)*drhoc2dPc)*dPcdt; fc3 = mdci*(hci hc12) + UDc1*lc1*(Tca1 Tcr1) ... Ac*lc1*(drhoc1hc1dPc hc12*drhoc1dPc 1)*dPcdt; fc4 = mdci*(hc12 hco) + UDc2*lc2*(Tca2 Tcr2) ... (Ac*lc1*(hc12 hco)*drhoc1dPc ... + Ac*lc2*((hc2 hco)*drhoc2dPc + rhoc2*dhc2dPc 1))*dPcdt; fc5 = mdca*Cpca*(Tca21 Tcao) + UDc1*lc1*(Tcr1 Tca1); fc6 = mdca*Cpca*(Tcai Tca21) + UDc2*lc2*(Tcr2 Tca2); xdc1 = fc1/c12; ldc1 = fc3/c31 c32/c31*xdc1; xdc2 = (c44*(fc2 c21*ldc1 c22*xdc1) c24*(fc4 c41*ldc1 ... c42*xdc1))/(c23*c44 c24*c43); Tdcr2 = (c43*(fc2 c21*ldc1 c22*xdc1) + c23*(fc4 c41*ldc1 ... c42*xdc1))/(c23*c44 c24*c43); Tdca1 = fc5/c55; Tdca2 = fc6/c66; block.Derivatives.Data = [ldc1; xdc1; xdc2; Tdcr2; Tdca1; Tdca2]; %end Derivatives %% %% Terminate: %% Functionality : Called at the end of simulation for cleanup %% Required : Yes %% CMEX counterpart: mdlTerminate %% %function Terminate(block) %end Terminate % End of COND_Dynamic.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% COND_2ph.m % COND_2ph.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% function [Tcr rhoc rhochc drhocdPc drhochcdPc drhocdxc drhochcdxc hcb] ... = COND_2ph(P,xc1) % Thermodyanmic properties of the twophase fluid in a condenser. % % This function is utilized in COND_Dynamics.m and COND_pxz.mat file is % necessary. % Written by ChoonJae Ryu, Energy and Gas Dynamics Lab (Prof. William E.

PAGE 200

200 % Lear). % Department of Mechanical and Aerospace Engineering, University of Florida load COND_pxz.mat dP = 1e3; xtol = 1e6; P1 = P dP/2; P2 = P + dP/2; if xc1 > 1xtol x1 = xc1 xtol/2; x2 = 1; elseif xc1 < xtol x1 = 0; x2 = xc1 + xtol/2; else x1 = xc1 xtol/2; x2 = xc1 + xtol/2; end dx = x2 x1; pp = [P1 P P2]; xx = [x1 xc1 x2]; rcp = interp2(xc,Pc',rhoc1,xx,pp'); rhcp = interp2(xc,Pc',rhoc1hc1,xx,pp'); Tcr = interp2(xc,Pc',Tcr1,xx(2),pp(2)); hcb = interp2(xc,Pc',hc12,xx(2),pp(2)); rhoc = rcp(2,2); rhochc = rhcp(2,2); drhocdPc = (rcp(3,2) rcp(1,2))/dP; drhochcdPc = (rhcp(3,2) rhcp(1,2))/dP; drhocdxc = (rcp(2,3) rcp(2,1))/dx; drhochcdxc = (rhcp(2,3) rhcp(2,1))/dx; % End of COND_2ph.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Absorber ABS_Driver.m % ABS_Driver.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% clear all %% initial values %% % Initial inputs % Pa = 8.339618761000000e+002; xai = 0.62; hai = 3.306850577071115e+002; mdai = 0.216574986421376; Taai = 2.887111000000000e+002;

PAGE 201

201 mdaa = 1.658345174363795; % Geometry and parameters % La = 1.3208; Aa = 0.007862719188542; UDa1 = 3.089813894119425; UDa2 = 4.269195056210765; % Initial values of state variables % la1 = 1.147922798833574; xa1 = 0.620000000000000; xa2 = 0.620000000000000; Tar2 = 3.061617943044990e+002; Taa1 = 2.963320428308150e+002; Taa2 = 2.895931428308149e+002; %% Step Inputs %% stime = 10; % Pressure Paf = Pa; Paslope = 0; % step input for dPadt Past = stime; % sec. start time % Ammonia mass fraction and enthalpy xaif = xai; % final value xaist = stime; % step time % Enthalpy haif = hai; haist = stime; % Mass flow rate refrigerant mdaif = mdai; % final value mdaist = stime; % step time % Mass flow rate secondary fluid mdaaf = mdaa; % final value mdaast = stime; % step time % Inlet temperature secondary fluid Taaif = Taai; % final value Taaist = stime; % step time %% Run Simulink Model %% sim('ABS_Sim'); % End of ABS_Driver.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% ABS_Dynamics.m % ABS_Dynamics.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% function ABS_Dynamics(block) % Dynamic modeling of an absorber

PAGE 202

202 % % Written by ChoonJae Ryu, Energy and Gas Dynamics Lab (Prof. Lear). % Department of Mechanical and Aerospace Engineering, University of Florida %% %% The setup method is used to setup the basic attributes of the %% Sfunction such as ports, parameters, etc. Do not add any other %% calls to the main body of the function. %% setup(block); %endfunction %% Function: setup =================================================== %% Abstract: %% Set up the Sfunction block's basic characteristics such as: %% Input ports %% Output ports %% Dialog parameters %% Options %% %% Required : Yes %% CMex counterpart: mdlInitializeSizes %% function setup(block) % Register parameters block.NumDialogPrms = 15; % Register number of ports block.NumInputPorts = 1; block.NumOutputPorts = 1; % Setup port properties to be inherited or dynamic block.SetPreCompInpPortInfoToDynamic; block.SetPreCompOutPortInfoToDynamic; % Override input port properties block.InputPort(1).Dimensions = 7; % size_inputs %block.InputPort(1).DatatypeID = 0; % double %block.InputPort(1).Complexity = 'Real'; block.InputPort(1).DirectFeedthrough = 0; % size_feedthrough % Override output port properties block.OutputPort(1).Dimensions = 5; % size_outputs %block.OutputPort(1).DatatypeID = 0; % double %block.OutputPort(1).Complexity = 'Real'; % Register sample times % [0 offset] : Continuous sample time % [positive_num offset] : Discrete sample time % % [1, 0] : Inherited sample time % [2, 0] : Variable sample time block.SampleTimes = [0 0];

PAGE 203

203 % States block.NumContStates = 6; % xe = [le1 Pe heo Tew1 Tew2]' % Specify the block simStateComliance. The allowed values are: % 'UnknownSimState', < The default setting; warn and assume DefaultSimState % 'DefaultSimState', < Same sim state as a builtin block % 'HasNoSimState', < No sim state % 'CustomSimState', < Has GetSimState and SetSimState methods % 'DisallowSimState' < Error out when saving or restoring the model sim state block.SimStateCompliance = 'DefaultSimState'; %% ----------------------------------------------------------------%% The Mfile Sfunction uses an internal registry for all %% block methods. You should register all relevant methods %% (optional and required) as illustrated below. You may choose %% any suitable name for the methods and implement these methods %% as local functions within the same file. See comments %% provided for each function for more information. %% ----------------------------------------------------------------%block.RegBlockMethod('PostPropagationSetup', @DoPostPropSetup); block.RegBlockMethod('InitializeConditions', @InitializeConditions); %block.RegBlockMethod('Start', @Start); block.RegBlockMethod('Outputs', @Outputs); % Required %block.RegBlockMethod('Update', @Update); block.RegBlockMethod('Derivatives', @Derivatives); %block.RegBlockMethod('Terminate', @Terminate); % Required %end setup %% %% PostPropagationSetup: %% Functionality : Setup work areas and state variables. Can %% also register runtime methods here %% Required : No %% CMex counterpart: mdlSetWorkWidths %% %function DoPostPropSetup(block) %block.NumDworks = 1; % % block.Dwork(1).Name = 'x1'; % block.Dwork(1).Dimensions = 1; % block.Dwork(1).DatatypeID = 0; % double % block.Dwork(1).Complexity = 'Real'; % real % block.Dwork(1).UsedAsDiscState = true; % % block.Dwork(2).Name = 'x2'; % block.Dwork(2).Dimensions = 1; % block.Dwork(2).DatatypeID = 0; % double % block.Dwork(2).Complexity = 'Real'; % real % block.Dwork(2).UsedAsDiscState = true; %

PAGE 204

204 %% %% InitializeConditions: %% Functionality : Called at the start of simulation and if it is %% present in an enabled subsystem configured to reset %% states, it will be called when the enabled subsystem %% restarts execution to reset the states. %% Required : No %% C MEX counterpart: mdlInitializeConditions %% function InitializeConditions(block) la1_0 = block.DialogPrm(1).Data; xa1_0 = block.DialogPrm(2).Data; xa2_0 = block.DialogPrm(3).Data; Tar2_0 = block.DialogPrm(4).Data; Taa1_0 = block.DialogPrm(5).Data; Taa2_0 = block.DialogPrm(6).Data; block.ContStates.Data = [la1_0; xa1_0; xa2_0; Tar2_0; Taa1_0; Taa2_0]; %end InitializeConditions %% %% Start: %% Functionality : Called once at start of model execution. If you %% have states that should be initialized once, this %% is the place to do it. %% Required : No %% CMEX counterpart: mdlStart %% %function Start(block) %endfunction %% %% Outputs: %% Functionality : Called to generate block outputs in %% simulation step %% Required : Yes %% CMEX counterpart: mdlOutputs %% function Outputs(block) % parameters and constants La = block.DialogPrm(7).Data; % m Aa = block.DialogPrm(8).Data; % m2 UDa1 = block.DialogPrm(9).Data; % kW/mK UDa2 = block.DialogPrm(10).Data; % kW/mK % inputs Pa = block.InputPort(1).Data(1); if Pa == 0 % only for initial output (t = 0) Pa = block.DialogPrm(11).Data; dPadt = 0; xai = block.DialogPrm(12).Data;

PAGE 205

205 hai = block.DialogPrm(13).Data; mdai = block.DialogPrm(14).Data; Taai = block.DialogPrm(15).Data; else dPadt = block.InputPort(1).Data(2); xai = block.InputPort(1).Data(3); hai = block.InputPort(1).Data(4); mdai = block.InputPort(1).Data(5); Taai = block.InputPort(1).Data(7); end % state variables, xc = []' la1 = block.ContStates.Data(1); xa1 = block.ContStates.Data(2); if xa1>1 xa1=1; end xa2 = block.ContStates.Data(3); if xa2>1 xa2=1; end Tar2 = block.ContStates.Data(4); Taa1 = block.ContStates.Data(5); Taa2 = block.ContStates.Data(6); la2 = La la1; % Thermodyanmic properties [Tar1 rhoa1 rhoa1ha1 drhoa1dPa drhoa1ha1dPa drhoa1dxa1 drhoa1ha1dxa1 ha12] ... = ABS_2ph(Pa,xa1,hai); xa12 = xa1; %Ta12 = Tb_Bogart(Pa,xa12); %ha12 = liqprop_ibrahim(Ta12,Pa,xa12); [rhoa2 ha2 drhoa2dPa dha2dPa drhoa2dxa2 dha2dxa2 ... drhoa2dTar2 dha2dTar2] = liqp(Tar2,Pa,xa2); xao = xa2; hao = 2*ha2 ha12; %Tao = h2t_liq(Pa,hao,xao); Taa21 = 2*Taa2 Taai; Taao = 2*Taa1 Taa21; a12 = Aa*la1*rhoa1; a21 = Aa*(rhoa1*(xa12 xao) + rhoa2*(xao xa2)); a22 = Aa*la1*(xa12 xao)*drhoa1dxa1; a23 = Aa*la2*((xa2 xao)*drhoa2dxa2 + rhoa2); a24 = Aa*la2*(xa2 xao)*drhoa2dTar2; a31 = Aa*(rhoa1ha1 rhoa1*ha12); a32 = Aa*la1*(drhoa1ha1dxa1 ha12*drhoa1dxa1); a41 = Aa*(rhoa1*(ha12 hao) + rhoa2*(hao ha2)); a42 = Aa*la1*(ha12 hao)*drhoa1dxa1; a43 = Aa*la2*((ha2 hao)*drhoa2dxa2 + rhoa2*dha2dxa2); a44 = Aa*la2*((ha2 hao)*drhoa2dTar2 + rhoa2*dha2dTar2); fa1 = mdai*(xai xa12); fa2 = mdai*(xa12 xao) (Aa*la1*(xa12 xao)*drhoa1dPa ...

PAGE 206

2 06 + Aa*la2*(xa2 xao)*drhoa2dPa)*dPadt; fa3 = mdai*(hai ha12) + UDa1*la1*(Taa1 Tar1) ... Aa*la1*(drhoa1ha1dPa ha12*drhoa1dPa 1)*dPadt; fa4 = mdai*(ha12 hao) + UDa2*la2*(Taa2 Tar2) ... (Aa*la1*(ha12 hao)*drhoa1dPa ... + Aa*la2*((ha2 hao)*drhoa2dPa + rhoa2*dha2dPa 1))*dPadt; xda1 = fa1/a12; lda1 = fa3/a31 a32/a31*xda1; xda2 = (a44*(fa2 a21*lda1 a22*xda1) a24*(fa4 a41*lda1 ... a42*xda1))/(a23*a44 a24*a43); Tdar2 = (a43*(fa2 a21*lda1 a22*xda1) + a23*(fa4 a41*lda1 ... a42*xda1))/(a23*a44 a24*a43); mdao = mdai Aa*(rhoa1 rhoa2)*lda1 ... Aa*la1*(drhoa1dPa*dPadt + drhoa1dxa1*xda1) ... Aa*la2*(drhoa2dPa*dPadt + drhoa2dxa2*xda2 + drhoa2dTar2*Tdar2); block.OutputPort(1).Data = [mdao; xao; hao; Taao; la1]; %block.OutputPort(1).Data = block.Dwork(1).Data + block.InputPort(1).Data; %end Outputs %% %% Update: %% Functionality : Called to update discrete states %% during simulation step %% Required : No %% CMEX counterpart: mdlUpdate %% %function Update(block) %block.Dwork(1).Data = block.InputPort(1).Data; %end Update %% %% Derivatives: %% Functionality : Called to update derivatives of %% continuous states during simulation step %% Required : No %% CMEX counterpart: mdlDerivatives %% function Derivatives(block) % parameters or constants La = block.DialogPrm(7).Data; % m Aa = block.DialogPrm(8).Data; % m2 UDa1 = block.DialogPrm(9).Data; % kW/mK UDa2 = block.DialogPrm(10).Data; % kW/mK rhoaa = 998; % kg/m3 Cpaa = 4.18; % kJ/kgK Aaa = Aa; % m2 % inputs [Pc dPcdt xci mdci mdca Tcai] Pa = block.InputPort(1).Data(1);

PAGE 207

207 dPadt = block.InputPort(1).Data(2); xai = block.InputPort(1).Data(3); hai = block.InputPort(1).Data(4); mdai = block.InputPort(1).Data(5); mdaa = block.InputPort(1).Data(6); Taai = block.InputPort(1).Data(7); % state variables, xc = []' la1 = block.ContStates.Data(1); xa1 = block.ContStates.Data(2); if xa1>1 xa1=1; end xa2 = block.ContStates.Data(3); if xa2>1 xa2=1; end Tar2 = block.ContStates.Data(4); Taa1 = block.ContStates.Data(5); Taa2 = block.ContStates.Data(6); la2 = La la1; % Thermodyanmic properties [Tar1 rhoa1 rhoa1ha1 drhoa1dPa drhoa1ha1dPa drhoa1dxa1 drhoa1ha1dxa1 ha12] ... = ABS_2ph(Pa,xa1,hai); xa12 = xa1; %Ta12 = Tb_Bogart(Pa,xa12); %ha12 = liqprop_ibrahim(Ta12,Pa,xa12); [rhoa2 ha2 drhoa2dPa dha2dPa drhoa2dxa2 dha2dxa2 ... drhoa2dTar2 dha2dTar2] = liqp(Tar2,Pa,xa2); xao = xa2; hao = 2*ha2 ha12; Taa21 = 2*Taa2 Taai; Taao = 2*Taa1 Taa21; a12 = Aa*la1*rhoa1; a21 = Aa*(rhoa1*(xa12 xao) + rhoa2*(xao xa2)); a22 = Aa*la1*(xa12 xao)*drhoa1dxa1; a23 = Aa*la2*((xa2 xao)*drhoa2dxa2 + rhoa2); a24 = Aa*la2*(xa2 x ao)*drhoa2dTar2; a31 = Aa*(rhoa1ha1 rhoa1*ha12); a32 = Aa*la1*(drhoa1ha1dxa1 ha12*drhoa1dxa1); a41 = Aa*(rhoa1*(ha12 hao) + rhoa2*(hao ha2)); a42 = Aa*la1*(ha12 hao)*drhoa1dxa1; a43 = Aa*la2*((ha2 hao)*drhoa2dxa2 + rhoa2*dha2dxa2); a44 = Aa*la2*((ha2 hao)*drhoa2dTar2 + rhoa2*dha2dTar2); a55 = rhoaa*Aaa*Cpaa*la1; a66 = rhoaa*Aaa*Cpaa*la2; fa1 = mdai*(xai xa12); fa2 = mdai*(xa12 xao) (Aa*la1*(xa12 xao)*drhoa1dPa ...

PAGE 208

208 + Aa*la2*(xa2 xao)*drhoa2dPa)*dPadt; fa3 = mdai*(hai ha12) + UDa1*la1*(Taa1 Tar1) ... Aa*la1*(drhoa1ha1dPa ha12*drhoa1dPa 1)*dPadt; fa4 = mdai*(ha12 hao) + UDa2*la2*(Taa2 Tar2) ... (Aa*la1*(ha12 hao)*drhoa1dPa ... + Aa*la2*((ha2 hao)*drhoa2dPa + rhoa2*dha2dPa 1))*dPadt; fa5 = mdaa*Cpaa*(Taa21 Taao) + UDa1*la1*(Tar1 Taa1); fa6 = mdaa*Cpaa*(Taai Taa21) + UDa2*la2*(Tar2 Taa2); xda1 = fa1/a12; lda1 = fa3/a31 a32/a31*xda1; xda2 = (a44*(fa2 a21*lda1 a22*xda1) a24*(fa4 a41*lda1 ... a42*xda1))/(a23*a44 a24*a43); Tdar2 = (a43*(fa2 a21*lda1 a22*xda1) + a23*(fa4 a41*lda1 ... a42*xda1))/(a23*a44 a24*a43); Tdaa1 = fa5/a55; Tdaa2 = fa6/a66; block.Derivatives.Data = [lda1; xda1; xda2; Tdar2; Tdaa1; Tdaa2]; %end Derivatives %% %% Terminate: %% Functionality : Called at the end of simulation for cleanup %% Required : Yes %% CMEX counterpart: mdlTerminate %% %function Terminate(block) %end Terminate % End of ABS_Dynamics.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% ABS_2ph.m % ABS_2ph.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% function [Ta rhoa rhoaha drhoadPa drhoahadPa drhoadxa drhoahadxa hab] ... = ABS_2ph(P,xa1,hai) load ABS_pxz1.mat m = 100; dP = 1e3; xtol = 1e6; P1 = P dP/2; P2 = P + dP/2; if xa1 > 1xtol x1 = xa1 xtol/2; x2 = 1; elseif xa1 < xtol x1 = 0;

PAGE 209

209 x2 = xa1 + xtol/2; else x1 = xa1 xtol/2; x2 = xa1 + xtol/2; end dx = x2 x1; za = 0:1/n:1; [xxa ppa zza] = meshgrid(xa,Pa,za); %z = 0:0.01:1; z = [0 1]; [xi pii zi] = meshgrid(xa1,P,z); %Taz = griddata3(xxa,ppa,zza,Tar1,xi,pi,zi); ha1z = interp3(xxa,ppa,zza,ha1,xi,pii,zi); %haz(1,:) = ha1z(1,1,:); %zai = interp1(haz,z,hai); %haz0 = vapprop_ibrahim(Taz(1),P,xa1); %haz1 = liqprop_ibrahim(Taz(2),P,xa1); zai = (hai ha1z(1))/(ha1z(2) ha1z(1)); xx = [x1 xa1 x2]; pp = [P1 P P2]; zz = zai:(1zai)/m:1; [xxi ppi zzi] = meshgrid(xx,pp,zz); T1 = interp3(xxa,ppa,zza,Tar1,xxi,ppi,zzi); Ta = mean(T1(2,2,:)); rho1 = interp3(xxa,ppa,zza,rhoa1,xxi,ppi,zzi); rhoa = mean(rho1(2,2,:)); rp1 = mean(rho1(1,2,:)); rp2 = mean(rho1(3,2,:)); rx1 = mean(rho1(2,1,:)); rx2 = mean(rho1(2,3,:)); rho1h1 = interp3(xxa,ppa,zza,rhoa1ha1,xxi,ppi,zzi); rhoaha = mean(rho1h1(2,2,:)); rhp1 = mean(rho1h1(1,2,:)); rhp2 = mean(rho1h1(3,2,:)); rhx1 = mean(rho1h1(2,1,:)); rhx2 = mean(rho1h1(2,3,:)); drhoadPa = (rp2 rp1)/dP; drhoahadPa = (rhp2 rhp1)/dP; drhoadxa = (rx2 rx1)/dx; drhoahadxa = (rhx2 rhx1)/dx; ha12(:,:) = ha1(:,:,101); hab = interp2(xa,Pa',ha12,xa1,P); % End of ABS_2ph.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

PAGE 210

210 Generator GEN_Driver.m % GEN_Driver.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% clear all %% initial values %% % Initial inputs % Pg = 1.791850466000000e+003; xgi = 0.62; hgi = 90.175825309765756; mdgi = 0.185995325421376; Tgai = 5.728500000000000e+002; mdga = 1.152879667; % Geometry and parameters % Lg = 0.6096; Ag = 0.041071438769110; Cpga = 1.251729794677541; UDg1 = 0.965702895542424; UDg2 = 0.853902219139207; % Initial values of state variables % lg1 = 0.009865125203645; xg1 = 0.62; xg2 = 0.62; Tgr1 = 3.418070673356066e+002; Tga1 = 5.062929379110602e+002; Tga2 = 5.398429379110602e+002; %% Step Inputs %% stime = 10; % Pressure Pgf = Pg ; Pgslope = 0; % step input for dPadt Pgst = stime; % sec. start time % Ammonia mass fraction and enthalpy xgif = xgi; % final value xgist = stime; % step time % Enthalpy hgif = hgi; hgist = stime; % Mass flow rate refrigerant mdgif = mdgi; % final value mdgist = stime; % step time % Mass flow rate secondary fluid mdgaf = mdga; % final value mdgast = stime; % step time

PAGE 211

211 % Inlet temperature secondary fluid Tgaif = Tgai; % final value Tgaist = stime; % step time %% Run Simulink Model %% sim('GEN_Sim'); % End of GEN_Driver.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% GEN_Dynamics.m % GEN_Dynamics.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% function GEN_Dynamics(block) % Dynamic modeling of a generator % % Written by ChoonJae Ryu, Energy and Gas Dynamics Lab (Prof. Lear). % Department of Mechanical and Aerospace Engineering, University of Florida %% %% The setup method is used to setup the basic attributes of the %% Sfunction such as ports, parameters, etc. Do not add any other %% calls to the main body of the function. %% setup(block); %endfunction %% Function: setup =================================================== %% Abstract: %% Set up the Sfunction block's basic characteristics such as: %% Input ports %% Output ports %% Dialog parameters %% Options %% %% Required : Yes %% CMex counterpart: mdlInitializeSizes %% function setup(block) % Register parameters block.NumDialogPrms = 17; % Register number of ports block.NumInputPorts = 1; block.NumOutputPorts = 1; % Setup port properties to be inherited or dynamic block.SetPreCompInpPortInfoToDynamic; block.SetPreCompOutPortInfoToDynamic; % Override input port properties block.InputPort(1).Dimensions = 7; % size_inputs

PAGE 212

212 %block.InputPort(1).DatatypeID = 0; % double %block.InputPort(1).Complexity = 'Real'; block.InputPort(1).DirectFeedthrough = 0; % size_feedthrough % Override output port properties block.OutputPort(1).Dimensions = 9; % size_outputs %block.OutputPort(1).DatatypeID = 0; % double %block.OutputPort(1).Complexity = 'Real'; % Register sample times % [0 offset] : Continuous sample time % [positive_num offset] : Discrete sample time % % [1, 0] : Inherited sample time % [2, 0] : Variable sample time block.SampleTimes = [0 0]; % States block.NumContStates = 6; % xe = [le1 Pe heo Tew1 Tew2]' % Specify the block simStateComliance. The allowed values are: % 'UnknownSimState', < The default setting; warn and assume DefaultSimState % 'DefaultSimState', < Same sim state as a builtin block % 'HasNoSimState', < No sim state % 'CustomSimState', < Has GetSimState and SetSimState methods % 'DisallowSimState' < Error out when saving or restoring the model sim state block.SimStateCompliance = 'DefaultSimState'; %% ----------------------------------------------------------------%% The Mfile Sfunction uses an internal registry for all %% block methods. You should register all relevant methods %% (optional and required) as illustrated below. You may choose %% any suitable name for the methods and implement these methods %% as local functions within the same file. See comments %% provided for each function for more information. %% ----------------------------------------------------------------%block.RegBlockMethod('PostPropagationSetup', @DoPostPropSetup); block.RegBlockMethod('InitializeConditions', @InitializeConditions); %block.RegBlockMethod('Start', @Start); block.RegBlockMethod('Outputs', @Outputs); % Required %block.RegBlockMethod('Update', @Update); block.RegBlockMethod('Derivatives', @Derivatives); %block.RegBlockMethod('Terminate', @Terminate); % Required %end setup %% %% PostPropagationSetup: %% Functionality : Setup work areas and state variables. Can %% also register runtime methods here %% Required : No %% CMex counterpart: mdlSetWorkWidths

PAGE 213

213 %% %function DoPostPropSetup(block) %block.NumDworks = 1; % % block.Dwork(1).Name = 'x1'; % block.Dwork(1).Dimensions = 1; % block.Dwork(1).DatatypeID = 0; % double % block.Dwork(1).Complexity = 'Real'; % real % block.Dwork(1).UsedAsDiscState = true; % % block.Dwork(2).Name = 'x2'; % block.Dwork(2).Dimensions = 1; % block.Dwork(2).DatatypeID = 0; % double % block.Dwork(2).Complexity = 'Real'; % real % block.Dwork(2).UsedAsDiscState = true; % %% %% InitializeConditions: %% Functionality : Called at the start of simulation and if it is %% present in an enabled subsystem configured to reset %% states, it will be called when the enabled subsystem %% restarts execution to reset the states. %% Required : No %% CMEX counterpart: mdlInitializeConditions %% function InitializeConditions(block) lg1_0 = block.DialogPrm(1).Data; Tgr1_0 = block.DialogPrm(2).Data; xg1_0 = block.DialogPrm(3).Data; xg2_0 = block.DialogPrm(4).Data; Tga1_0 = block.DialogPrm(5).Data; Tga2_0 = block.DialogPrm(6).Data; block.ContStates.Data = [lg1_0; Tgr1_0; xg1_0; xg2_0; Tga1_0; Tga2_0]; %end InitializeConditions %% %% Start: %% Functionality : Called once at start of model execution. If you %% have states that should be initialized once, this %% is the place to do it. %% Required : No %% CMEX counterpart: mdlStart %% %function Start(block) %endfunction %% %% Outputs: %% Functionality : Called to generate block outputs in %% simulation step %% Required : Yes %% CMEX counterpart: mdlOutputs

PAGE 214

214 %% function Outputs(block) % parameters and constants Lg = block.DialogPrm(7).Data; % m Ag = block.DialogPrm(8).Data; % m2 UDg1 = block.DialogPrm(9).Data; % kW/mK UDg2 = block.DialogPrm(10).Data; % kW/mK Cpga = block.DialogPrm(11).Data; % inputs Pg = block.InputPort(1).Data(1); if Pg == 0 Pg = block.DialogPrm(12).Data; dPgdt = 0; xgi = block.DialogPrm(13).Data; hgi = block.DialogPrm(14).Data; mdgi = block.DialogPrm(15).Data; mdga = block.DialogPrm(16).Data; Tgai = block.DialogPrm(17).Data; else dPgdt = block.InputPort(1).Data(2); xgi = block.InputPort(1).Data(3); hgi = block.InputPort(1).Data(4); mdgi = block.InputPort(1).Data(5); mdga = block.InputPort(1).Data(6); Tgai = block.InputPort(1).Data(7); end % state variables, xc = []' lg1 = block.ContStates.Data(1); Tgr1 = block.ContStates.Data(2); xg1 = block.ContStates.Data(3); if xg1>1 xg1=1; end xg2 = block.ContStates.Data(4); if xg2>1 xg2=1; end Tga1 = block.ContStates.Data(5); Tga2 = block.ContStates.Data(6); lg2 = Lg lg1; % Thermodyanmic properties [rhog1 hg1 drhog1dPg dhg1dPg drhog1dxg1 dhg1dxg1 drhog1dTgr1 dhg1dTgr1] ... = liqp(Tgr1,Pg,xg1); Tgr2 = Tga2 + 2*(Tga2 Tgai)*mdga*Cpga/UDg2/lg2; [rhog2 hgo rhog2hg2 drhog2dPg drhog2hg2dPg drhog2dxg2 drhog2hg2dxg2 ... hgov hgol xgov xgol] = GEN_2ph(Pg,xg2,Tgr2); xg12 = xg1; Tg12 = Tb_Bogart(Pg,xg12); h g12 = liqprop_ibrahim(Tg12,Pg,xg12); xgo = xg2; Tga21 = 2*Tga2 Tgai; Tgao = 2*Tga1 Tga21; g13 = Ag*lg1*rhog1;

PAGE 215

215 g21 = Ag*rhog1*(xg12 xgo); g22 = Ag*lg1*(xg12 xgo)*drhog1dTgr1; g23 = Ag*lg1*(xg12 xgo)*drhog1dxg1; g24 = Ag*lg2*rhog2; g31 = Ag*rhog1*(hg1 hg12); g32 = Ag*lg1*((hg1 hg12)*drhog1dTgr1 + rhog1*dhg1dTgr1); g33 = Ag*lg1*((hg1 hg12)*drhog1dxg1 + rhog1*dhg1dxg1); g41 = Ag*(rhog1*(hg12 hgo) (rhog2hg2 rhog2*hgo)); g42 = Ag*lg1*(hg12 hgo)*drhog1dTgr1; g43 = Ag*lg1*(hg12 hgo)*drhog1dxg1; g44 = Ag*lg2*(drhog2hg2dxg2 hgo*drhog2dxg2); fg1 = mdgi*(xgi xg12); fg2 = mdgi*(xg12 xgo) Ag*lg1*(xg12 xgo)*drhog1dPg*dPgdt; fg3 = mdgi*(hgi hg12) + UDg1*lg1*(Tga1 Tgr1) ... Ag*lg1*((hg1 hg12)*drhog1dPg + rhog1*dhg1dPg 1)*dPgdt; %if fg3 < 1e10 fg3=0; end fg4 = mdgi*(hg12 hgo) + UDg2*lg2*(Tga2 Tgr2) ... (Ag*lg1*(hg12 hgo)*drhog1dPg ... + Ag*lg2*(drhog2hg2dPg hgo*drhog2dPg 1))*dPgdt; %if fg4 < 1e10 fg4=0; end G = [0 0 g13 0; g21 g22 g23 g24; g31 g32 g33 0; g41 g42 g43 g44]; Fg = [fg1; fg2; fg3; fg4]; xdg = G\ Fg; ldg1 = xdg(1); Tdgr1 = xdg(2); xdg1 = xdg(3); xdg2 = xdg(4); mdgo = mdgi Ag*(rhog1 rhog2)*ldg1 ... Ag*lg1*(drhog1dPg*dPgdt + drhog1dxg1*xdg1 + drhog1dTgr1*Tdgr1) ... Ag*lg2*(drhog2dPg*dPgdt + drhog2dxg2*xdg2); block.OutputPort(1).Data = [mdgo; xgo; hgo; Tgao; lg1; hgov; hgol; xgov; xgol]; %block.OutputPort(1).Data = block.Dwork(1).Data + block.InputPort(1).Data; %end Outputs %% %% Update: %% Functionality : Called to update discrete states %% during simulation step %% Required : No %% CMEX counterpart: mdlUpdate %% %function Update(block) %block.Dwork(1).Data = block.InputPort(1).Data; %end Update

PAGE 216

216 %% %% Derivatives: %% Functionality : Called to update derivatives of %% continuous states during simulation step %% Required : No %% CMEX counterpart: mdlDerivatives %% function Derivatives(block) % parameters or constants Lg = block.DialogPrm(7).Data; % m Ag = block.DialogPrm(8).Data; % m2 UDg1 = block.DialogPrm(9).Data; % kW/mK UDg2 = block.DialogPrm(10).Data; % kW/mK rhoga = 2; % kg/m3 Cpga = block.DialogPrm(11).Data; % kJ/kgK Aga = Ag; % m2 % inputs [Pc dPcdt xci mdci mdca Tcai] Pg = block.InputPort(1).Data(1); dPgdt = block.InputPort(1).Data(2); xgi = block.InputPort(1).Data(3); hgi = block.InputPort(1).Data(4); mdgi = block.InputPort(1).Data(5); mdga = block.InputPort(1).Data(6); Tgai = block.InputPort(1).Data(7); % state variables, xc = []' lg1 = block.ContStates.Data(1); Tgr1 = block.ContStates.Data(2); xg1 = block.ContStates.Data(3); if xg1>1 xg1=1; end xg2 = block.ContStates.Data(4); if xg2>1 xg2=1; end Tga1 = block.ContStates.Data(5); Tga2 = block.ContStates.Data(6); lg2 = Lg lg1; % Thermodyanmic properties [rhog1 hg1 drhog1dPg dhg1dPg drhog1dxg1 dhg1dxg1 drhog1dTgr1 dhg1dTgr1] ... = liqp(Tgr1,Pg,xg1); Tgr2 = Tga2 + 2*(Tga2 Tgai)*mdga*Cpga/UDg2/lg2; [rhog2 hgo rhog2hg2 drhog2dPg drhog2hg2dPg drhog2dxg2 drhog2hg2dxg2 ... hgov hgol xgov xgol] = GEN_2ph(Pg,xg2,Tgr2); xg12 = xg1; Tg12 = Tb_Bogart(Pg,xg12); hg12 = liqprop_ibrahim(Tg12,Pg,xg12); xgo = xg2; Tga21 = 2*Tga2 Tgai; Tgao = 2*Tga1 Tga21; g13 = Ag*lg1*rhog1; g21 = Ag*rhog1*(xg12 xgo);

PAGE 217

217 g22 = Ag*lg1*(xg12 xgo)*drhog1dTgr1; g23 = Ag*lg1*(xg12 xgo)*drhog1dxg1; g24 = Ag*lg2*rhog2; g31 = Ag*rhog1*(hg1 hg12); g32 = Ag*lg1*((hg1 hg12)*drhog1dTgr1 + rhog1*dhg1dTgr1); g33 = Ag*lg1*((hg1 hg12)*drhog1dxg1 + rhog1*dhg1dxg1); g41 = Ag*(rhog1*(hg12 hgo) (rhog2hg2 rhog2*hgo)); g42 = Ag*lg1*(hg12 hgo)*drhog1dTgr1; g43 = Ag*lg1*(hg12 hgo)*drhog1dxg1; g44 = Ag*lg2*(drhog2hg2dxg2 hgo*drhog2dxg2); g55 = rhoga*Aga*Cpga*lg1; g66 = rhoga*Aga*Cpga*lg2; fg1 = mdgi*(xgi xg12); fg2 = mdgi*(xg12 xgo) Ag*lg1*(xg12 xgo)*drhog1dPg*dPgdt; fg3 = mdgi*(hgi hg12) + UDg1*lg1*(Tga1 Tgr1) ... Ag*lg1*((hg1 hg12)*drhog1dPg + rhog1*dhg1dPg 1)*dPgdt; %if fg3 < 1e10 fg3=0; end fg4 = mdgi*(hg12 hgo) + UDg2*lg2*(Tga2 Tgr2) ... (Ag*lg1*(hg12 hgo)*drhog1dPg ... + Ag*lg2*(drhog2hg2dPg hgo*drhog2dPg 1))*dPgdt; %if fg4 < 1e10 fg4=0; end fg5 = mdga*Cpga*(Tga21 Tgao) + UDg1*lg1*(Tgr1 Tga1); fg6 = mdga*Cpga*(Tgai Tga21) + UDg2*lg2*(Tgr2 Tga2); G = [0 0 g13 0; g21 g22 g23 g24; g31 g32 g33 0; g41 g42 g43 g44]; Fg = [fg1; fg2; fg3; fg4]; xdg = G\ Fg; Tdga1 = fg5/g55; Tdga2 = fg6/g66; block.Derivatives.Data = [xdg; Tdga1; Tdga2]; %end Derivatives %% %% Terminate: %% Functionality : Called at the end of simulation for cleanup %% Required : Yes %% CMEX counterpart: mdlTerminate %% %function Terminate(block) %end Terminate % End of GEN_Dynamics.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% GEN_2ph.m % GEN_2ph.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

PAGE 218

218 function [rhog ho rhoghg drhogdPg drhoghgdPg drhogdxg drhoghgdxg ... hog hol xog xol] = GEN_2ph(P,xg2,T2) load GEN_pxz2.mat m = 100; dP = 1e3; xtol = 1e6; P1 = P dP/2; P2 = P + dP/2; if xg2 > 1xtol x1 = xg2 xtol/2; x2 = 1; elseif xg2 < xtol x1 = 0; x2 = xg2 + xtol/2; else x1 = xg2 xtol/2; x2 = xg2 + xtol/2; end dx = x2 x1; zg = 0:1/n:1; [xxg ppg zzg] = meshgrid(xg,Pg,zg); z = 0:0.01:1; [xi pii zi] = meshgrid(xg2,P,z); Tgmeanz = interp3(xxg,ppg,zzg,Tmean,xi,pii,zi); Tgmz(1,:) = Tgmeanz(1,1,:); zgo = interp1(Tgmz,z,T2); [xhi phi zhi] = meshgrid(xg2,P,zgo); ho = interp3(xxg,ppg,zzg,hg2,xhi,phi,zhi); hog = interp3(xxg,ppg,zzg,hgg,xhi,phi,zhi); hol = interp3(xxg,ppg,zzg,hgl,xhi,phi,zhi); xog = interp3(xxg,ppg,zzg,xgg,xhi,phi,zhi); xol = interp3(xxg,ppg,zzg,xgl,xhi,phi,zhi); xx = [x1 xg2 x2]; pp = [P1 P P2]; zz = 0:zgo/m:zgo; [xxi ppi zzi] = meshgrid(xx,pp,zz); rho2 = interp3(xxg,ppg,zzg,rhog2,xxi,ppi,zzi); rhog = mean(rho2(2,2,:)); rp1 = mean(rho2(1,2,:)); rp2 = mean(rho2(3,2,:)); rx1 = mean(rho2(2,1,:)); rx2 = mean(rho2(2,3,:)); rho2h2 = interp3(xxg,ppg,zzg,rhog2hg2,xxi,ppi,zzi); rhoghg = mean(rho2h2(2,2,:)); rhp1 = mean(rho2h2(1,2,:)); rhp2 = mean(rho2h2(3,2,:)); rhx1 = mean(rho2h2(2,1,:));

PAGE 219

219 rhx2 = mean(rho2h2(2,3,:)); drhogdPg = (rp2 rp1)/dP; drhoghgdPg = (rhp2 rhp1)/dP; drhogdxg = (rx2 rx1)/dx; drhoghgdxg = (rhx2 rhx1)/dx; % End of GEN_2ph.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Evaporator EVAP_Driver.m % EVAP_Driver.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% clear all %% initial values %% % initial inputs % Pe = 8.339618761000000e+002; xei = 0.998000000000000; hei = 1.252714784168311e+002; mdei = 0.061827110000000; Teai = 3.288222000000000e+002; mdea = 1.152879667000000; % Geometry and parameters % Le = 0.914400000000000; Ae = 0.059148967930983; Cpea = 2.382488973429181; UAe = 2.783884462820257; Les = 1.004577863913866; % Initial values of state variables % xe = 0.998000000000000; Tea = 3.166055500000000e+002; %% Step inputs %% stime = 10; % Pressure Pef = Pe;%/1.01; Peslope = 0;%0.5; % step input for dPadt Pest = stime; % sec. start time % Ammonia mass fraction and enthalpy xeif = xei*1.001; % final value xeist = stime; % step time % Enthalpy heif = hei;%*1.1; heist = stime; % Mass flow rate refrigerant mdeif = mdei;%*1.1; % final value

PAGE 220

220 mdeist = stime; % step time % Mass flow rate secondary fluid mdeaf = mdea;%*1.1; % final value mdeast = stime; % step time % Inlet temperature secondary fluid Teaif = Teai; % final value Teaist = stime; % step time %% Run Simulink Model %% sim('EVAP_Sim'); % EVAP_Driver.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% EVAP_Dynamics.m % EVAP_Dynamics.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% function EVAP_Dynamics(block) % Dynamic modeling of an evaporator % % Written by ChoonJae Ryu, Energy and Gas Dynamics Lab (Prof. Lear). % Department of Mechanical and Aerospace Engineering, University of Florida %% %% The setup method is used to setup the basic attributes of the %% Sfunction such as ports, parameters, etc. Do not add any other %% calls to the main body of the function. %% setup(block); %endfunction %% Function: setup =================================================== %% Abstract: %% Set up the Sfunction block's basic characteristics such as: %% Input ports %% Output ports %% Dialog parameters %% Options %% %% Required : Yes %% CMex counterpart: mdlInitializeSizes %% function setup(block) % Register parameters block.NumDialogPrms = 12; % Register number of ports block.NumInputPorts = 1; block.NumOutputPorts = 1; % Setup port properties to be inherited or dynamic

PAGE 221

221 block.SetPreCompInpPortInfoToDynamic; block.SetPreCompOutPortInfoToDynamic; % Override input port properties block.InputPort(1).Dimensions = 7; % size_inputs %block.InputPort(1).DatatypeID = 0; % double %block.InputPort(1).Complexity = 'Real'; block.InputPort(1).DirectFeedthrough = 0; % size_feedthrough % Override output port properties block.OutputPort(1).Dimensions = 5; % size_outputs %block.OutputPort(1).DatatypeID = 0; % double %block.OutputPort(1).Complexity = 'Real'; % Register sample times % [0 offset] : Continuous sample time % [positive_num offset] : Discrete sample time % % [1, 0] : Inherited sample time % [2, 0] : Variable sample time block.SampleTimes = [0 0]; % States block.NumContStates = 3; % xe = [le1 Pe heo Tew1 Tew2]' % Specify the block simStateComliance. The allowed values are: % 'UnknownSimState', < The default setting; warn and assume DefaultSimState % 'DefaultSimState', < Same sim state as a builtin block % 'HasNoSimState', < No sim state % 'CustomSimState', < Has GetSimState and SetSimState methods % 'DisallowSimState' < Error out when saving or restoring the model sim state block.SimStateCompliance = 'DefaultSimState'; %% ----------------------------------------------------------------%% The Mfile Sfunction uses an internal registry for all %% block methods. You should register all relevant methods %% (optional and required) as illustrated below. You may choose %% any suitable name for the methods and implement these methods %% as local functions within the same file. See comments %% provided for each function for more information. %% ----------------------------------------------------------------%block.RegBlockMethod('PostPropagationSetup', @DoPostPropSetup); block.RegBlockMethod('InitializeConditions', @InitializeConditions); %block.RegBlockMethod('Start', @Start); block.RegBlockMethod('Outputs', @Outputs); % Required %block.RegBlockMethod('Update', @Update); block.RegBlockMethod('Derivatives', @Derivatives); %block.RegBlockMethod('Terminate', @Terminate); % Required %end setup %%

PAGE 222

222 %% PostPropagationSetup: %% Functionality : Setup work areas and state variables. Can %% also register runtime methods here %% Required : No %% CMex counterpart: mdlSetWorkWidths %% %function DoPostPropSetup(block) %block.NumDworks = 1; % % block.Dwork(1).Name = 'x1'; % block.Dwork(1).Dimensions = 1; % block.Dwork(1).DatatypeID = 0; % double % block.Dwork(1).Complexity = 'Real'; % real % block.Dwork(1).UsedAsDiscState = true; % % block.Dwork(2).Name = 'x2'; % block.Dwork(2).Dimensions = 1; % block.Dwork(2).DatatypeID = 0; % double % block.Dwork(2).Complexity = 'Real'; % real % block.Dwork(2).UsedAsDiscState = true; % %% %% InitializeConditions: %% Functionality : Called at the start of simulation and if it is %% present in an enabled subsystem configured to reset %% states, it will be called when the enabled subsystem %% restarts execution to reset the states. %% Required : No %% CMEX counterpart: mdlInitializeConditions %% function InitializeConditions(block) Les_0 = block.DialogPrm(1).Data; xe_0 = block.DialogPrm(2).Data; Tea_0 = block.DialogPrm(3).Data; block.ContStates.Data = [Les_0; xe_0; Tea_0]; %end InitializeConditions %% %% Start: %% Functionality : Called once at start of model execution. If you %% have states that should be initialized once, this %% is the place to do it. %% Required : No %% CMEX counterpart: mdlStart %% %function Start(block) %endfunction %% %% Outputs:

PAGE 223

223 %% Functionality : Called to generate block outputs in %% simulation step %% Required : Yes %% CMEX counterpart: mdlOutputs %% function Outputs(block) % parameters or constants Le = block.DialogPrm(4).Data; % m Ae = block.DialogPrm(5).Data; % m2 UAe = block.DialogPrm(6).Data; % kW/mK % inputs Pe = block.InputPort(1).Data(1); if Pe == 0 Pe = block.DialogPrm(8).Data; dPedt = 0; xei = block.DialogPrm(9).Data; hei = block.DialogPrm(10).Data; mdei = block.DialogPrm(11).Data; Teai = block.DialogPrm(12).Data; else dPedt = block.InputPort(1).Data(2); xei = block.InputPort(1).Data(3); hei = block.InputPort(1).Data(4); mdei = block.InputPort(1).Data(5); Teai = block.InputPort(1).Data(7); end % state variables, xc = []' Les = block.ContStates.Data(1); xe = block.ContStates.Data(2); if xe>1 xe=1; end Tea = block.ContStates.Data(3); % Thermodyanmic properties xeo = xe; [Ter1 heo rhoe1 rhoe1he1 drhoe1dPe drhoe1he1dPe drhoe1dxe drhoe1he1dxe ... rhoe2 rhoe2he2 drhoe2dPe drhoe2he2dPe drhoe2dxe drhoe2he2dxe] ... = EVAP_2ph(Pe,xe,hei,Le,Les); Teao = 2*Tea Teai; e12 = Ae*(Les*rhoe1 + (Le Les)*rhoe2); e21 = Ae*(rhoe1he1 rhoe1*heo + rhoe2*heo rhoe2he2); e22 = Ae*Les*(drhoe1he1dxe heo*drhoe1dxe) ... + Ae*(Le Les)*(drhoe2he2dxe heo*drhoe2dxe); fe1 = mdei*(xei xeo); fe2 = mdei*(hei heo) + UAe*(Tea Ter1) ... (Ae*Les*(drhoe1he1dPe heo*drhoe1dPe 1) ... + Ae*(Le Les)*(drhoe2he2dPe heo*drhoe2dPe 1))*dPedt; xde = fe1/e12; Ldes = (e12*fe2 e22*fe1)/e12/e21;

PAGE 224

224 mdeo = mdei Ae*(rhoe1 rhoe2)*Ldes ... (Ae*Les*drhoe1dxe + Ae*(Le Les)*drhoe2dxe)*xde ... (Ae*Les*drhoe1dPe + Ae*(Le Les)*drhoe2dPe)*dPedt; block.OutputPort(1).Data = [mdeo; xeo; heo; Teao; Les]; %block.OutputPort(1).Data = block.Dwork(1).Data + block.InputPort(1).Data; %end Outputs %% %% Update: %% Functionality : Called to update discrete states %% during simulation step %% Required : No %% CMEX counterpart: mdlUpdate %% %function Update(block) %block.Dwork(1).Data = block.InputPort(1).Data; %end Update %% %% Derivatives: %% Functionality : Called to update derivatives of %% continuous states during simulation step %% Required : No %% CMEX counterpart: mdlDerivatives %% function Derivatives(block) % parameters or constants Le = block.DialogPrm(4).Data; % m Ae = block.DialogPrm(5).Data; % m2 UAe = block.DialogPrm(6).Data; % kW/mK rhoea = 3; % kg/m3 Cpea = block.DialogPrm(7).Data; % kJ/kgK Aea = Ae; % m2 % inputs Pe = block.InputPort(1).Data(1); dPedt = block.InputPort(1).Data(2); xei = block.InputPort(1).Data(3); hei = block.InputPort(1).Data(4); mdei = block.InputPort(1).Data(5); mdea = block.InputPort(1).Data(6); Teai = block.InputPort(1).Data(7); % state variables, xc = []' Les = block.ContStates.Data(1); xe = block.ContStates.Data(2); if xe>1 xe=1; end Tea = block.ContStates.Data(3); % Thermodyanmic properties xeo = xe; [Ter1 heo rhoe1 rhoe1he1 drhoe1dPe drhoe1he1dPe drhoe1dxe drhoe1he1dxe ...

PAGE 225

225 rhoe2 rhoe2he2 drhoe2dPe drhoe2he2dPe drhoe2dxe drhoe2he2dxe] ... = EVAP_2ph(Pe,xe,hei,Le,Les); Teao = 2*Tea Teai; e12 = Ae*(Les*rhoe1 + (Le Les)*rhoe2); e21 = Ae*(rhoe1he1 rhoe1*heo + rhoe2*heo rhoe2he2); e22 = Ae*Les*(drhoe1he1dxe heo*drhoe1dxe) ... + Ae*(Le Les)*(drhoe2he2dxe heo*drhoe2dxe); e33 = rhoea*Aea*Cpea*Le; fe1 = mdei*(xei xeo); fe2 = mdei*(hei heo) + UAe*(Tea Ter1) ... (Ae*Les*(drhoe1he1dPe heo*drhoe1dPe 1) ... + Ae*(Le Les)*(drhoe2he2dPe heo*drhoe2dPe 1))*dPedt; fe3 = mdea*Cpea*(Teai Teao) + UAe*(Ter1 Tea); xde = fe1/e12; Tdea = fe3/e33; Ldes = (e12*fe2 e 22*fe1)/e12/e21; block.Derivatives.Data = [Ldes; xde; Tdea]; %end Derivatives %% %% Terminate: %% Functionality : Called at the end of simulation for cleanup %% Required : Yes %% CMEX counterpart: mdlTerminate %% %function Terminate(block) %end Terminate % End of EVAP_Dynamics.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% EVAP_2ph.m % EVAP_2ph.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% function [Tmean ho r1 rh1 dr1dP drh1dP dr1dx drh1dx ... r2 rh2 dr2dP drh2dP dr2dx drh2dx] = EVAP_2ph(P,xe1,hei,Le,Les) load EVAP_pxz2.mat m = 100; dP = 1e3; xtol = 1e6; P1 = P dP/2; P2 = P + dP/2;

PAGE 226

226 if xe1 > 1xtol x1 = xe1 xtol/2; x2 = 1; elseif xe1 < xtol x1 = 0; x2 = xe1 + xtol/2; else x1 = xe1 xtol/2; x2 = xe1 + xtol/2; end dx = x2 x1; ze = 0:1/n:1; [xxe ppe zze] = meshgrid(xe,Pe,ze); z = [0 1]; [xi pii zi] = meshgrid(xe1,P,z); he1z = interp3(xxe,ppe,zze,he,xi,pii,zi); zei = (hei he1z(1))/(he1z(2) he1z(1)); zeo = zei + Le/Les*(1 zei); [xhi phi zhi] = meshgrid(xe1,P,zeo); ho = interp3(xxe,ppe,zze,he,xhi,phi,zhi); xx = [x1 xe1 x2]; pp = [P1 P P2]; z1 = zei:(1zei)/m:1; z2 = zeo:(1zeo)/m:1; zz = zei:(zeozei)/m:zeo; % For temperature only [xi1 pi1 zi1] = meshgrid(xx,pp,z1); rho1 = interp3(xxe,ppe,zze,rhoe,xi1,pi1,zi1); r1 = mean(rho1(2,2,:)); rp1 = mean(rho1(1,2,:)); rp2 = mean(rho1(3,2,:)); rx1 = mean(rho1(2,1,:)); rx2 = mean(rho1(2,3,:)); rho1h1 = interp3(xxe,ppe,zze,rhoehe,xi1,pi1,zi1); rh1 = mean(rho1h1(2,2,:)); rhp1 = mean(rho1h1(1,2,:)); rhp2 = mean(rho1h1(3,2,:)); rhx1 = mean(rho1h1(2,1,:)); rhx2 = mean(rho1h1(2,3,:)); dr1dP = (rp2 rp1)/dP; drh1dP = (rhp2 rhp1)/dP; dr1dx = (rx2 rx1)/dx; drh1dx = (rhx2 rhx1)/dx; [xi2 pi2 zi2] = meshgrid(xx,pp,z2); rho2 = interp3(xxe,ppe,zze,rhoe,xi2,pi2,zi2); r2 = mean(rho2(2,2,:)); rp1 = mean(rho2(1,2,:)); rp2 = mean(rho2(3,2,:)); rx1 = mean(rho2(2,1,:)); rx2 = mean(rho2(2,3,:)); rho2h2 = interp3(xxe,ppe,zze,rhoehe,xi2,pi2,zi2); rh2 = mean(rho2h2(2,2,:)); rhp1 = mean(rho2h2(1,2,:));

PAGE 227

227 rhp2 = mean(rho2h2(3,2,:)); rhx1 = mean(rho2h2(2,1,:)); rhx2 = mean(rho2h2(2,3,:)); dr2dP = (rp2 rp1)/dP; drh2dP = (rhp2 rhp1)/dP; dr2dx = (rx2 rx1)/dx; drh2dx = (rhx2 rhx1)/dx; [xxi ppi zzi] = meshgrid(xe1,P,zz); Te1 = interp3(xxe,ppe,zze,Te,xxi,ppi,zzi); Tmean = mean(Te1(1,1,:)); % End of EVAP_2ph.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Solution Heat Exchanger SHX_Driver.m % SHX_Driver.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% clear all %% initial values %% % initial inputs % Ph = 1.791850466000000e+003; hsci = 99.908784049784416; xsci = 0.62; hshi = 1.777520026131992e+002; xshi = 0.468975972268840; mdsci = 0.185995325421376; mdsho = 0.154747876421376; % Geometry and parameters % Ls = 2.032; Asc = 0.002696964962284; UAs = 1.795847738116563; % Initial values of state variables % Tsc = 3.206828516611801e+002; xsc = 0.62; Tsh = 3.403698485178869e+002; xsh = 0.46897597226884; %% Step inputs %% stime = 0; % Pressure Phf = Ph;%*1.01; Phslope = 0;%0.5; % step input for dPadt Phst = stime; % sec. start time % Ammonia mass fraction and enthalpy xscif = xsci;%*1.1; % final value xscist = stime; % step time % Enthalpy

PAGE 228

228 hscif = hsci;%*0.9; hscist = stime; % Ammonia mass fraction and enthalpy xshif = xshi;%*1.1; % final value xshist = stime; % step time % Enthalpy hshif = hshi;%*1.1; hshist = stime; % Mass flowrate refrigerant mdscif = mdsci*1.1; % final value mdscist = stime; % step time % Mass flow rate secondary fluid mdshof = mdsho;%*1.1; % final value mdshost = stime; % step time %% Run Simulink Model %% sim('SHX_Sim'); % End of SHX_Driver.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% SHX_Driver.m % SHX_Dynamics.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% function SHX_Dynamics(block) % Dynamic modeling of an SHX % This file works with SHX_sfunc2.mdl and SHX_ts_init1.m % % This SFunction is utilized in the SIMULINK file named SHX_sfunc2.mdl % Written by ChoonJae Ryu, Energy and Gas Dynamics Lab (Prof. Lear). % Department of Mechanical and Aerospace Engineering, University of Florida %% %% The setup method is used to setup the basic attributes of the %% Sfunction such as ports, parameters, etc. Do not add any other %% calls to the main body of the function. %% setup(block); %endfunction %% Function: setup =================================================== %% Abstract: %% Set up the Sfunction block's basic characteristics such as: %% Input ports %% Output ports %% Dialog parameters %% Options %% %% Required : Yes %% CMex counterpart: mdlInitializeSizes %%

PAGE 229

229 function setup(block) % Register parameters block.NumDialogPrms = 14; % Register number of ports block.NumInputPorts = 1; block.NumOutputPorts = 1; % Setup port properties to be inherited or dynamic block.SetPreCompInpPortInfoToDynamic; block.SetPreCompOutPortInfoToDynamic; % Override input port properties block.InputPort(1).Dimensions = 8; % size_inputs %block.InputPort(1).DatatypeID = 0; % double %block.InputPort(1).Complexity = 'Real'; block.InputPort(1).DirectFeedthrough = 0; % size_feedthrough % Override output port properties block.OutputPort(1).Dimensions = 6; % size_outputs %block.OutputPort(1).DatatypeID = 0; % double %block.OutputPort(1).Complexity = 'Real'; % Register sample times % [0 offset] : Continuous sample time % [positive_num offset] : Discrete sample time % % [1, 0] : Inherited sample time % [2, 0] : Variable sample time block.SampleTimes = [0 0]; % States block.NumContStates = 4; % xe = [le1 Pe heo Tew1 Tew2]' % Specify the block simStateComliance. The allowed values are: % 'UnknownSimState', < The default setting; warn and assume DefaultSimState % 'DefaultSimState', < Same sim state as a builtin block % 'HasNoSimState', < No sim state % 'CustomSimState', < Has GetSimState and SetSimState methods % 'DisallowSimState' < Error out when saving or restoring the model sim state block.SimStateCompliance = 'DefaultSimState'; %% ----------------------------------------------------------------%% The Mfile Sfunction uses an internal registry for all %% block methods. You should register all relevant methods %% (optional and required) as illustrated below. You may choose %% any suitable name for the methods and implement these methods %% as local functions within the same file. See comments %% provided for each function for more information. %% ----------------------------------------------------------------%block.RegBlockMethod('PostPropagationSetup', @DoPostPropSetup);

PAGE 230

230 block.RegBlockMethod('InitializeConditions', @InitializeConditions); %block.RegBlockMethod('Start', @Start); block.RegBlockMethod('Outputs', @Outputs); % Required %block.RegBlockMethod('Update', @Update); block.RegBlockMethod('Derivatives', @Derivatives); %block.RegBlockMethod('Terminate', @Terminate); % Required %end setup %% %% PostPropagationSetup: %% Functionality : Setup work areas and state variables. Can %% also register runtime methods here %% Required : No %% CMex counterpart: mdlSetWorkWidths %% %function DoPostPropSetup(block) %block.NumDworks = 1; % % block.Dwork(1).Name = 'x1'; % block.Dwork(1).Dimensions = 1; % block.Dwork(1).DatatypeID = 0; % double % block.Dwork(1).Complexity = 'Real'; % real % block.Dwork(1).UsedAsDiscState = true; % % block.Dwork(2).Name = 'x2'; % block.Dwork(2).Dimensions = 1; % block.Dwork(2).DatatypeID = 0; % double % block.Dwork(2).Complexity = 'Real'; % real % block.Dwork(2).UsedAsDiscState = true; % %% %% InitializeConditions: %% Functionality : Called at the start of simulation and if it is %% present in an enabled subsystem configured to reset %% states, it will be called when the enabled subsystem %% restarts execution to reset the states. %% Required : No %% CMEX counterpart: mdlInitializeConditions %% function InitializeConditions(block) hsc_0 = block.DialogPrm(1).Data; xsc_0 = block.DialogPrm(2).Data; hsh_0 = block.DialogPrm(3).Data; xsh_0 = block.DialogPrm(4).Data; block.ContStates.Data = [hsc_0; xsc_0; hsh_0; xsh_0]; %end InitializeConditions %% %% Start: %% Functionality : Called once at start of model execution. If you

PAGE 231

231 %% have states that should be initialized once, this %% is the place to do it. %% Required : No %% CMEX counterpart: mdlStart %% %function Start(block) %endfunction %% %% Outputs: %% Functionality : Called to generate block outputs in %% simulation step %% Required : Yes %% CMEX counterpart: mdlOutputs %% function Outputs(block) Ls = block.DialogPrm(5).Data; % m Asc = block.DialogPrm(6).Data; % m2 UAs = block.DialogPrm(7).Data; % kW/mK Ash = Asc; % m2 Ph = block.InputPort(1).Data(1); if Ph == 0 Ph = block.DialogPrm(8).Data; dPhdt = 0; hsci = block.DialogPrm(9).Data; xsci = block.DialogPrm(10).Data; hshi = block.DialogPrm(11).Data; xshi = block.DialogPrm(12).Data; m dsci = block.DialogPrm(13).Data; mdsho = block.DialogPrm(14).Data; else dPhdt = block.InputPort(1).Data(2); hsci = block.InputPort(1).Data(3); xsci = block.InputPort(1).Data(4); hshi = block.InputPort(1).Data(5); xshi = block.InputPort(1).Data(6); mdsci = block.InputPort(1).Data(7); mdsho = block.InputPort(1).Data(8); end Tsc = block.ContStates.Data(1); xsc = block.ContStates.Data(2); if xsc>1 xsc=1; end Tsh = block.ContStates.Data(3); xsh = block.ContStates.Data(4); if xsh>1 xsh=1; end hsc = liqprop_ibrahim(Tsc,Ph,xsc); hsh = liqprop_ibrahim(Tsh,Ph,xsh); xsco = xsc; if xsco>1 xsco=1; end xsho = xsh; if xsho>1 xsho=1; end hsco = 2*hsc hsci; hsho = 2*hsh hshi;

PAGE 232

232 [rhosc hsc drhoscdPh dhscdPh drhoscdxsc dhscdxsc ... drhoscdTsc dhscdTsc] = liqp(Tsc,Ph,xsc); [rhosh hsh drhoshdPh dhshdPh drhoshdxsh dhshdxsh ... drhoshdTsh dhshdTsh] = liqp(Tsh,Ph,xsh); s11 = Asc*Ls*(xsc xsco)*drhoscdTsc; s12 = Asc*Ls*((xsc xsco)*drhoscdxsc + rhosc); s1p = Asc*Ls*(xsc xsco)*drhoscdPh; s21 = Asc*Ls*((hsc hsco)*drhoscdTsc + rhosc*dhscdTsc); s22 = Asc*Ls*((hsc hsco)*drhoscdxsc + rhosc*dhscdxsc); s2p = Asc*Ls*((hsc hsco)*drhoscdPh + rhosc*dhscdPh 1); s33 = Ash*Ls*(xsh xshi)*drhoshdTsh; s34 = Ash*Ls*((xsh xshi)*drhoshdxsh + rhosh); s3p = Ash*Ls*(xsh xshi)*drhoshdPh; s43 = Ash*Ls*((hsh hshi)*drhoshdTsh + rhosh*dhshdTsh); s44 = Ash*Ls*((hsh hshi)*drhoshdxsh + rhosh*dhshdxsh); s4p = Ash*Ls*((hsh hshi)*drhoshdPh + rhosh*dhshdPh 1); fs1 = mdsci*(xsci xsco) s1p*dPhdt; fs2 = mdsci*(hsci hsco) + UAs*(Tsh Tsc) s2p*dPhdt; fs3 = mdsho*(xshi xsho) s3p*dPhdt; fs4 = mdsho*(hshi hsho) + UAs*(Tsc Tsh) s4p*dPhdt; detSC = s11*s22 s12*s21; Tdsc = (s22*fs1 s12*fs2)/detSC; xdsc = (s11*fs2 s21*fs1)/detSC; detSH = s33*s44 s34*s43; Tdsh = (s44*fs3 s34*fs4)/detSH; xdsh = (s33*fs4 s43*fs3)/detSH; mdsco = mdsci Asc*Ls*(drhoscdPh*dPhdt + drhoscdTsc*Tdsc ... + drhoscdxsc*xdsc); mdshi = mdsho + Ash*Ls*(drhoshdPh*dPhdt + drhoshdTsh*Tdsh ... + drhoshdxsh*xdsh); block.OutputPort(1).Data = [mdsco; xsco; hsco; mdshi; xsho; hsho]; %block.OutputPort(1).Data = block.Dwork(1).Data + block.InputPort(1).Data; %end Outputs %% %% Update: %% Functionality : Called to update discrete states %% during simulation step %% Required : No %% CMEX counterpart: mdlUpdate %% %function Update(block) %block.Dwork(1).Data = block.InputPort(1).Data;

PAGE 233

233 %end Update %% %% Derivatives: %% Functionality : Called to update derivatives of %% continuous states during simulation step %% Required : No %% CMEX counterpart: mdlDerivatives %% function Derivatives(block) % parameters or constants Ls = block.DialogPrm(5).Data; % m Asc = block.DialogPrm(6).Data; % m2 UAs = block.DialogPrm(7).Data; % kW/mK Ash = Asc; % m2 % inputs [Pc dPcdt xci mdci mdca Tcai] Ph = block.InputPort(1).Data(1); dPhdt = block.InputPort(1).Data(2); hsci = block.InputPort(1).Data(3); xsci = block.InputPort(1).Data(4); hshi = block.InputPort(1).Data(5); xshi = block.InputPort(1).Data(6); mdsci = block.InputPort(1).Data(7); mdsho = block.InputPort(1).Data(8); % state variables, xc = []' Tsc = block.ContStates.Data(1); xsc = block.ContStates.Data(2); if xsc>1 xsc=1; end Tsh = block.ContStates.Data(3); xsh = block.ContStates.Data(4); if xsh>1 xsh=1; end hsc = liqprop_ibrahim(Tsc,Ph,xsc); hsh = liqprop_ibrahim(Tsh,Ph,xsh); xsco = xsc; if xsco>1 xsco=1; end xsho = xsh; if xsho>1 xsho=1; end hsco = 2*hsc hsci; hsho = 2*hsh hshi; [rhosc hsc drhoscdPh dhscdPh drhoscdxsc dhscdxsc ... drhoscdTsc dhscdTsc] = liqp(Tsc,Ph,xsc); [rhosh hsh drhoshdPh dhshdPh drhoshdxsh dhshdxsh ... drhoshdTsh dhshdTsh] = liqp(Tsh,Ph,xsh); s11 = Asc*Ls*(xsc xsco)*drhoscdTsc; s12 = Asc*Ls*((xsc xsco)*drhoscdxsc + rhosc); s1p = Asc*Ls*(xsc xsco)*drhoscdPh; s21 = Asc*Ls*((hsc hsco)*drhoscdTsc + rhosc*dhscdTsc); s22 = Asc*Ls*((hsc hsco)*drhoscdxsc + rhosc*dhscdxsc); s2p = Asc*Ls*((hsc hsco)*drhoscdPh + rhosc*dhscdPh 1); s33 = Ash*Ls*(xsh xshi)*drhoshdTsh;

PAGE 234

234 s34 = Ash*Ls*((xsh xshi)*drhoshdxsh + rhosh); s3p = Ash*Ls*(xsh xshi)*drhoshdPh; s43 = Ash*Ls*((hsh hshi)*drhoshdTsh + rhosh*dhshdTsh); s44 = Ash*Ls*((hsh hshi)*drhoshdxsh + rhosh*dhshdxsh); s4p = Ash*Ls*((hsh hshi)*drhoshdPh + rhosh*dhshdPh 1); fs1 = mdsci*(xsci xsco) s1p*dPhdt; fs2 = mdsci*(hsci hsco) + UAs*(Tsh Tsc) s2p*dPhdt; fs3 = mdsho*(xshi xsho) s3p*dPhdt; fs4 = mdsho*(hshi hsho) + UAs*(Tsc Tsh) s4p*dPhdt; detSC = s11*s22 s12*s21; Tdsc = (s22*fs1 s12*fs2)/detSC; xdsc = (s11*fs2 s21*fs1)/detSC; detSH = s33*s44 s34*s43; Tdsh = (s44*fs3 s34*fs4)/detSH; xdsh = (s33*fs4 s43*fs3)/detSH; block.Derivatives.Data = [Tdsc; xdsc; Tdsh; xdsh]; %end Derivatives %% %% Terminate: %% Functionality : Called at the end of simulation for cleanup %% Required : Yes %% CMEX counterpart: mdlTerminate %% %function Terminate(block) %end Terminate % End of SHX_Dynamics.m %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

PAGE 235

235 LIST OF REFERENCES [ 1 ] Anxionnaz, R., 1945 Inst allation a T urbi nes a G az a C ircuit S emi O uvert French Patent 999133. [ 2 ] Anxionnaz, R., 1948 Improvements in or R elating to G as T urbine P lant with S emi O pen C ircuit British Patent 651166. [ 3 ] Lear, W E. and Sherif S. A., Combined C ooling and P ower P lant with W ater E xtraction, US Patent 20060037337. [ 4 ] Kalina, A. I., 1983, Combined C ycle and W aste H eat R ecovery P ower S ystems B ased on a N ovel T hermodynamic E nergy C ycle U tilizing L ow T emper ature H eat for P ower G eneration ASME P aper N o. 83 JPG C GT 3 [ 5 ] Ptek J. and Klomfar, J., 1995, Simple F unctions for F ast C alculations of S elected T hermodynamic P roperties of the A mmoniaW ater S ystem, Int. J Refrig ., 18 ( 4 ) pp 228 234. [ 6 ] Tillner Roth R., and Friend, D. G., 1998, Survey and A ssessment of A vailable M easurements on T hermodynamic P roperties of the M i xture { W ater + A mmonia}, J. P hys C hem R ef D at a, 27( 1 ), pp. 4561 [ 7 ] Gillespi, P. C. Wilding, W. V. and Wilson, G. M. 1985, Vapor L iquid E quilibrium M easurements on the A mmoniaW ater S ystem from 313K to 589K, Project 75881, Itec Research Co. Inc. Provo, Utah [ 8 ] Macriss, R. A., Eakin, B. E. Ellington, R. T. and Huebler, J. 1964, Physical and T hermodynamic P roperties of A mmonia W ater M ixtures, Research bulletin, No. 34, Ins titute of G as T echnology, Chicago, I L. [9 ] Park, Y M. and Sonntag, R. E. 1990, Thermodynamic P roperties of A mmoniaW ater M ixtures: a G eneralized E quationof S tate A pproach, ASHRAE Tran s ., 9 6 pp. 150159. [1 0 ] Tillner Roth R., and Friend, D. G., 1998, A Helmholtz F ree E nergy F ormulation of the T hermodynamic P roperties of the M ixture { W ater + A mmonia}, J. Phys. Chem. Ref. Data, 27(1) pp 6396 [1 1 ] El Sayed, Y. M. and Tribus, M., 1985, Thermodynamic P roperties of W ater A mmonia M ixtures T heoretic al I mplementation for U se in P ower C ycles A nalysis, ASME Pub. AES 1 pp. 89 95. [1 2 ] Bogart, M., 1981, Ammonia Absorption Refrigeration in Industrial Processes Gulf Pub Co., Houston, TX. [1 3 ] Xu, F., and Goswami, D. Y. 1999, Thermodynamic P roperties o f A mmoniaW ater M ixtures for P ower C ycle A pplications, Energy, 24(6), pp 52 5 536

PAGE 236

236 [1 4 ] Tamm, G. O., 2003, Experimental I nvestigation of A n A mmoniaB ased C ombined P ower and C ooling C ycle, Ph.D. Dissertat ion, University of Florida, Gainesville, FL. [1 5 ] Ibrahim O. M., and Klein, S. A., 1993, Thermodynamic P roperties of A mmoniaW ater M ixtures, ASHRAE T ran s., 99 pp. 1495 1502. [1 6 ] Ryu C., Tiffany, D. R. Crittenden, J. F. Lear, W. E., and Sherif, S. A. 2010, Dynamic M odeling of A N ovel C ooling, H eat, P ower, and W ater M icroturbine C ombined C ycle, J. E nerg R esour ASME 132(2), pp. 0210061 9. [1 7 ] Khan, J. 2006, Design and O ptimization of A D istributed G eneration S ystem with the P roduction of W ater and R efrigeration, Ph.D. Dissertation, Universi ty of Florida, Gainesville, FL. [1 8 ] Adewusi, S. A. and Zubair, S M. 2004, Second L aw B ased T hermodynamic A nalysis of A mmoniaW ater A bsorption S ystems, Energ C onvers M anage. 45( 1516) pp. 23552369 [ 19] Kim, B ., and Park, J 2007, Dynamic S imula tion of A S ingle E ffect A mmoniaW ater A bsorption C hiller, Int. J Refrig 30 (3) pp. 535 545 [2 0 ] MacArthur, J. W. and Grald, E. W., 1989, Unsteady C ompressible T wo P hase F low M odel for P redicting C yclic H eat P ump P erformance and A C omparison with E xpe rimental D ata, Int. J. Refrig., 12 (1) pp. 2941 [2 1 ] He, X. D., Liu, S., and Asada, H. H., 1997, Modeling of Vapor Compression Cycles for Multivariable Feedback Control of HVAC Systems, J. Dyn. Syst.T ASME, 119 (2), pp. 183 191. [2 2 ] Jensen, J. M., an d Tummescheit, H., 2002, Moving Boundary Models of Dynamic Simulations of TwoPhase Flows, Proceedings Int. Modelica Conf. 2 pp. 235244. [2 3 ] Wedekind, G. L., Bhatt, B. L., and Beck, B. T., 1978, A System Mean Void Fraction Model for Predicting Various Transient Phenomena Associated with Two Phase Evaporating and Condensing Flows, Int. J. Multiphas. Flow, 4 (1), pp. 97114. [2 4 ] Zivi, S. M., 1964, Estimation of Steady State Steam Void Fraction by Means of the Principle of Minimum Entropy Production, J. Heat Transfer, 86 pp. 247 252. [2 5 ] Schulz, S. C. G. 1971 Equations of S tate for the S ystem A mmonia W ater for U se with C omputers, Proceedings Int. Congr Refrig 2 pp. 431 436. [2 6 ] Ziegler B., and Trepp, Ch., 1984, Equation of S tate for A mmonia W ater M ix tures, Int. J. Refrig 7 ( 2 ), pp. 101 106

PAGE 237

237 [2 7 ] M. Conde Eng., 2006, Thermophysical Properties of {NH3 + H2O} Mixtures for the Industrial Design of Absorption Refrigeration Equipment Manuel Conde Engin eering, Switzerland Zurich. [2 8 ] Lear, W. E. and Laganelli, A. L., 1999, High Pressure Regenerative Turbine Engine: 21st Century Propulsion, Final Test Report Contract No. NAS327396.14 [ 29] Khan, J. R., Lear, W. E., Sherif, S. A., and Crittenden, J. F., 2008, "Performance of a Novel Combined Cooling and Power Gas Turbine with Water Harvesting," J. Eng. Gas. Turb. Power, 130 (4), pp. 0417021 10 [3 0 ] Goswami, D. Y., Kreith, F., and Kreider, J. F., 2000, Principles of Solar Engineering, 2nd ed. Taylo r and Francis, Philadelphia, PA [3 1 ] Parsons R. A., 1998, ASHRAE Handbook Refrigeration SI ed., ASHRAE, Atlanta, GA [3 2 ] Gustafson, M. W., Baylor, J. S., and Epstein, G., 1993, "Estimating Air Conditioning Load Control Effectiveness Using An Engineering Model," IEEE T Power Syst. 8 ( 3 ) pp. 972978. [3 3 ] Threlkeld, J. L., 1962, Thermal Environmental Engineering, Prentice Hall, Inc., Englewood Cliffs, NJ [3 4 ] MacFarlane, R. S., and Lear, W. E., 1997, System Impact of H2O Production and Injection on a Novel Semi Closed Cycle Gas Turbine, Paper No. AIAA970374. [3 5 ] Mattingly, J. D., 1996, Elements of Gas Turbine Propulsion McGraw Hill, Inc., New York. [3 6 ] Carcasci, C., and Facchini, B., 2000, Comparison b etween Two Gas Turbine Solutions to Increase Combined Power Plant Efficiency, Energ. Co nvers. Manage. 41 (8), pp. 757 773. [3 7 ] Fiaschi, D., Lombardi, L., and Tapinassi, L., 2004, The Recuperative Auto Thermal Reforming and Recuperative Reforming Gas Turbine Power Cycles with CO2 Removal Part II: The Recuperative Reforming Cycle, J. Eng. Gas. Turb. Power, 126(1), pp. 6268. [3 8 ] Boza, J. J., Lear, W. E., and Sherif, S. A., 2003, Performance of a Novel Semi Closed Gas Turbine Refrigeration Combined Cycle, ASME Paper No. 2003GT 38576. [ 39] Proeschel, R., 2002, http://www.proepowersystems.com/proe90.htm

PAGE 238

238 [4 0 ] M assardo, F., Cazzola, W., and Lagorio, G., 2004, Widget Temp: A Novel Web Based Approach for the Thermoeconomic Analysis and Optimization of Conventional and Innovative Cycles ASME Paper No. 2004GT 54115 [4 1 ] Lear, W. E., Ryu, C. J., Crittenden, J. F., Srinivasan, A., Ellis, W., Tiffany, D. R., and Sherif, S. A., 2008, System Design of a Novel Combined Cooling, Heat, Power, and Water Microturbine Combined Cycle, ASME Paper No. GT200851454 [42] Huang, M. C., Chen, B. R., Hsiao, M. J., and Chen, S. L., 2007, Application of Thermal Battery in the Ice Storage Air Conditioning System as a Subcooler, Int. J. Refrig. 30(2), pp. 245253 [43] MacArthur, J. W. and Grald, E. W., 1989, Unsteady Compressible TwoPhase Flow Model for Predicting Cyclic Heat Pump Performance and a Comparison with E xperimental Data, Int J. Refrig. 12 ( 1 ) pp. 2941. [4 4 ] Seborg, D. E., Edgar, T. F., and Mellichamp, D. A., 2004, Process Dynamics and Control 2nd ed. Wiley New York

PAGE 239

239 BIOGRAPHICAL SKETCH Choon Jae Ryu was born in Busan, South Korea, in 1977. He graduated from G yeongsang University auxiliary high s chool in Jinju, Gyeongnam Province, South Korea, in 1995. In 2001, he received a Bachelor of Engineering in automotive e ngineering from Korea Maritime University with a S umma Cum Laude honor in engineering He joined Dr. Rodney S. Rouff s research group in 2002 as a predoctoral fellow and studied about carbon nano tube in Northwestern University in Evanston, IL. After finishing research in Northwestern University, he joined Dr. SungTak Ro s research group in 2003. His research area was about mild hybrid electric vehicle. In 2005, he r eceived a Master of Science in mechanical and aerospace engineering from Seoul National University. He worked for Siemens Automotive Company in Icheon, Gyeonggi Province, South Korea and Next Generation Vehicle (NGV), a division of Hyundai Motor Company, in Seoul, South Korea after graduation. He started his Ph.D studies at the University of Florida in the D epartment of M echanical and A erospac e E ngineering an d joined Dr. William E. Lear s group in 2007. He obtained his Doctor of Philos ophy in mechanical engineering from University of Florida in 2011.