<%BANNER%>

Experimental Investigation of An Ammonia-Based Combined Power and Cooling Cycle


PAGE 1

EXPERIMENTAL INVESTI GATION OF AN AMMONIA BASED COMBINED POWER AND COOLING CYCLE By GUNNAR OLAVI TAMM A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR T HE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2003

PAGE 2

ii ACKNOWLEDGMENTS There are a number of individuals who have contributed to the completion of this dissertation and the enrichment of my educational experience at the University of Florida. The gracious assistance, guidance and friendship of these individua ls are recognized. As my Ph.D. committee chair, Dr. D. Yogi Goswami provided guidance but allowed me to follow my own research design. This afforded me the opportunity and responsibility to think independently and motivate myself on a real project. Dr. Gos wami also encouraged interaction with industry, writing journal papers, soliciting grants and attending conferences, which have exposed me to the professional stage and built confidence in my transition towards an engineering career. My Ph.D. committee mem bers, Dr. Jacob N. Chung, Dr. James F. Klausner, Dr. Ulrich H. Kurtzweg and Dr. S. A. Sherif, provided direction early on and helped me set realistic goals. The suggestions made regarding clarity, consistency, and analysis of results in my Ph.D. proposal w ere useful in preparing this dissertation and other technical papers on the study. The senior engineering technician at the Solar Energy and Energy Conversion Laboratory, Chuck Garretson, exhibits a demeanor that motivates self reliance. Skilled and willin g to teach, Chuck is an asset to those willing to learn. His practical assistance, advice and instruction were critical for my project. Barbara Graham provided much editorial and secretarial assistance to the solar lab group in writing papers and reports on our projects. In addition, her activism in solar and

PAGE 3

iii environmental issues helped to stimulate the group and increase our involvement in related events. The weekly meeting of the solar lab group was a good arena to practice presentation and discussion sk ills. Attentive graduate students in the group, such as Sanjay Vijayaraghavan, are commended for giving honest criticism, constructive arguments and intelligent suggestions during and after these meetings. Sanjay and Chris Martin, also a graduate student, are recognized for their assistance with my experimental work. As my predecessor, Shaoguang Lu assisted with the first design of the experimental setup and the initial procurement of system components, granting me challenging opportunities in his wake. I r eiterate my appreciation to those that contributed to this dissertation and to making the solar lab a constructive, pleasant, and humorous working environment. Thanks go to my family for their encouragement and support throughout my lengthy educational exp erience. I also acknowledge financial support from the U.S. Department of Energy, Florida Solar Energy Center and NASA.

PAGE 4

iv TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. ii LIST OF TABLES ................................ ................................ ................................ .............. ix LIST OF FIGURES ................................ ................................ ................................ ............. x NOMENCLATURE ................................ ................................ ................................ .......... xiv ABSTRACT ................................ ................................ ................................ ................... xviii CHAPTER 1 MOTIVATION ................................ ................................ ................................ ................ 1 Energy Breakdown and Renewable Resources ................................ ............................... 1 Low Temperature Resourc es ................................ ................................ .......................... 4 Solar Resources ................................ ................................ ................................ ........ 4 Geothermal Resources ................................ ................................ ............................. 6 Waste Heat Resources ................................ ................................ ............................ 10 Biomass Resources ................................ ................................ ................................ 10 Available Methods Fo r Thermal Energy Conversion ................................ ................... 10 Direct Power Production ................................ ................................ ........................ 11 Indirect Power Production ................................ ................................ ...................... 12 2 BACKGROUND AND SUMMARY OF PREVIOUS WORK ................................ .... 17 Overvi ew of the Cycle ................................ ................................ ................................ .. 17 Thermodynamics of the Cycle ................................ ................................ ............... 18 Comparison to Other Cycles ................................ ................................ .................. 20 Theoretical Background ................................ ................................ ................................ 21 Properties of Ammonia Water Mixtur es ................................ ................................ 21 Efficiency Based on Source and Sink Temperatures ................................ ............. 21 Efficiency Based on Energy Transfers ................................ ................................ ... 23 Previous Theoretical Studies ................................ ................................ ......................... 25 Parametric Analysis ................................ ................................ ............................... 26 Optimization ................................ ................................ ................................ ........... 28 Irreversibility Analysis ................................ ................................ ........................... 33 Summary of Theoretical Studies ................................ ................................ ............ 35

PAGE 5

v 3 EXPERIMENTAL SYSTEM ................................ ................................ ........................ 36 Ammonia Water Side ................................ ................................ ................................ ... 37 Hot Water Side ................................ ................................ ................................ .............. 44 Coolant Side ................................ ................................ ................................ .................. 44 Instrumentation ................................ ................................ ................................ ............. 47 Thermocouples ................................ ................................ ................................ ....... 48 Pressure Tra nsducers ................................ ................................ .............................. 49 Gas Chromatograph and Syringe Sampling ................................ ........................... 49 Flow Meters ................................ ................................ ................................ ........... 51 Data Acquisition Hardware and Analysis Software ................................ ..................... 51 Safety ................................ ................................ ................................ ............................ 52 4 EXPERIMENTAL METHODOLOGY ................................ ................................ ......... 54 Parameters in Simulations and Experiments ................................ ................................ 54 Limits and Selection of Operating Conditions ................................ .............................. 56 Heat Source Temperature ................................ ................................ ....................... 59 Boiler Pressure ................................ ................................ ................................ ....... 61 Ambient Temperature ................................ ................................ ............................ 62 Basic Solution Ammonia Mass Fraction ................................ ................................ 63 Heat Source Flow Ratio ................................ ................................ ......................... 63 S ystem Behavior ................................ ................................ ................................ ........... 63 Uncertainty of Measurements ................................ ................................ ....................... 65 5 EXPERIMENTAL RESULTS ................................ ................................ ....................... 69 Vapor Generation ................................ ................................ ................................ .......... 69 Observations ................................ ................................ ................................ ........... 70 Recovered Heat ................................ ................................ ................................ ...... 71 Boiler Heat Input ................................ ................................ ................................ .... 73 Vapor Fraction Leaving the Separator ................................ ................................ ... 75 Vapor and Weak Solution Ammonia Mass Fractions ................................ ............ 77 Absorptio n ................................ ................................ ................................ ..................... 80 Observations ................................ ................................ ................................ ........... 80 Coolant Flow Rate and Temperature ................................ ................................ ..... 81 Absorber Heat Rejection ................................ ................................ ........................ 82 Potential Work Output, Cooling Capacity and Cycle Effic iencies ............................... 83 Parametric Dependence ................................ ................................ .......................... 84 Pump Work Input ................................ ................................ ................................ ... 87 Turbine Work Output ................................ ................................ ............................. 88 Cooling Capacity ................................ ................................ ................................ .... 90 First Law Eff iciency ................................ ................................ ............................... 91 Second Law Efficiency ................................ ................................ .......................... 92 6 RECOMMENDATIONS FOR FUTURE WORK ................................ ........................ 93

PAGE 6

vi Target Areas for Improvement ................................ ................................ ...................... 93 Turbine/Generator Set and Refrige ration Unit ................................ ....................... 93 Absorber Design ................................ ................................ ................................ ..... 94 Boiler Heat Exchanger ................................ ................................ ........................... 94 Ammonia Water Pump ................................ ................................ .......................... 95 Coolant Temperature Fluctuations ................................ ................................ ......... 95 Heat Source Capacity ................................ ................................ ............................. 96 Control Methods ................................ ................................ ................................ ..... 96 Operating Limits ................................ ................................ ................................ .... 97 Insulation ................................ ................................ ................................ ................ 98 Component Modeling in Simulations ................................ ................................ .... 98 Suggested Strategy for System Modifications ................................ .............................. 98 Short Term Plan ................................ ................................ ................................ ..... 98 Long Term Plan ................................ ................................ ................................ .... 100 7 CONCLUSIONS ................................ ................................ ................................ .......... 101 APPENDIX A AMMONIA TOXIC ITY ................................ ................................ ............................. 102 B BINARY FLUID PROPERTY EVALUATION ................................ ........................ 103 Characteristics of Ammonia and Water ................................ ................................ ...... 103 Ideal Models ................................ ................................ ................................ ................ 104 Semi Empirical Correlations ................................ ................................ ....................... 105 Cubic Equations of State ................................ ................................ ...................... 107 Virial Equations of State ................................ ................................ ...................... 108 Gibbs Excess Energy ................................ ................................ ............................ 108 Law of Corresponding States ................................ ................................ ............... 109 Perturbation Theory ................................ ................................ ............................. 110 Group Contribution Method ................................ ................................ ................. 111 Property Correlations Used in the Current Study ................................ ....................... 111 Ammonia and Water Mixtures ................................ ................................ ............. 111 E thylene Glycol and Water Mixtures ................................ ................................ .. 119 C EXPERIMENTAL COMPONENT LIST ................................ ................................ ... 122 Data Acquisition and Electrical ................................ ................................ .................. 122 Fluids ................................ ................................ ................................ ........................... 123 Heat Exchange ................................ ................................ ................................ ............ 123 Instrumentation ................................ ................................ ................................ ........... 124 Pumping ................................ ................................ ................................ ...................... 126 Safety ................................ ................................ ................................ .......................... 127 Valves ................................ ................................ ................................ ......................... 128 Vessels ................................ ................................ ................................ ........................ 129

PAGE 7

vii D EXPERIMENTAL PROCEDURES ................................ ................................ ........... 130 Calibration of Instrumentation ................................ ................................ .................... 130 Thermocouple Calibration ................................ ................................ ................... 130 Pressure Transducer Calibration ................................ ................................ .......... 131 Gas Chromatograph Calibra tion ................................ ................................ .......... 133 Flow Meter Calibration ................................ ................................ ........................ 134 Syringe Sampling Techniques ................................ ................................ .................... 136 Observations on the Behavior of Ammonia Water Mixtures ................................ ..... 138 Normal Operat ing Procedures ................................ ................................ ..................... 139 Startup ................................ ................................ ................................ .................. 139 Operation and Testing ................................ ................................ .......................... 142 Shutdown ................................ ................................ ................................ .............. 144 System Maintenance and Initial Setup ................................ ................................ ........ 145 Data Acquisition Setup ................................ ................................ ........................ 145 System Draining and Cleaning ................................ ................................ ............. 147 System Charging ................................ ................................ ................................ .. 149 Idling in Cold Weather ................................ ................................ ......................... 151 Idling in Hot Weather ................................ ................................ ........................... 151 Emergency Procedures ................................ ................................ ................................ 152 E UNCERTAINTY ANALYSIS ................................ ................................ .................... 153 Theoretical Background of Analysis Method ................................ ............................. 153 Uncertainty of Measured Values ................................ ................................ .......... 153 Uncertainty of Calculated Values ................................ ................................ ........ 156 Uncertainty of State Point Measurements ................................ ................................ ... 157 Temperature Measurements ................................ ................................ ................. 157 Pressure Measurements ................................ ................................ ........................ 159 Ammonia Mass Fraction Measurements ................................ .............................. 162 Flow Rate Measurements ................................ ................................ ..................... 165 Uncertainty of Energy Transfers and Efficiencies ................................ ...................... 167 Boiler Heat Input ................................ ................................ ................................ .. 169 Absorber Heat Rejection ................................ ................................ ...................... 170 Cooling Capacity ................................ ................................ ................................ .. 171 Internal Heat Recovery ................................ ................................ ......................... 172 Pump Work Input ................................ ................................ ................................ 173 Turbine Work Output ................................ ................................ ........................... 174 Vapor Fraction ................................ ................................ ................................ ..... 174 Boiler Heat Exchanger Effectiveness ................................ ................................ ... 175 Recovery Heat Exchanger Effectiveness ................................ ............................. 175 First Law Efficiency ................................ ................................ ............................. 176 Second Law Efficiency ................................ ................................ ........................ 177 Conclusions of the Uncertainty Analysis ................................ ................................ .... 178 F COMPUTER PROGRAM FOR DESIGN AND DATA ANALYSIS ........................ 179

PAGE 8

viii Program Features ................................ ................................ ................................ ........ 179 Property Evaluation ................................ ................................ .............................. 179 Design and Simulation ................................ ................................ ......................... 180 Experimental Data Processing ................................ ................................ ............. 180 Experimental Data Organization and Comparison ................................ ............... 181 Uncertainty Analysis ................................ ................................ ............................ 182 Units Preferences ................................ ................................ ................................ 182 Calculation Preferences and Calibration Constants ................................ ............. 182 Program Structure ................................ ................................ ................................ ....... 182 Property E valuation Functions ................................ ................................ .................... 183 Critical Temperature and Pressure of the Mixture ................................ ............... 184 Bubble and Dew Point Temperature, Pressure and Mass Fraction ...................... 184 Enthalpy of Pure Components and the Mixture ................................ ................... 186 Entropy of Pure Components and the Mixture ................................ ..................... 187 Specific Volume of Pure Components and the Mixture ................................ ...... 188 Determination of All Properties at a State Poi nt ................................ .................. 189 Thermodynamic Analysis Functions ................................ ................................ .......... 193 Determination of All State Points in the Cycle ................................ .................... 193 Calculation of Energy Transfers, Exergy Losses and Efficiencies ...................... 199 Supporting Files for Program Operation ................................ ................................ ..... 203 Storage of Temporary Values ................................ ................................ .............. 203 List of Raw Data Files ................................ ................................ .......................... 205 LIST OF REFERENCES ................................ ................................ ................................ 206 BIOGRAPHICAL SKETCH ................................ ................................ ........................... 211

PAGE 9

ix LIST OF TABLES Table page 4.1 Fluid and component parameters in the simulation and experiment .......................... 56 4.2 Limits of primary operating conditions ................................ ................................ ...... 59 4.3 Exp erimental control measures for maintaining operating parameters ...................... 65 A.1 Ammonia exposure limits ................................ ................................ ........................ 102 B.1 Coefficients for pure water and pure ammonia ................................ ........................ 1 14 B.2 Coefficien ts for the ammonia water Gibbs energy functions ................................ .. 116 B.3 Coefficients for determining bubble and dew point temperatures, critical temperatures, and critical pressures of ammonia water mixtures ........................ 117 B .4 Empirical constants in ethylene glycol and water mixture property evaluation ...... 121 E.1 Chauvenets criterion for elimination of data points ................................ ................ 155 E.2 Uncertainties in single measurements within co nfidence intervals .......................... 156 E.3 Device history of pressure measurement in the system ................................ ............ 162 E.4 Liquid flow meter uncertainties ................................ ................................ ................ 167 E.5 The uncertainties of calculated values and their dependence on measurements for a typical test of the experimental system, within 90% confidence ......................... 168

PAGE 10

x LIST OF FIGURES Figure page 1.1 U.S. energy consumption in 1999 ................................ ................................ ................. 1 1.2 Installed U.S. capacity of renewable electricity production in 1999 ............................ 2 1.3 U.S. ener gy research funding in 2001 ................................ ................................ ........... 3 1.4 The consumption of geothermal and solar energy in the U.S. ................................ ...... 8 1.5 Temperature entropy diagram for the Stirling cycle ................................ ................... 13 1.6 Schematic diagram of the Kalina cycle ................................ ................................ ....... 16 2.1 Schematic of the power and cooling cycle concept ................................ .................... 19 2.2 Temperature entropy diagram of an ideal cycle with a sensible heat source and heat sink ................................ ................................ ................................ ......................... 22 2.3 Effect of turbine inlet pressure on the thermal efficiency of the cycle ....................... 27 2.4 Effect of turbine inlet pressure on the cooling capacity of the cycle .......................... 27 2.5 Efficiencies of the optimized cycle at various heat source temperatures, optimized for second law efficiency ................................ ................................ ............................. 29 2.6 Pressure ratio of the optimized cycle at various heat source temperatures, optimized for second law efficiency ................................ ................................ ....................... 30 2.7 Ratio of refrigeration to work of the optimized cycle at various heat source temperatures, optimized for second law efficiency ................................ ............... 31 2.8 Exergy destruction of the optimized cycle at various heat source temperat ures, optimized for second law efficiency ................................ ................................ ...... 32 3.1 Schematic of the experimental system concept ................................ .......................... 36 3.2 Schematic of the experimental system components for the system side .................... 38 3.3 Photograph of the system side and hot water side of the experiment ......................... 39

PAGE 11

xi 3.4 Photograph of the system side of the experiment ................................ ....................... 39 3.5 Photograph of the system side pump and expansion tank ................................ .......... 40 3.6 Photograph of the system side heat exchangers. ................................ ......................... 40 3.7 Schematic of the absorber design ................................ ................................ ................ 42 3.8 Photograph of the flow indicator to the storage tank ................................ .................. 43 3.9 Photograph of the hot water side of the experiment ................................ ................... 43 3.10 Schematic of the experimental system components for the hot water side .............. 45 3.11 Schematic of the experimental system components for the coolant side .................. 46 3.12 Photograph of the coolant side of the experiment ................................ ..................... 47 3.13 Location of data measurements for system analysis ................................ ................. 48 3.14 Photograph s of pressure measurement instruments on the system. .......................... 49 3.15 Photograph of a syringe sampling port and sight glass on the absorber ................... 50 3.16 Photograph of the gas chromatograph setup ................................ ............................. 50 3.17 Photograph of the temperature displays and coolant flow meter .............................. 51 3.18 Photograph of the storage tank and emergency tank ................................ ................ 53 4.1 Schematic of the cycle concept used in experim ent simulations, showing the main fluid parameters ................................ ................................ ................................ ...... 55 4.2 Expected turbine work output per unit basic solution flow, based on simulation with no losses ................................ ................................ ................................ ................. 57 4.3 Expected refrigeration output per uni t basic solution flow, based on simulation with no losses ................................ ................................ ................................ ................. 58 4.4 Vapor fraction of ammonia water fluid leaving the boiler, based on simulation with no losses ................................ ................................ ................................ ................. 60 4.5 Boiler heat input per unit basi c solution flow, based on simulation with no losses ... 61 4.6 Expected absorber heat rejection per unit basic solution flow, based on simulation with no losses ................................ ................................ ................................ ......... 62 4.7 Expected first law efficiency, b ased on simulation with no losses and with typical losses ................................ ................................ ................................ ...................... 67

PAGE 12

xii 4.8 Expected second law efficiency, based on simulation with no losses and with typical losses ................................ ................................ ................................ ...................... 68 5.1 Qualitative representation of the change in properties during the boiling of the strong solution ................................ ................................ ................................ ................... 70 5.2 Expected internal heat recovery per unit basic solution flow, based on simulation with typical losses ................................ ................................ ................................ .. 72 5.3 Boiler heat inpu t per unit basic solution flow, for various vapor fractions leaving the boiler ................................ ................................ ................................ ...................... 74 5.4 Boiler heat exchanger effectiveness for the initial and improved systems ................. 74 5.5 Vapor fractions in the initial system tests for various boiler temperatures and pressures, with a basic solution ammonia mass fraction of 45.6% ........................ 76 5.6 Vapor fractions in the improved system tests for various boiler temperatures and pressures, with a basic solution ammonia mass fraction of 38.3% ........................ 76 5.7 Ammonia mass fraction in the weak solution for various boiler exit temperatures and pressures, with a basic solution ammonia mass fraction of 45.6% ........................ 78 5.8 Ammonia mass fraction in the vapor for various boiler exit temperatures and pressures, with a basic solution ammonia mass fraction of 45.6% ........................ 78 5.9 Ammonia mass fraction in the weak solution and vapor for a basic solution ammonia mass fraction of 45.6%, for various boiler pressures ................................ ............. 79 5.10 Coolant flow rates for various temperature differences between the coolant inlet and the absorber pool ................................ ................................ ................................ .... 82 5.11 Absorber heat r ejection per unit basic solution flow, for various vapor fractions leaving the boiler ................................ ................................ ................................ .... 83 5.12 Simulated first law efficiency for various turbine and pump isentropic efficiencies, and with other typical losses ................................ ................................ .................. 85 5.13 Simulated second law efficiency for various turbine and pump isentropic efficiencies, and with other typical losses ................................ .............................. 85 5.14 Simulated cooling capacity per unit basic solution flow, for various turbine isen tropic efficiencies and heat source temperatures ................................ ............. 86 5.15 Pump work input per unit basic solution flow, for various pressure rises from the absorber to the boiler ................................ ................................ .............................. 87 5.16 Measured pressure loss acr oss the vapor bubble inlet to the absorber ...................... 89

PAGE 13

xiii 5.17 Measured pressure loss across the weak solution spray inlet to the absorber ........... 89 5.18 Expected work output per unit basic solution flow from a 90% efficient turbine, for observed vapor fractions and expansion ratios ................................ ...................... 90 5.19 Expected first law efficiency with a 90% efficient turbine, for observed vapor fractions and expansion ratios ................................ ................................ ................ 91 5.20 E xpected second law efficiency with a 90% efficient turbine, for observed vapor fractions and expansion ratios ................................ ................................ ................ 92 B.1 Bubble and dew point diagram for a non azeotropic mixture ................................ .. 104

PAGE 14

xiv NOMENCLATURE A area [m 3 ] C gas chromatograph calibration constant C p isobaric heat capacity [kJ/K] COP coefficient of performance D diffusion coefficient [m 2 /s] E exergy [kW] D E change in exergy [kW] F general function or general property G Gibb s energy [kW] H enthalpy [kW] HE heat exchanger effectiveness K number of samples in the set N number of samples in the population P pressure [bar] Q heat transfer rate [kW] or volumetric flow rate [m 3 /s] R molar specific gas constant [kJ/K] or rati o of maximum deviation to standard deviation S entropy [kW] T temperature [ C] V volume [m 3 ] or voltage [V]

PAGE 15

xv W work transfer rate [kW] c p isobaric specific heat capacity [kJ/kgK] g molar specific Gibbs energy [kJ/kmol] or acceleration due to gravit y h molar specific enthalpy [kJ/kmol] hst heat source temperature [K] m mass flow rate [kg/s] q specific heat transfer rate [kJ/kg] s molar specific entropy [kJ/kmol] v molar specific volume [m 3 /kmol] w specific work transfer rate [kJ/kg] x mass fr action of ammonia in the liquid [kg NH 3 (liq) / kg total (liq)] or general measurement x average of a general measurement ~ x mole fraction of ammonia in the liquid [mol NH 3 (liq) / mol total (liq)] y mass fr action of ammonia in the vapor [kg NH 3 (vap) / kg total (vap)] ~ y mole fraction of ammonia in the vapor [mol NH 3 (vap) / mol total (vap)] Greek 1 first law (thermal) efficiency 2 second law (exergy) efficiency p pump isentropic efficiency t turbine isentropic efficiency dynamic viscosity [Pa s] density [kg/m 3 ]

PAGE 16

xvi s standard deviation w uncertainty subscripts 0 reference value or dead state B reference value H high H 2 O water L liquid or low NH 3 ammonia a ammonia or absorber abs absorber avg average b bubble or boiler c critical or cooling or calibration cycle based on the whole cycle d dew ex exit f final or float g vapo r h heating hot hot water hs heat source i pertaining to component i or initial

PAGE 17

xvii in inlet m mixture or measured or pertaining to the mean max upper limit min lower limit p pump r reduced rec recovery refrig based on refrigeration s standard con ditions or superheater str strong solution sys system t turbine tur turbine vap vapor w water weak weak solution superscripts E excess L liquid g vapor in inlet mix mixing

PAGE 18

xviii 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 EXPERIMENTAL INVESTIGATION OF AN AMMONIA BASED COMBINED POWER AND CO OLING CYCLE By Gunnar Olavi Tamm May 2003 Chair: D. Yogi Goswami Major Department: Mechanical and Aerospace Engineering A novel ammonia water thermodynamic cycle, capable of producing both power and refrigeration, was proposed by D. Yogi Goswami. The binary mixture exhibits variable boiling temperatures during the boiling process, which leads to a good thermal match between the heating fluid and working fluid for efficient heat source utilization. The cycle can be driven by low temperature sources such as solar, geothermal, and waste heat from a conventional power cycle, reducing the reliance on high temperature sources such as fossil fuels. A theoretical simulation of the cycle at heat source temperatures obtainable from low and mid temperature solar c ollectors showed that the ideal cycle could produce power and refrigeration at a maximum exergy efficiency, defined as the ratio of the net work and refrigeration output to the change in availability of the heat source, of over 60%. The exergy efficiency i s a useful measure of the cycles performance as it compares the effectiveness of different cycles in harnessing the same source.

PAGE 19

xix An experimental system was constructed to demonstrate the feasibility of the cycle and to compare the experimental results wit h the theoretical simulations. In this first phase of experimentation, the turbine expansion was simulated with a throttling valve and a heat exchanger. Results showed that the vapor generation and absorption condensation processes work experimentally. The potential for combined turbine work and refrigeration output was evidenced in operating the system. Analysis of losses led to modifications in the system design, which were implemented to yield improvements in heat exchange, vapor generation, pump perform ance and overall stability. The research that has been conducted verifies the potential of the power and cooling cycle as an alternative to using conventional fossil fuel technologies. The research that continues is to further demonstrate the concept and d irect it towards industry. On the large scale, the cycle can be used for industrial power production or as a central power plant for a community, with refrigeration produced as required by the application. On the small scale, an affordable residential or c ommercial unit could allow independent electricity generation for the home or business while also cooling it.

PAGE 20

1 CHAPTER 1 MOTIVATION To identify the intended domain of the proposed power and cooling cycle, an overview of low temperature energy resources is presented in this chapter, giving insight into the requirements of systems that convert them to useable energy A discussion of available methods for low temperature resource conversion follows as background and motivation for the development of the proposed power and cooling cycle as an alternative to the established technologies. Energy Breakdown and Renewable R esources The global energy consumption in 1999 was 381.9x10 15 Btu, with the U.S. contributing a quarter of the total energy usage. Fossil fuels account for a majority of the energy supply, and nuclear and renewable resources make up the remainder. natural gas 23.0% coal 22.4% geothermal 0.4% wind <0.1% solar <0.1% biomass 3.3% hydro 3.6% nuclear 8.0% petroleum 39.1% Figure 1.1. U.S. energy consumption in 1999 (Energy Information Administration, 2001).

PAGE 21

2 Renewable energy is defined as a continuous and self propagating flow of energy in the natural environment. There is no initiation energy required to release the stored potential, as in the case of finite or non renewable energy resources (Twidell and Weir, 1986). As renewable energy is cyclical in the environment, the exploitation of it is naturally with less environmental impact than in using fini te energy resources. Figure 1.2. Installed U.S. capacity of renewable electricity production in 1999 (Energy Information Administration, 2001). The installed capacity of renewable electricity production is shown in Fig. 1.2. Hydro is by far the most de veloped renewable resource. The major reason that the global dependence on finite energy resources has not shifted faster to renewable energy is cost. In 2001, only 10% of the U.S. electricity needs were met with renewable resources, while

PAGE 22

3 54% came from co al, 20% from nuclear and 16% from natural gas (Wardell, 2001 a ). In looking at 2001 production costs for power companies, electricity from coal only costs $0.021 per kilowatt hour and nuclear even less at $0.017/kWh. Meanwhile, natural gas and wind generat ion costs twice as much at $0.04/kWh, and other renewable resources cost much more (Wardell, 2001 b ). As the renewable market grows and technology improves, these figures will improve but it will take some time. The U.S. government realizes the environmenta l and political importance of diversifying its energy market from fossil fuels into alternatives, as does the rest of the world. Federal research spending in the U.S. in 2001 was $955 million, with Fig. 1.3 showing the largest allotment for coal. However, considering the large current U.S. dependence on coal, it is positive to note that only 27% of research dollars were for coal research and 30% for renewable resources. hydro 0.5% geothermal 2.6% wind 3.7% natural gas 4.2% hydrogen 4.7% nuclear 6.3% coal 27.2% fusion 25.7% solar 9.4% biomass 8.4% petroleum 7.3% Figure 1.3. U.S. energy research funding in 2001 (Helvarg, 2001).

PAGE 23

4 Low Temperature Resources The proposed power and cooling cycle is designed for use with low temperature resources, which can include solar and geothermal. The combustion of biomass, although a renewable resource, produces a resource temperature tha t exceeds the scope of the proposed cycles design. Indirectly, finite resources can also be used. The proposed cycle can serve as a bottoming cycle for conventional power plants, which reject spent fluids that are cooled beyond primary use but can serve a s the source for a low temperature end cycle. In this way more work can be extracted to improve the overall resource utilization of conventional systems. The following is an overview of resources that can be used to operate the proposed cycle. Namely solar geothermal, and waste heat resources are discussed, to establish their potentials for utilization and to give background on their development. Solar Resources The approximate solar radiation arriving on the earths surface is 1000 W/m 2 reduced by atmosp heric interaction from 1377 W/m 2 on average arriving extra terrestrially (Twidell and Weir, 1986). The actual radiant flux density varies by 1.5% according to solar fluctuations and by 4% according to the distance of the earth from the sun. Ultraviolet light accounts for 9% of the electromagnetic spectrum of solar radiation, visible light for 45% and infrared for 46%. The total daily solar flux to the earth is about 3 30 MJ/m 2 depending on the location, time of the year and weather (Twidell and Weir, 1 986). Solar radiation arrives on the earths surface as beam and diffuse radiation. On a clear day, the beam radiation can account for 90% of the total while on an overcast day diffuse radiation can account for 100% of the total (Twidell and Weir, 1986). O nly beam radiation can be focused.

PAGE 24

5 Owing to the electromagnetic nature of solar radiation, it can be converted directly to electricity through photovoltaic means. Photovoltaic cells were first introduced in 1954, and developed largely for the space industr y. Photovoltaic devices operate with the principal of solar radiation exciting electrons to separate from certain materials and move in the form of a current to produce electricity. Photocells typically produce potentials of 0.5 V, which can be compounded and current densities of 200 A/m 2 with clear sky radiation of 1000 W/m 2 (Twidell and Weir, 1986). Typical efficiencies of commercial units are 10 15%, producing 1 2 kWh/m 2 per day. The National Renewable Energy Laboratory set a laboratory thin film solar c ell conversion record of 18.8% in 1998 (Wolcott, 1999). Other PV technologies can extend efficiencies by a few more percent, although at a very high cost. The cost of solar cells has dropped sharply from $100 per Watt in 1974 to about $5 per Watt in 2001. However, the cost should fall to less than $1 per Watt to become commercially competitive in the electricity market. Solar thermal technology is much more competitive currently than photovoltaics for residential use. Domestic hot water accounts for 18% of a homes energy usage, which has prompted the development of cheap solar thermal collection at a cost of $100 to $300 per square meter (International Energy Agency, 1991; Valenti, 1999). With flat plate collectors, temperatures around 100 C can be achieve d in the collection medium. In regards to capital cost and energy savings, residential and commercial solar thermal heating systems have reasonable payback periods and have gained popularity globally. The typical annual efficiency of solar hot water system s is 35 40% (International Energy Agency, 1991).

PAGE 25

6 Solar thermal use for power generation has been limited to mid and high temperature collection, as efficient power cycles for low temperature resources have not been thoroughly explored. The power production cost has been found to go down dramatically with system size, so most systems are on the large commercial scales. Systems using parabolic trough concentrators get transfer fluid temperatures around 700 C. Parabolic dish concentrators can achieve temperat ures up to 3000 K (Twidell and Weir, 1986). For large solar systems, two methods are used for mid temperature cycles. With distributed collection, many collectors are networked which heat a fluid to high temperatures. This fluid can be steam, which is expa nded through a turbine, or an intermediary fluid such as ammonia, which is dissociated to store chemical energy. The recombination gives off heat that can be continuously extracted, even at night. The heat from this can power a steam turbine. Power towers are the other method, with an array of reflectors focused on a single collection point (Twidell and Weir, 1986). Both of these methods for solar thermal electricity are infeasible on residential scales, as the cost would be tremendous to the homeowner. The typical efficiency for mid and high temperature solar thermal power generation is about 20% (International Energy Agency, 1991). The cost of solar collection is minimized in using flat plate collectors, although the temperatures produced in the transfer f luid are not sufficiently high to efficiently fuel conventional power cycles. The proposed cycle addresses this issue by being able to utilize low temperature resources, providing a cost savings potential by being able to use inexpensive flat plate collect ors. Geothermal Resources Geothermal energy is loosely defined as the natural heat of the earth. Practically, geothermal energy is economically viable if it is concentrated into finite regions, and

PAGE 26

7 close enough to the surface so it can be readily extracted The heat flow outwards from the earth is relatively constant geographically, with pockets of variation where there are cracks in the crust in tectonically active regions. This is where geothermal resources abound. The three primary geothermal sources ca n be characterized as hot igneous, conduction dominated and hydrothermal (Chilingar et al., 1982; International Energy Agency, 1991). Hot igneous sources include molten rock and dry rock. Molten rock is liquid magma chambers that are too high in temperatur e (650 1200 C) and too deep (>3 km) to harvest any energy from with current technologies. Dry rock is solidified magma, closer to the surface at less than 3 km, and cooler at less than 650 C. A dry rock geothermal system has been proven at Los Alamos by creating a heating water loop through it, but at too much cost to become practical (Chilingar et al., 1982). Conduction type sources transfer heat to the surface through conduction from great depths. In order to reach temperatures over 100 C to become use able in power cycles, the well depth would need to exceed 5 10 km which becomes too expensive to be practical. The hydrothermal sources are the most convenient and common geothermal sources to harvest. Hot water and steam saturate porous and permeable rock s, with natural convection currents moving the hot fluids towards the surface while cooler fluids sink. The total heat flow from the earth is estimated at 42 x 10 12 W, or roughly 82 mW/m 2 on the surface. The internal generation is largely by radioactive de cay of isotopes of uranium, thorium and potassium. The thermal content of the earth is vast, estimated at 12.6 x 10 24 MJ above a reference temperature of 15 C. However, the fraction of this geothermal energy that can be utilized is rather small. Geotherma l energy for electrical

PAGE 27

8 generation has been exploited in Italy for nearly 100 years, beginning in 1904. New Zealand has been exploiting geothermal power for 50 years, and the western U.S. for 40 years (Armstead, 1978; Chilingar et al., 1982). Fig. 1.4 show s the early development of geothermal power was well before solar. Most U.S. geothermal sites are on federal lands, which were opened up for private leasing by energy companies in 1970 with the Geothermal Steam Act (Kuwada, 1972). In 1993 the global geothe rmal power generation was only 5915 MWe, of which 43.9% was generated in the U.S., 15% in the Philippines, 12.7% in Mexico, 10.8% in Italy, and the balance in other countries. However the industry is gaining ground internationally, as the total generation was up 72% from 10 years prior (Dickson and Fanelli, 1995). Figure 1.4. The consumption of geothermal and solar energy in the U.S. (Energy Information Administration, 2001).

PAGE 28

9 A deterrent for the geothermal industry growth is the economical risk in identi fying the source. Location of geothermal sources can prove costly. Indications of potential sources may include geysers, seismic or volcanic activity, and hot springs. Given the initial surface testing, shallow boreholes are made and finally deep wells. Th e success rate rests at about 10%, with a $1 million cost for each site that is tested, in a 1982 estimate (Chilingar et al., 1982). The high cost has hindered extensive exploration. Once a geothermal site is found, it can also be difficult to determine th e total available energy of the site, which is necessary to gauge the worth of the investment. Strictly speaking, geothermal energy is not locally renewable. Exploitation of geofluids will undoubtedly prompt a geological response, which could result in a f inite lifetime for a geothermal site before it cools off (Dickson and Fanelli, 1995). So although the global geothermal potential is virtually limitless, local concentrations with high geothermal gradients are bounded and not renewable (Armstead, 1978). Es timates must be made for the geothermal gradient, heat flow to the surface, and heat capacity of the fluid holding rocks, which are often only accurate to within one order of magnitude. The potential site must be selected prudently. As an example of the re source potential stored at a geothermal site, the Appalachian region of Virginia, North Carolina and Tennessee could provide the geothermal energy equivalent of 30 35 trillion barrels of crude oil from its dry rock heat store (Chilingar et al., 1982). The availability of geothermal resources is not disputed. However, economical exploitation of such an immense resource becomes the issue in resource location and utilization. The immediate environmental impact of geothermal systems is also not negligible. As t he geofluid is not pure water, there may be issues with

PAGE 29

10 surface water pollution, air pollution from gases dissolved in the geofluid that are freed as it cools, and underground pollution. Other concerns likewise are not taken lightly, although it is anticip ated that geothermal energy will have much less long term environmental impact than conventional energy sources (Dickson and Fanelli, 1995). Waste Heat Resources Waste heat resources can be loosely defined as the unused energy of conventional power cycles. This can be from conventional fossil fuel plants, from biomass combustion plants, or nuclear reactors. The waste heat is rejected in fluids that are too low in temperature for useful conversion in the primary cycle but can be hot enough to power a bottomi ng cycle. This allows extraction of more work from the original energy source, and raises the overall efficiency of the system. Biomass Resources Biomass is defined as energy derived from the sun and converted through plant photosynthesis into organic mate rial. The stored energy can be in the form of plants or animal waste. The largest potential for biomass harvesting is in wood products. The estimated annual global resource of biomass is 170 billion tons, with less than 1% currently harvested (Internationa l Energy Agency, 1991). The most common conversion scheme for biomass is direct combustion, although the fuel can be converted to more practical liquid, gas or solid forms. Biomass combustion occurs at 500 600 C, which is beyond the design of the proposed cycles use. However, biomass is mentioned here as it is used as an energy source in comparative cycles. Available Methods For Thermal Energy Conversion The thermal conversion process is initiated with the collection of the resource heat. The heat in the collection fluid is either converted to power directly, or the fluid

PAGE 30

11 serves as a storage medium to later power another working fluid. The advantage of the storage medium is that it allows for off hour conversion of solar radiation. The disadvantage is that multiple intermediary conversion processes all exhibit additional inefficiencies. Some methods of conversion to electricity are specific to the resource, and some are applicable to multiple resources. Available methods for thermal conversion of low and mi d temperature resources are presented for comparison to the proposed cycle. Direct Power Production Conversion methods that exhibit direct power production include processes in which the fluid heated by the source is the only intermediary, such that power is produced from it directly as the working fluid. Solar steam expansion High temperature steam is possible using solar concentrators. Water is pumped to high pressures, then boiled in the concentrator. The high temperature and pressure steam then expand s through a turbine producing power. The steam is condensed and can be recycled back through the system to preserve exergy. Geothermal steam expansion Geothermal plants operate with a variety of cycles. In the simplest cycle, steam is separated out from t he geofluid and allowed to expand through a turbine, exiting to the atmosphere. These systems tend to be the cheapest in capital cost ($1000/kW net ) for geothermal use, although they require roughly twice the steam input per unit power production than syste ms that condense the expanded steam. The goal of the condensation is to expand the steam through the turbine to much lower pressures than atmospheric, allowing for greater turbine work output. The condensed steam must then be pumped to atmospheric pressure s for discharge. Drawbacks of the condensing design include additional components needed to pump non

PAGE 31

12 condensable gases from the condenser, and cooling towers to extract heat from the condenser. This all adds to the capital cost by roughly 50%, although ove rall the cost is less per amount of steam extracted (Hudson, 1995). Solar chimney Based on the principles of buoyancy, solar heated air rises through a converging tower, speeding up as it rises. The fast moving air rises through a turbine, which is couple d to a generator to produce electricity. The efficiencies of solar chimneys are on the order of 1 2% percent, and they must be extremely tall to produce significant output. The design is simple but becomes impractical at great heights. Indirect Power Produ ction Conversion methods that exhibit indirect power production include processes in which the fluid heated by the source is an intermediary transfer fluid, which subsequently transfers energy to the working fluid that produces power. The power production can occur simultaneously with the heat extraction form the source, or the charged transfer fluid can be stored for later generation. Stirling engine. The allure of the Stirling engine has been the high potential efficiency, operating ideally with the same efficiency of a Carnot cycle between the same source and sink reservoir temperatures, as shown in Fig 1.5. Although the cycle was first developed in 1816, it subsided to the internal combustion engine until recent growing interest (Morrison, 1999). The bas ic cycle includes an isothermal compression, heat addition at constant volume, isothermal expansion, and finally heat rejection at constant volume. The Stirling engine is an external combustion engine and tends to produce less pollution than internal combu stion engines (Moran and Shapiro, 1992). It can be powered by concentrated solar power, hot igneous geothermal resources, or with the heat from

PAGE 32

13 burning biomass (Morrison, 1999). Biomass is combusted in the 500 600 C range, similar to mid temperature solar and geothermal collection, at which the Stirling cycle exhibits high thermal efficiencies over 60%. Low temperature solar and geothermal can also be used to fuel Stirling engines, but at lower thermal efficiency limits of 20 35%. e n t r o p y t e m p e r a t u r e 1 H 3 4 2 L c o n s t a n t v o l u m e c o n s t a n t v o l u m e Figure 1.5. Temperatur e entropy diagram for the Stirling cycle. Rankine cycle. Rankine cycles used as secondary or binary cycles in low temperature applications have been limited to geothermal resources. Binary Rankine systems are the most common of geothermal power cycles. Sol ar power generation at low temperatures is too costly using Rankine cycles, and so solar thermal applications have been limited to mid and high temperature regimes. For low temperature resource fluids under 150 C in geothermal systems, it becomes more eco nomical to use a secondary Rankine cycle for power production than a direct power production method. Typically the secondary fluid for geothermal applications has been isobutane (Bliem, 1980), although other fluids can be used

PAGE 33

14 (Armstead, 1978). The working fluid has a lower boiling point than water, such that the evaporation process in the Rankine cycle typically occurs at low temperatures. The high pressure working fluid is vaporized and expands through a turbine, finally being condensed by heat exchange w ith cooling water. For geothermal applications the heat source is a geothermal brine, which is reinjected into the ground after heat is extracted from it by the working fluid. Rankine cycles are appealing in situations that require the geofluid to remain l iquid, as it can be pressurized during the extraction and reinjection processes to prevent flashing. Flashing is harmful to the well as it allows for fouling by dissolved substances, leaving a residue on components (Hudson, 1995; Kuwada, 1972). Improvement s to the geothermal Rankine cycle can be made in the form of supplemental heating. It has been shown that raising a geothermal low temperature resource temperature with an additional heating from solar, biomass, fossil fuel, or other means results in a lar ge payback from the primary Rankine cycle. The extra energy output from the combined system is about 4 times the supplemental energy added to it (Subbiah and Natarajan, 1988). Low to mid temperature geothermal systems operating with binary Rankine cycles c an utilize source temperatures of 85 175 C. Typical plant costs depend highly on the geothermal fluid temperature. For a 140 C resource temperature, the plant capital cost will be about $1900/kW net excluding the cost of the wells (Dickson and Fanelli, 1 995). Kalina cycle. Improvements in the thermal efficiency over the conventional Rankine cycle can be achieved by increasing the number of boiling steps, operating at

PAGE 34

15 supercritical conditions, or using a mixed working fluid. The mixed working fluid provide s a varied boiling temperature, but with a conventional Rankine cycle this will cause a varied condensation temperature for only moderate overall improvement (Kalina, 1984). The Kalina cycle, shown in Fig. 1.6, is designed for use as a bottoming cycle or w ith low temperature heat sources. A multicomponent working fluid is used, with a concentration dependent boiling temperature. This allows for a good thermal match in the boiler with the sensible heat source and yields less exergy loss. Better heat utilizat ion allows for lower component costs of the evaporator and condenser, as less surface area is required for the same power output. Besides the variable boiling temperature being able to extract more heat from the heat source, there is also more internal hea t recovery. A high pressure stream can be boiled by a low pressure condensing stream, since the low pressure stream can have phase changes at higher temperatures than the high pressure stream according to the ammonia mass fraction of the mixture. This cann ot happen in a single component Rankine cycle (Enick et al., 1998). Parametric analysis and optimization were performed for the Kalina cycle operating with a binary ammonia and water mixture, with properties determined according to the method outlined by E l Sayed and Tribus. Analysis showed turbine inlet concentration and separator temperature as the dominant parameters. The vapor exiting the turbine is then used for distillation of the basic solution. This is more efficient than using it for evaporation (M arston, 1990). The addition of the Kalina cycle as a bottoming cycle raises the overall thermal efficiency of the conventional gas turbine system to 50 52%. The thermal efficiency of

PAGE 35

16 the Kalina cycle alone is sometimes misleading. It is more appropriate to present a second law or exergy efficiency for low temperature cycles. The Kalina cycle has been shown to be 16 to 19% more efficient than a comparative Rankine cycle operating with the same heat source and sink. Payback is estimated at 1.5 years versus 6. 5 for Rankine cycle systems used as bottoming cycles, based on $0.06 per kWH electricity cost (Kalina, 1984). Figure 1.6. Schematic diagram of the Kalina cycle (Kalina, 1984).

PAGE 36

17 CHAPTER 2 BACKGROUND AND SUMMA RY OF PREVIOUS WORK This chapter introduces the combined power and cooling cycle as an alternative to the established methods of low temperature energy conversion. Following an overview of the cycle concept, previous theoreti cal work is outlined and key results are summarized as a motivation for the current experimental work. Overview of the Cycle A new energy conversion cycle has been proposed by D. Y. Goswami (1995), which can be used as a bottoming cycle using waste heat fr om a conventional power cycle or an independent cycle using low temperature sources. The motivation is to improve the effectiveness of harnessing renewable energy resources such as solar and geothermal energy. With the subsequent reduction in cost from suc h improvement in energy conversion, renewable resources can become more competitive to conventional energy technologies, and the global reliance on fossil fuels may be reduced. The proposed cycle derives both usable power and cooling from the heat of the e nergy source. On the large scale, the combined cycle could be used for industrialized power production or as a centralized power plant for a community, with refrigeration produced as required by the application. On the small scale, an affordable residentia l or commercial unit could allow independent electricity generation for the home or business while also cooling it, all with the energy of the sun.

PAGE 37

18 Thermodynamics of the Cycle Multi component working fluids in power cycles exhibit variable boiling tempera tures during the boiling process which make them suitable for a sensible heat source (Ibrahim and Klein, 1996; Kalina, 1984). The temperature difference between the heat source and the working fluid remains small to allow for a good thermal match between t he source and working fluid, such that less irreversibility results during the heat addition process. The proposed cycle combines the Rankine and absorption refrigeration cycles, using a binary ammonia water mixture as the working fluid. An ammonia water m ixture is used as it exhibits desirable thermodynamic properties, such as a large heat capacity. Ammonia is relatively inexpensive, can accommodate system design modifications well and separates easily from internal lubricating oils (Norton, 2001; Pillis, 1993). Lower viscosities of ammonia require smaller piping sizes and are less taxing on pumps. Ammonia is also environmentally friendly, with no ill effects on global warming or the ozone layer (Pillis, 1993). Ammonia water mixtures as the working fluid in power cycles have shown higher efficiencies than with the conventional Rankine cycle using water or another single component fluid alone (Bogart, 1981; Thorin et al., 1998). Drawbacks of using ammonia include it being toxic (see Table A.1 in Appendix A) a nd corrosive, with a mild flammability range of 16 to 27% in air (Pillis, 1993). A schematic of the cycle is shown in Fig. 2.1. The relatively strong basic solution of ammonia water leaves the absorber as saturated liquid at the cycle low pressure. It is p umped to the system high pressure and is preheated before entering the boiler by recovering heat from the weak solution returning to the absorber. As the boiler operates between the bubble and dew point temperatures of the mixture at the system high

PAGE 38

19 pressu re, partial boiling produces a high concentration saturated vapor and relatively low concentration saturated liquid. The liquid weak solution gives up heat in the recovery unit and throttles into the absorber. The rectifier condenses out water to further p urify the vapor, by rejecting heat to a secondary strong solution stream, before entering the boiler. The vapor is superheated and expanded through the turbine to produce work. Due to the low boiling point of ammonia the vapor expands to low temperatures y ielding the potential for refrigeration. The vapor is finally absorbed back into the liquid, giving off heat that is rejected as the cycle heat output. Figure 2.1. Schematic of the power and cooling cycle concept. The main parameters that can be varied to influence the cycle are the heat source temperature, system high pressure, basic solution ammonia mass fraction, ratio of working and heating fluid flow rates, and absorber pressure and temperature. Saturation in the absorber reduces the number of inde pendent main parameters to five that govern

PAGE 39

20 the cycle. Rectifier and superheater temperatures can also be modified, and the conditions of heat transfer from the source to the ammonia water mixture as well. The cycle can be driven by different heat sources including solar, geothermal, and low temperature waste heat. The use of mid and low temperature solar collectors to drive the combined cycle was investigated by Goswami and Xu (1999), while using geothermal energy as a heat source was analyzed by Lu (2001 ) and Tamm et al. (2001). Typical working conditions of a 400 K boiler superheated to 410 K, and an ambient at 280 K yield a first law efficiency of 23.5% if work and cooling are added as the net output. In comparison, the equivalent Carnot efficiency is 3 1.7%. Conventional power cycles operating between the same temperatures have lower first law efficiencies than the proposed cycle, as no cooling output is included. At higher temperatures, however, their thermal efficiencies are better. The thermal efficie ncy is deceiving though, and the strength of this cycle lies rather in the heat source utilization. It exhibits much higher second law efficiencies than conventional power cycles at the same temperatures. Comparison to Other Cycles For utilization of low t emperature resources, the proposed cycle offers several advantages in comparison to other thermal energy conversion methods. In most instances, low temperature heat from geothermal and solar resources is not an option to provide direct steam expansion, as the resource temperature is too low. The solar chimney is highly inefficient unless built to large scales, at which point it becomes impractical. The Stirling and Rankine cycles become dominant only for mid and high temperature resources. The Kalina cycle is really the only one that merits a closer look. The primary advantage of the proposed cycle over the Kalina cycle is the possibility of refrigeration output. In the proposed cycle, the heat rejection occurs at

PAGE 40

21 much lower temperatures than in the Kalina c ycle because the vapor is absorbed into the bulk fluid while giving off latent heat. Thus the vapor is allowed to expand to much lower temperatures than in the Kalina cycle, allowing for refrigeration to be a byproduct of the power production process. The Kalina cycle also operates best for higher heat source temperatures, such that it is not as suitable for low temperature solar, geothermal or waste heat resources. Theoretical Background Thermodynamic definitions and principles to evaluate the cycle are di scussed in this section, for comparing the proposed cycle to other established cycles that also use low temperature, sensible heat sources. The theoretical limit in the performance of the cycle is given in terms of source and sink temperatures, while the a ctual performance is gauged from calculated energy transfers. Properties of Ammonia Water Mixtures There are several studies on the evaluation of ammonia water mixture properties in the literature. A convenient semi empirical scheme is used here that combi nes the Gibbs free energy method for mixtures and bubble and dew point temperature correlations for phase equilibrium. The calculated results have been compared to experimental mixture properties in the literature with good agreement (Xu and Goswami, 1999) The property evaluation method is further described in Appendix B. Efficiency Based on Source and Sink Temperatures The Carnot efficiency of a cycle operating between two reservoirs at constant temperature is given as Eq. 2.1, where T H is the source temp erature and T L is the sink temperature.

PAGE 41

22 H L T T = 1 1 h (2.1) The heat source is typically sensible, however, and exhibits a temperature change during the energy transfer. Assuming that the heat source enthalpy is a linear function of the temperat ure, the cycle thermal efficiency can be derived as Eq. 2.2 for a sensible heat source, using an entropic average temperature. = f H i H f H i H L T T T T T , , 1 ) ln( 1 h (2.2) e n t r o p y t e m p e r a t u r e 1 3 4 2 L i H f H i L f Figure 2.2. Temperature entropy diagram of an ideal cycle with a sensible heat source and heat sink. If the sink temperature varies as well, as shown in Fig. 2.2, then with the same assumptions the cycle thermal efficiency can be expressed as Eq. 2.3 for a sensible heat source and sink. Variation of the heat sink temperature in addition to the heat source temperature allows for greater exergy potential.

PAGE 42

23 = ) ln( ) ln( 1 , , , , 1 f L i L f L i L f H i H f H i H T T T T T T T T h (2.3) The thermal efficiency in Eq. 2.3 multiplied by the heat source input gives the maximum possible power output for a cycle operating between a sensible source and sink. T he bracketed terms represent the minimum part of the source input that must be rejected as heat. Efficiency Based on Energy Transfers Knowing properties at the state points in the cycle of Fig. 2.1 allows for mass and energy conservation equations to be wr itten over components and for the cycle as a whole. With minimal assumptions, the work and heat transfers can be determined for the cycle in steady state. The cycle first law or thermal efficiency is defined as the net useful energy product divided by the total energy input, given by Eq. 2.4 or equivalently by Eq. 2.5. h c net Q Q W + = 1 h (2.4) s b c p t q q q w w + + = 1 h (2.5) The net work includes both the turbine output and pump input. Q c is the refrigeration capacity and Q h is the total heat added to the cycle from the heat source in both the boiler and superheater. These are calculated based on simple mass and energy balances over the cycle components. This definition for first law efficiency can be deceiving, as the availability in refrigeration is l ess than that in work. The addition of work and cooling output in a single cycle is not documented in the literature, as such a combined power and refrigeration

PAGE 43

24 cycle is a new idea. Finding and justifying an appropriate definition for first and second law efficiencies including both work and cooling is a new concept, and is being explored as motivated by this cycle. A first comparison can be made by scaling the refrigeration term with the coefficient of performance (COP) of a refrigeration cycle operating b etween the appropriate temperatures. The Carnot COP is used as a limit, although typically the COP is much lower at about 3. Eq. 2.6 gives a modified first law efficiency if power output is the primary goal and Eq. 2.7 can be considered as a new cooling ef ficiency or coefficient of performance, if refrigeration is the desired output. h refrig c net Q COP Q W + = 1 h (2.6) h c refrig net cycle Q Q COP W COP + = (2.7) Exergy, or availability, is defined as the maximum reversible work a substance can do during the process of reachi ng equilibrium with its environment. The second law efficiency, or exergy efficiency, for a solar heat source is defined as the exergy output divided by the exergy input to the cycle (Cengel and Boles, 1998). The exergy input is taken as the exergy change of the heat source. The exergy output is the exergy of the net work and the exergy of the refrigeration. The second law efficiency for a solar heat source is given by Eq. 2.8. It measures how much useful output can be derived from a change in heat source e xergy. hs c net E E W D + = 2 h (2.8)

PAGE 44

25 The exergy of refrigeration, E c is the refrigeration capacity divided by the coefficient of performance of a Carnot refrigeration cycle operating between the ambient and cycle low temperatures, as given by Eq. 2.9 (Sz argut et al., 1988). ( ) c c c c T T T Q E = 0 (2.9) Note that in Eq. 2.8, the definition was based on the exergy change of the heat source. Solar systems recycle any unused heat source fluid back to the solar collector, so no exergy is wasted. On the other hand, the unused heat source fluid in geothermal systems is dumped, therefore a more appropriate second law definition for these systems is given by Eq. 2.10. This measures how efficiently the maximum amount of available energy is converted to useful outp ut. in hs c net E E W + = 2 h (2.10) In the definitions of Eqs. 2.9 and 2.10, there is some difficulty in justifying the addition of work with an equivalent cooling term. There is no standard definition available for the second law efficiency of a combined power and cooling cycle. Other definitions can be argued and are being investigated for this cycle, but the above are used for a first analysis. Previous Theoretical Studies The theoretical work that has been performed on this cycle is documented by Goswam i and Xu (1999), Lu (2001) and Tamm et al. (2001), and is summarized here as background and motivation for the current work. A parametric analysis of the cycle and optimization of its performance are reviewed, with the inclusion of losses in an

PAGE 45

26 irreversibi lity study. This summary verifies the feasibility of the cycle concept and establishes the necessity for an experimental investigation. Parametric Analysis Operating conditions were individually varied in a straightforward parametric analysis to study the effects on the energy transfers and efficiencies of the combined cycle (Goswami and Xu, 1999). The parametric analysis gave insight into the behavior of the cycle, and showed that optimization of the cycle would be possible for first or second law efficien cy, as well as work or cooling output. Figure 2.3 is a sample of the parametric study, showing a peak in the thermal efficiency as defined by Eq. 2.4. Figure 2.3 shows that for higher ammonia mass fractions in the basic solution, more vaporization occurs f or the given boiler temperature and pressure. This higher vapor fraction allows for greater flow through the turbine, and more work production for a higher thermal efficiency. Figure 2.4 concludes that more vapor is available for refrigeration output also, per kg of basic solution that is boiled. Figures 2.3 and 2.4 were evaluated for a boiler at 400 K (260 F), superheater at 410 K (278 F), absorber at 280 K (44 F) and rectifier at 360 K (188 F). The low pressure in the system at the absorber was set at 2 bar (14.3 psig). For increasing turbine inlet or system high pressure, the figures show that the work and cooling outputs peak. Determining the location of these peaks is necessary a priori to optimize a working system's performance. The effect of the h igher pressure in limiting vapor production begins to dominate as the boiler exit fluid is shifted towards saturated liquid. The peak shifts for higher basic solution ammonia mass fractions as a two phase equilibrium can be sustained at higher pressures fo r higher mass fractions.

PAGE 46

27 14 16 18 20 22 24 16 18 20 22 24 26 28 30 32 34 turbine inlet pressure (bar) x=0.47 x=0.5 x=0.53 thermal efficiency (%) Figure 2.3. Effect of turbine inlet pressure on the thermal efficiency (%) of the cycle (Goswami and Xu, 1999). 0 5 10 15 20 25 30 16 18 20 22 24 26 28 30 32 34 turbine inlet pressure (bar) x=0.47 x=0.5 x=0.5 3 cooling capacity (kJ/kg) Figure 2.4. Effect of turbine inlet pressure on the cooling c apacity (kJ/kg) of the cycle (Goswami and Xu, 1999). Note that a series of similar plots can be determined by changing any of the operating parameters. Each plot could conceivably provide a visual location of the optimum over the range of interest for the single parameter. For practical operation, the

PAGE 47

28 cycle has several parameters that are varied together, presenting a multidimensional surface on which an optimum can be found. A mathematical approach at locating this optimum is necessary. Optimization Initia l parametric studies of the cycle showed the potential for the cycle to be optimized for first or second law efficiency, as well as work or cooling output. For a solar heat source, optimization of the second law efficiency is most appropriate, since the sp ent heat source fluid is recycled through the solar collectors and the unused exergy is not wasted. For a geothermal heat source, the second law efficiency allows an interpretation of the quality of energy usage from the resource fluid. The second law effi ciency is a useful measure of the cycles performance as it may be used to compare the effectiveness of different cycles in harnessing the same source. The optimization work presented here was performed for solar heat sources, and the second law efficienc y is that given by Eq. 2.8. Work has been conducted for geothermal systems as well (Lu, 2001), with the second law efficiency as in Eq. 2.10. Optimization methodology A Generalized Reduced Gradient (GRG) scheme is used for the optimization (Tamm et al., 2 001). The GRG method searches a feasible region bounded by equality and inequality constraints. Moving finitely towards a better value with a newly determined search direction at every step, the optimum is ultimately reached within a limit of convergence. The optimization scheme searches over eight free variables for the optimal second law efficiency as defined by Eq. 2.4. The parameters are the absorber or ambient temperature, boiler, superheater and rectifier temperature, the boiler pressure (high pressu re), absorber pressure (low pressure), and heat source inlet and exit temperatures.

PAGE 48

29 From these eight free variables all other state points in the cycle can be determined with minimal and reasonable assumptions, neglecting potential and kinetic energies. Op timization results The optimization results verified that the cycle could be optimized using the Generalized Reduced Gradient method. The desired heat source temperature will vary according to the intended use of the cycle. The effects of heat source temp erature on the optimized cycle performance are shown in Figs. 2.5 to 2.8, which are optimized for the second law efficiency at the chosen heat source temperature. Except for the heat source and sink temperatures, which are set, the operating parameters are determined from the optimal solutions. 3 0 0 3 5 0 4 0 0 4 5 0 h e a t s o u r c e t e m p e r a t u r e ( K ) 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 e f f i c i e n c y ( % ) e x e r g y f i r s t l a w Q c / Q h W n e t / Q h Figure 2.5. Efficiencies of the optimized cycle at various heat source temperatures, optimized for second law efficiency (Tamm et al., 2001). The refrigeration as a fraction of the heat addition, Q c /Q h changes li ttle as the heat source temperature increases as shown in Fig. 2.5. As the heat source temperature approaches the ambient temperature, refrigeration approaches zero. The highest

PAGE 49

30 refrigeration fraction is near a source temperature of 390 K (242 F). The ref rigeration fraction decreases to zero near 480 K (404 F), as the higher temperature vapor can no longer be expanded to sub ambient temperatures. 3 0 0 3 5 0 4 0 0 4 5 0 h e a t s o u r c e t e m p e r a t u r e ( K ) 0 2 4 6 8 1 0 1 2 1 4 1 6 p r e s s u r e r a t i o Figure 2.6. Pressure ratio of the optimized cycle at various heat source temperatures, optimized for second law efficiency (Tamm et al., 2001). The net power as a fraction of heat addition, W net /Q h increases as the heat source temperature increases. As turbine work output is related mainly to the pressure ratio across the turbine, the net power curve can be ex plained in relation to the pressure ratio in Fig. 2.6, which shows a continuous increase with heat source temperature. The first law efficiency curve, which is a sum of the refrigeration and power curves, shows similar behavior to the power curve up to the maximum value of 23.6% at 400 K (260 F). After the maximum point the efficiency starts decreasing slowly in a similar manner to the refrigeration curve.

PAGE 50

31 The second law efficiency shows a maximum value of 65.2% at 380 K (224 F). The sharp increase in Fig 2.5 of the second law efficiency between 320 and 380 K (116 to 224 F) is due to the increase of both power and refrigeration outputs. The second law efficiency reaches a maximum where the refrigeration output begins to decrease above 400 K (260 F). Fig ure 2.7 shows the refrigeration to net power ratio versus heat source temperature. In the temperature range between 320 and 360 K (116 to 188 F) this ratio changes rapidly, while above 360 K (188 F) the ratio decreases slowly, reaching 0.12 at 460 K (368 F). Thus, increasing the heat source temperature favors production of power rather than refrigeration. 3 0 0 3 5 0 4 0 0 4 5 0 h e a t s o u r c e t e m p e r a t u r e ( K ) 0 0 2 0 4 0 6 0 8 1 Q c / W n e t Figure 2.7. Ratio of refrigeration to work of the optimized cycle at various heat source temperatures, optimized for second law efficiency (Tamm et al., 2001). Figure 2.8 shows normalized exergy destruction in the cycle as a function of the heat source temperature. The total exergy destruction in the cycle increases with an

PAGE 51

32 increase in the heat source temperature, such that more of the heat source ava ilability is wasted. It can be seen in Fig. 2.8 that the exergy destruction in both the absorber and heat exchanger changes little as the source temperature increases. The superheater has almost no exergy destruction because of its small heat load. The boi ler exergy destruction is much lower than that of the absorber. 3 0 0 3 5 0 4 0 0 4 5 0 h e a t s o u r c e t e m p e r a t u r e ( K ) 0 0 0 5 0 1 0 1 5 0 2 e x e r g y d e s t r u c t i o n / Q h X X X X X X X X X X X X X X X t o t a l a b s o r b e r r e c t i f i e r s u p e r h e a t e r b o i l e r h e a t e x c h a n g e r X Figure 2.8. Exergy destruction of the optimized cycle at various heat source temperatures, optimized for second law efficiency (Tamm et al., 2001). From the exergy analysis, if the heat so urce is between 320 and 460 K (116 to 368 F), then the best operating heat source temperature is around 380 K (224 F), since it gives the maximum second law efficiency. Exergy destruction in the rectifier increases throughout in Fig. 2.8, as the heating load also increases in the rectifier. The sharp increase of rectifier exergy destruction is due to more rectification needed for higher heat source temperatures in order to obtain a high purity vapor stream and thus refrigeration. At even higher temperatur es, it becomes

PAGE 52

33 too costly to produce refrigeration from a second law efficiency standpoint, such that for optimized cases there is less refrigeration according to Fig. 2.5. Figure 2.8 shows that finally there is less rectification and thus fewer losses in the rectifier at high heat source temperatures. It has been shown that the cycle can be optimized for a range of heat source temperatures. Similarly, the cycle can be optimized for each heat sink or ambient temperature, and other parameters. This way the c ycle can be customized to the intended application for optimal performance. Irreversibility Analysis In realistic systems, there are irreversibilities associated with every component as with this ammonia water cycle. These irreversibilities will have negat ive effects on the performance of the cycle. The effects of each loss were studied individually and jointly on the cycle performance (Tamm et al., 2001). Typical working conditions used in this analysis were 400 K (260 F) and 30 bar (421 psig) at the boil er exit, 360 K (188 F) rectification, 410 K (278 F) at the turbine inlet, 280 K (44 F) and 2 bar (14.3 psig) in the absorber, and a basic solution ammonia mass fraction of 0.53. A typical turbine efficiency of 90% was assumed as suggested in the literat ure (Drbal et al., 1996). The thermal efficiency drops from 23.3% to 19.7%, a decrease of 15.4%. Due to the irreversibility in the turbine, although the pressure ratio is the same, the exhaust temperature of the turbine is higher. Less energy is converted into mechanical work in the turbine, and the turbine work output drops from 76.1 kW to 68.5 kW, a decrease of 10.0%. At the same time, a higher turbine exhaust temperature provides less cooling capacity. The cooling capacity drops 29.2% from 26.0 kW to 18 .4 kW.

PAGE 53

34 A pump efficiency of 80% was assumed as suggested in the literature (Drbal et al., 1996). The pump work requirement increases from 3.4 kW to 4.2 kW. This small increase causes the thermal efficiency to drop slightly. A pressure loss of 5% of the inl et pressure was assumed across the boiler as suggested in the literature (Bhatt et al., 1994). The results show this pressure loss has almost no negative effect on the cycle performance. Only slightly more pump work is required to boost the boiler inlet pr essure to compensate for the pressure loss in the boiler. A pressure loss of 5% was assumed for the superheater (Bhatt et al., 1994). The results show only a minor negative effect on the cycle performance. Thermal efficiency decreases by 3% of that in the ideal cycle, from 23.3% to 22.7%. Due to the pressure loss in the superheater, the turbine inlet pressure drops. Therefore, less expansion is possible producing 1.2% less work and higher exhaust temperatures. The cooling capacity decreases by 6.5%. A press ure loss of 5% was assumed for both streams in the recovery heat exchanger (Bhatt et al., 1994). The effects on the cycle performance are minimal, with a negligible decrease in thermal efficiency owing to an increase in the pump work requirement. A pressur e loss of 5% was assumed in the refrigeration heat exchanger, for comparison to other component pressure losses. The thermal efficiency drops by 2.7% from 23.3% to 22.8%, as the higher turbine exhaust pressure limits the expansion possible. The work output decreases by 1.6%. The reduced turbine pressure ratio also raises the exhaust temperature, reducing the cooling capacity by 4.6%. In a typical cooler,

PAGE 54

35 however, the heat exchanger experiences a 3% pressure loss (Bhatt et al., 1994), which is the value used in the combined irreversibility study. Finally, the overall effect of the irreversibility associated with the cycle was analyzed for combined losses. The thermal efficiency decreases from 23.3% under ideal conditions to 18.5%. The turbine work output drop s by 11.8%, from 76.1 kW to 67.1 kW. The cooling capacity decreases by 37.7%, from 26.0 kW to 16.2 kW. It can be seen that the greatest loss is attributed to the imperfect expansion in the turbine. Summary of Theoretical Studies The parametric analysis sho wed the potential for the cycle to be optimized. Optimization of the operating parameters is possible for each heat source and heat sink temperature, using a Generalized Reduced Gradient (GRG) method. The cycle may be optimized for the first law efficiency second law efficiency, power output or cooling output, depending on the intended application and the heat source. For a solar heat source, optimization for the second law efficiency is most appropriate, since the spent heat source fluid is recycled back to the solar collectors. It is found from simulation that optimization for second law efficiency produces no refrigeration at high heat source temperatures, while for low heat source temperatures it does. Inclusion of realistic losses in the analysis reduc es the cycle thermal efficiency by 20.6%, with 11.8% less work output and 37.7% less cooling capacity. The largest source of irreversibility in the cycle is the imperfect turbine expansion.

PAGE 55

3 6 CHAPTER 3 EXPERIMENTAL SYSTEM An experimental system has been built to demonstrate the feasibility of the ammonia based combined power and cooling cycle. Operation of a working system gives practical experience and a means to improve the basic cycle desig n, in order to advance the concept towards industry. Figure 3.1. Schematic of the experimental system concept. In the current study, the shown turbine and refrigeration processes are simulated. The conceptual design of the experimental system is shown i n Figure 3.1. The experimental study is being approached in two phases. In Phase 1 the vapor generation and absorption processes are studied, while a heat exchanger and throttling valve model

PAGE 56

37 the expansion process until a turbine is installed in Phase 2 to form a complete power cycle. In comparison to the initial cycle concept of Fig. 2.1, the rectifier and consequently the superheater have been left out for simplicity as purification of the high concentration ammonia water vapor is not critical without a t urbine in place. The refrigeration unit is also simulated by the turbine heat exchanger. Phase 1 is presented in this dissertation, and is used in preparation for later Phase 2 studies. Over the course of the experimental work, an initial system was built. An experimental study revealed where modifications could be made, yielding an improved system that was also studied. Changes made in the system are noted below, and the results are compared in Chapter 5. Ammonia Water Side This main part of the system is shown in detail by the schematic in Fig. 3.2, with photographs in Figs. 3.3 and 3.4. A list of components is given in Appendix C, and experimental operating procedures are given in Appendix D. In the initial system, the strong solution of ammonia and water was pumped from the absorber with a rotary vane positive displacement pump capable of producing the high boiler pressures of interest. A diaphragm pump, shown in Fig. 3.5, is used in the improved system as the rotary vane pump was found to cavitate and fa iled after prolonged operation. Also, the diaphragm pump flow rate is not limited by the downstream pressure, in the range of operating pressures of interest. A strainer is at the pump inlet to prevent loose rust pieces from damaging the pump. The collecte d particles can be purged from the system through a valve. An expansion tank is positioned after the pump to dampen the pulses in the flow rate. Leaving the pump, the strong solution passes through a nickel brazed, stainless steel, vertically stacked plate heat exchanger where it

PAGE 57

38 recovers heat from the weak solution returning to the absorber. The heat exchange area in the recovery unit was increased by 286% for the improved system. The system heat exchangers are shown in Fig. 3.6. Figure 3.2. Schematic o f the experimental system components for the system side.

PAGE 58

39 Figure 3.3. Photograph of the system side and hot water side of the experiment. Figure 3.4. Photograph of the system side of the experiment.

PAGE 59

40 Figure 3.5. Photograph of the system side pump a nd expansion tank. Figure 3.6. Photograph of the system side heat exchangers.

PAGE 60

41 The boiler is also a vertically stacked plate heat exchanger, using resistance heated hot water to partially boil the strong solution which flows into a simple carbon steel ta nk where gravity separates out liquid from vapor. The heat exchange area in the boiler was also increased for the improved system, by 47%. The relatively low concentration weak solution exits the bottom while high concentration vapor leaves through the top of the separator. The weak solution flows through the solution heat exchanger to yield energy to the strong solution. After a throttling valve to reduce pressure, the weak solution enters at the top of the absorber through a spray nozzle across the coolin g elements inside, as shown in Fig. 3.7. The vapor is throttled and passes through a stacked plate heat exchanger against coolant flow. This expansion lowers the pressure and temperature of the vapor before it enters the carbon steel absorber through 18 ho les of 0.041 diameter distributed along a horizontal S shaped tube section, intended to bubble the vapor into the pool for better absorption. A storage tank is in place to contain the working fluid while the system is undergoing modifications or repairs. The storage tank is also used when the system is initially charged, and can be used during distillation of the working fluid to alter its concentration. A flow indicator, shown in Fig. 3.8, connects the storage tank with the system in order to monitor whet her liquid or vapor is flowing. Stainless steel tubing connects the components, with Swagelok pressure fittings and valves where needed. Care was taken to minimize flow losses in the system. All the tubing is insulated, as are the absorber and separator t anks. Additional insulation was added to flow meters and fittings in the improved design. Preventing heat losses in vapor lines discourages condensation on the tubing walls.

PAGE 61

42 Figure 3.7. Schematic of the absorber design. Material compatibility is a serio us concern with ammonia solutions. All components are selected to have no reaction with the working fluid. The tubing and fittings are made of stainless steel and the tanks are of mild steel. The absorber evaporator units are made of aluminum. Other heat e xchangers are with nickel brazed stainless steel. O rings in the valves are made of Kalrez, and are coated with Krytox lubricant. Gaskets, packing and septa elsewhere are suitably of neoprene, Teflon, EPDM or other compatible material.

PAGE 62

43 Figure 3.8. Ph otograph of the flow indicator to the storage tank. Figure 3.9. Photograph of the hot water side of the experiment.

PAGE 63

44 Hot Water Side The hot water side is shown in Fig. 3.9, with a schematic in Fig. 3.10. A 5 kW single pass boiler supplies hot water to a hot water storage tank, to build up ample high temperature storage. This way the system can provide more than 5 kW heat input, for a finite period of operation. The water is pressurized for the storage temperature to exceed 100 C (212 F), and can go as h igh as 150 C (302 F). A circulating pump sends the hot water to the plate heat exchanger used as a boiler of the ammonia water mixture, then back to the single pass boiler. Lines are set up which can divert the hot water from the plate heat exchanger to an advanced boiling apparatus of the ammonia water mixture, still in the design stage, that will allow a more detailed study of the boiling process in the cycle in future work. Hot water from this boiling apparatus will also be returned to the hot water s torage tank. The hot water temperature is controlled with a mixing valve, which mixes hot water drawn from the storage tank with some of the relatively cooler water returning from the system boiler. The valve regulates the flow from each of these lines to maintain a steady outlet temperature. The single pass boiler, storage tank, and associated tubing are well insulated. Coolant Side The coolant side is shown by the schematic in Figure 3.11, with a photo shown in Fig. 3.12. The thermostatically controlled chiller maintains a constant coolant temperature, within 3 C ( 5 F), in the chilled water storage tank. It is capable of producing coolant temperatures of 30 C ( 22 F). The coolant used is a 50/50 mixture of ethylene glycol and water, which has a f reezing point that is low enough for the

PAGE 64

45 application and liquid viscosity at such low temperatures that can be handled by conventional circulation pumps. The chiller serves to provide a range of coolant temperatures for simulating ambient sink conditions f or the absorption process. Figure 3.10. Schematic of the experimental system components for the hot water side.

PAGE 65

46 Figure 3.11. Schematic of the experimental system components for the coolant side. From the chilled water storage tank, the coolant is pum ped to the absorber and passes through a finned coil heat exchanger in the absorber. The finned coil heat exchanger consists of several aluminum car evaporators connected in parallel to minimize coolant flow losses. The fins on the evaporators enhance the heat transfer between the cooling liquid and ammonia solution. They also act as baffles to increase the residence time of the weak solution before it gets to the bottom of the tank so that the

PAGE 66

47 ammonia vapor and weak solution ha ve enough time to be mixed. A second coolant line provides the cooling for the heat exchanger in the ammonia vapor line to simulate the turbine. The storage tank and piping on the coolant side are well insulated. Figure 3.12. Photograph of the coolant side of the experiment. Instru mentation Pressures (P), temperatures (T), ammonia mass fractions (x) and flow rates (m) are recorded at locations shown in Figure 3.13. Properties at other state points can be extrapolated with minimal assumptions for energy transfers to be calculated. Ch anges in potential and kinetic energies are neglected. Some of the measurements are redundant and are used to compare calculated results from two sets of independent measurements. Properties of the hot water and coolant can be used to determine energy tran sfers and to verify those found from the ammonia water side.

PAGE 67

48 Figure 3.13. Location of data measurements for system analysis. Thermocouples In total, there are 19 temperatures monitored by T type thermocouples of which 14 are connected to the DaqBook da ta acquisition system to record real time and time averaged values with the DaqView software. The 14 are located as shown in Fig 3.13. Note that in the initial system, the liquid and vapor exit temperatures from the separator were not measured, such that only 12 thermocouple values were then recorded. Multiple

PAGE 68

49 channel temperature displays are near the experiment to allow quick readouts. A two point calibration was used for each of the thermocouples, as outlined in Appendix D. The error in the measurement i s discussed in Appendix E. Figure 3.14. Photographs of pressure measurement instruments on the system. Pressure Transducers Pressure transducers at 5 locations record data directly through the data acquisition system and compliment the analog gauge displays on the system. Transducer locations are noted in Fig 3.13. The transducers were calibrated using a one point method as outlined in Appendix D, with the uncertainty discussed in Appendix E. Gauge readings were used in place of certain transducers w hile they were being repaired, as listed according to the measurement device history in Table E.4 of Appendix E. Gas Chromatograph and Syringe Sampling Samples are taken by syringe through septum ports in the strong, weak and vapor lines. A sampling port o n the absorber is shown in Fig. 3.15. The samples are transported to a gas chromatograph (GC) in the lab by a locked syringe, and then analyzed using a thermal conductivity detector (TCD) in the GC. The GC setup is shown in Fig. 3.16.

PAGE 69

50 Procedures for syring e sampling, using the GC, and calibration are given in Appendix D. Uncertainty discussion is given in Appendix E. Figure 3.15. Photograph of a syringe sampling port and sight glass on the absorber. Figure 3.16. Photograph of the gas chromatograph set up.

PAGE 70

51 Flow Meters Variable area float type flow meters are in place to measure strong and weak solution, vapor, hot water and coolant flow rates. A secondary spring/piston type flow meter is connected to the vapor coolant for trials with higher flow rates. T he liquid flow meter readings were verified but not calibrated. Corrections for the measured flow rates are given in Appendix D. The uncertainty is discussed in Appendix E. Figure 3.17. Photograph of the temperature displays and coolant flow meter. Data Acquisition Hardware and Analysis Software DaqBook was used as the data acquisition electronic interface for the pressure and temperature measuring devices. Expansion cards are connected directly to the transducer and thermocouple wires from the system. More about the DaqBook components of the data acquisition system is presented in Appendix D. The readings provided by the DaqBook system are shown in real time and sent to file by the DaqView software. To process the raw binary or text files provided by the acquisition system, a data analysis program was written to extract and compile the

PAGE 71

52 multiple readings into more useful time averaged data. The key elements of the computer code are given in Appendix F. The program applies assumptions as needed to yield the pressure and temperature at each state point. With manual entry of concentration measurements from the GC and flow rates from the flow meters, the program accesses a routine similar to the one used in the theoretical studies to determine properties at each state point, and furthermore provides key results such as cycle efficiencies, work output and cooling capacity. The experimental results are then compared to a simulated cycle with the program. The program also allows for the design of experiments, e xtrapolating desired system operating conditions consisting of thermocouple, transducer and flow meter readings from a theoretical starting point consisting of heat source temperature, ambient sink temperature, boiler high pressure, absorber low pressure a nd a ratio of the heat source and ammonia water flow rates. Component parameters such as heat exchanger effectiveness, pump and turbine efficiencies, and other losses can be included and modified according to the real results of the experiment, such that t he design program evolves with the experimentation. The goal of the design program is to predict actual system parameters efficiently to minimize the experimentation time whenever new design conditions are to be tested. Safety The safety measures needed fo r system operation are extensive as ammonia is toxic and corrosive. The components have been mounted on the wall outside the lab building, allowing adequate ventilation in case of leaks. Pressure relief valves are included to prevent critical pressure buil dup in the system, which might otherwise rupture a component. Upon cracking, the relief valves discharge through tubing into a large

PAGE 72

53 emergency storage tank, which can be safely emptied at a later time. The emergency tank is shown in Fig. 3.18. However if a rupture does occur, overhead water sprinklers can immediately dilute the surrounding air to safe concentrations of ammonia. Figure 3.18. Photograph of the storage tank (left) and emergency tank. An emergency power cutoff switch can cut power to both t he system side and hot water pumps. A gas mask fitted with ammonia filters is available for emergency use. An eyewash station and shower are also located adjacent to the experiment. For normal system operation, however, only chemical spill resistant goggle s are necessary. Lab coats and face shields are recommended for certain maintenance operations. Emergency procedures are outline in Appendix D.

PAGE 73

54 CHAPTER 4 EXPERIMENTAL METHODO LOGY Strategies for the design of experiments, system operation and results analysis methods are discussed in this chapter to best match the limits and behavior of the system and to yield the results of interest. Parameters i n Simulations and Experiments In simulating the modified cycle that the experimental system is based on, the design parameters can be classified as either fluid or component parameters. The fluid parameters can sufficiently describe the cycle for the speci al case of ideal components. For realistic simulation, the complexity of the design increases as the components are given non ideal characteristics. Fig. 4.1 shows the fluid parameters used in the simulation of the cycle. The mass flow rates of the working fluid and hot water are extensive properties and can be combined into an intensive heat source flow ratio parameter, thus leaving 5 fluid parameters to simulate the ideal cycle. For the inclusion of realistic losses, 12 component parameters are used in th e simulation as shown in Table 4.1. The parameters include pump and turbine efficiency, pressure losses, effectiveness for heat transfers between two fluid streams for which the heat exchange process is specific, and approach temperature limits for heat tr ansfers with the ambient for which the heat exchange process is not yet specific. Typical values for the losses are also given in Table 4.1. For the experimental system trials, data is taken only at critical locations and certain component parameters are e stimated to complete the set of properties at each

PAGE 74

55 location. Minimal error is expected from these assumptions. The component parameters used in conjunction with the experimental data are shown in Table 4.1. Several pressure losses are estimated as pressure s are only measured at limited locations in the system. The pump efficiency is estimated since the immediate pump exit pressure is not recorded in each trial. The other component parameters of Table 4.1 that are assigned in the experimental trials are nece ssary to model the turbine and ambient conditions. Figure 4.1. Schematic of the cycle concept used in experiment simulations, showing the main fluid parameters.

PAGE 75

56 Table 4.1. Fluid and component parameters in the simulation and experiment. Simulation Ex periment Fluid Parameters heat source temperature Set Measured boiler pressure Set Measured ambient temperature Set Set (absorber T is measured) basic solution ammonia mass fraction Set Measured heat source flow ratio Set Measured Componen t Parameters ideal/typical approach T ambient and absorber 0 / 5 C Set Set approach T ambient and refrig. HE 0 / 5 C Set Set effectiveness recovery heat exchanger 100 / 85 % Set Measured effectiveness boiler heat exchanger 100 / 8 5 % Set Measured isentropic efficiency turbine 100 / 90 % Set Set isentropic efficiency pump 100 / 80 % Set Set pressure loss vapor inlet to absorber 0 / 5 % Set Measured pressure loss weak inlet to absorber 0 / 5 % Set Measured pres sure loss recovery HE strong side 0 / 5 % Set Set pressure loss recovery HE weak side 0 / 5 % Set Set pressure loss boiler HE system side 0 / 5 % Set Set pressure loss refrig HE system side 0 / 5 % Set Set Limits and Selection of Operatin g Conditions The operating conditions of the investigations were discretely selected to demonstrate the vapor generation and absorption processes, and to determine the power production and refrigeration potential of the system in comparison to simulations. Contour plots such as Figures 4.2 and 4.3 show this potential for a typical cycle design at two different ambient temperatures, and are used to determine the necessary operating conditions. These simulations, shown for the ideal cycle with no losses here, allow the anticipation of certain results and system behavior prior to experimentation.

PAGE 76

57 1 0 0 1 0 0 2 0 0 2 0 0 3 0 0 3 0 0 4 0 0 4 0 0 5 0 0 5 0 0 6 0 0 6 0 0 7 0 0 8 0 0 9 0 0 1 0 0 0 h e a t s o u r c e t e m p e r a t u r e ( C ) b o i l e r p r e s s u r e ( b a r ) 5 0 6 0 7 0 8 0 9 0 1 0 0 1 1 0 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 a m b i e n t t e m p e r a t u r e = 5 C b a s i c s o l u t i o n m a s s f r a c t i o n = 4 0 % h e a t s o u r c e f l o w r a t i o = 3 5 c y c l e w i t h n o l o s s e s 1 0 0 1 0 0 2 0 0 2 0 0 3 0 0 3 0 0 4 0 0 5 0 0 6 0 0 h e a t s o u r c e t e m p e r a t u r e ( C ) b o i l e r p r e s s u r e ( b a r ) 5 0 6 0 7 0 8 0 9 0 1 0 0 1 1 0 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 a m b i e n t t e m p e r a t u r e = 2 5 C b a s i c s o l u t i o n m a s s f r a c t i o n = 4 0 % h e a t s o u r c e f l o w r a t i o = 3 5 c y c l e w i t h n o l o s s e s Figure 4.2. Expected turbine work output (kJ/kg) per unit basic solution flow, based on simulation with no losses.

PAGE 77

58 0 3 0 3 0 3 0 3 1 5 1 5 1 5 1 5 2 7 2 7 2 7 2 7 3 9 3 9 3 9 3 9 5 1 5 1 h e a t s o u r c e t e m p e r a t u r e ( C ) b o i l e r p r e s s u r e ( b a r ) 5 0 6 0 7 0 8 0 9 0 1 0 0 1 1 0 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 a m b i e n t t e m p e r a t u r e = 5 C b a s i c s o l u t i o n m a s s f r a c t i o n = 4 0 % h e a t s o u r c e f l o w r a t i o = 3 5 c y c l e w i t h n o l o s s e s 0 3 0 3 0 3 0 3 0 9 0 9 0 9 0 9 1 5 1 5 1 5 2 1 h e a t s o u r c e t e m p e r a t u r e ( C ) b o i l e r p r e s s u r e ( b a r ) 5 0 6 0 7 0 8 0 9 0 1 0 0 1 1 0 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 a m b i e n t t e m p e r a t u r e = 2 5 C b a s i c s o l u t i o n m a s s f r a c t i o n = 4 0 % h e a t s o u r c e f l o w r a t i o = 3 5 c y c l e w i t h n o l o s s e s Figure 4.3. Expected refrigeration output (kJ/ kg) per unit basic solution flow, based on simulation with no losses.

PAGE 78

59 The plots above are 3 dimensional, but are viewed in two dimensions. In other words, each contour represents the locus of all work outputs of the same magnitude in Fig. 4.2, and cooling capacities of the same magnitude in Fig. 4.3. Prior to experimentation, the limits of the system were evaluated, to which the desired operating conditions must conform. The feasible range of operating conditions is dictated by experimental and practical li mits, and is more stringent than in simulations. The limits of operating conditions are given in Table 4.2. Limits of system components are listed in Appendix C, which cannot be exceeded by the operating conditions. Table 4.2. Limits of primary operating conditions. Lower limit Upper limit heat source temperature 50 C (122 F) 110 C (230 F) boiler pressure 3.8 bar (40 psig) 14.8 bar (200 psig) ambient temperature 5 C (41 F) 35 C (95 F) basic solution NH 3 mass fraction 30% 50% heat source flow ratio 0.5 8 Heat Source Temperature The heat source temperature was varied over a sufficiently large range to demonstrate trends in system performance and to compare them to simulations. Since the application can be used for geothermal sources, a practic al upper temperature limit is typically 140 C (284 F) for the boiler. However the experimental limit is taken as 110 C (230 F), as for low temperature solar sources, so not to tax the hot water storage and boiler beyond capacity. For low heat source te mperatures, vaporization is possible at low pressures. However, low pressures limit the expansion capability of the system so the heat source temperature lower limit was set at 50 C (122 F).

PAGE 79

60 The heat source temperature must be high enough to partially bo il the strong solution at the given pressure and ammonia mass fraction, in order to generate vapor for power and refrigeration production. However if too high, there will be no liquid leaving the separator and the cycle will resemble a binary fluid Rankine cycle. Limits of vapor generation for a typical solution concentration and absorber temperature are shown in Fig. 4.4, which gives the vapor fraction of the mixture leaving the boiler for various temperatures and pressures. The vapor fraction, or quality, is defined here as the ratio of the mass flow rate of vapor leaving the separator to the mass flow rate of the basic solution entering the boiler. 2 0 2 0 5 0 5 0 8 0 8 0 1 1 0 1 1 0 1 4 0 1 4 0 1 7 0 1 7 0 2 0 0 2 0 0 2 3 0 h e a t s o u r c e t e m p e r a t u r e ( C ) b o i l e r p r e s s u r e ( b a r ) 5 0 6 0 7 0 8 0 9 0 1 0 0 1 1 0 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 a m b i e n t t e m p e r a t u r e = 2 5 C b a s i c s o l u t i o n m a s s f r a c t i o n = 4 0 % h e a t s o u r c e f l o w r a t i o = 3 5 c y c l e w i t h n o l o s s e s Figure 4.4. Vapor fraction (%) of ammonia water fluid leaving the boiler, based on simulation with no los ses. The corresponding specific heat input is shown in Fig. 4.5. Knowing the experimental limitations for heat input, Fig. 4.5 shows whether the design is feasible a priori. With this system, 5 kW can be continuously supplied by the hot water heater. Any

PAGE 80

61 h igher heat input is finite and depends on the heat stored in the hot water storage tank. Similarly, Fig. 4.6 is used to remain within absorber heat rejection limits set by the coolant serving as the heat sink. 5 0 0 5 0 0 1 5 0 0 1 5 0 0 2 5 0 0 2 5 0 0 3 5 0 0 4 5 0 0 5 5 0 0 h e a t s o u r c e t e m p e r a t u r e ( C ) b o i l e r p r e s s u r e ( b a r ) 5 0 6 0 7 0 8 0 9 0 1 0 0 1 1 0 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 a m b i e n t t e m p e r a t u r e = 2 5 C b a s i c s o l u t i o n m a s s f r a c t i o n = 4 0 % h e a t s o u r c e f l o w r a t i o = 3 5 c y c l e w i t h n o l o s s e s Figure 4.5. Boiler heat input (kJ/kg) per u nit basic solution flow, based on simulation with no losses. Boiler Pressure The system performance is sensitive to the boiler pressure, and so it was varied over a large range to show diverse results. The boiler pressure must be such as to allow partial b oiling at the given boiler temperature and basic solution ammonia mass fraction. The upper limit is set at the separator pressure rating of 14.8 bar (200 psig). The lower limit is arbitrary, provided it is higher than the saturation pressure of the absorbe r. At low boiler pressures, not much expansion would be possible through the turbine and

PAGE 81

62 refrigeration would not be seen. The lower limit is thus taken as 3.8 bar (40 psig), but was higher in most studies. 4 5 0 0 3 5 0 0 3 5 0 2 5 0 0 2 5 0 0 1 5 0 0 1 5 0 0 5 0 0 5 0 0 h e a t s o u r c e t e m p e r a t u r e ( C ) b o i l e r p r e s s u r e ( b a r ) 5 0 6 0 7 0 8 0 9 0 1 0 0 1 1 0 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 a m b i e n t t e m p e r a t u r e = 2 5 C b a s i c s o l u t i o n m a s s f r a c t i o n = 4 0 % h e a t s o u r c e f l o w r a t i o = 3 5 c y c l e w i t h n o l o s s e s Figure 4.6. Expected absorber heat rejection (k J/kg) per unit basic solution flow, based on simulation with no losses. Ambient Temperature The ambient temperature is a key design parameter, although it was simulated through the measured absorber temperature and the estimated approach limit. The absorbe r temperature affects the absorber pressure as the absorber operates near saturation, and thus the expansion ratio that is available for the turbine is affected. The boiling process is not expected to depend highly on absorber temperature and so only a few ambient temperatures were studied. The limits of absorber temperatures reflect the ambient temperature limits, which were taken for a range of climates and seasons.

PAGE 82

63 Basic Solution Ammonia Mass Fraction The basic solution ammonia mass fraction may fully ra nge between 0 and 100% ammonia by mass, provided the mixture can be partially evaporated at the given boiler pressure and temperature. The strong solution should leave the absorber as liquid, and the saturation pressure must remain below the upper limit of the absorber for a given ambient temperature. For simulating the ambient temperatures of interest, a 30 50% ammonia mass fraction was selected as it gives a range of absorber saturation pressures most compatible with the current experimental system. It is desired to remain lower than the 3.1 bar (30 psig) absorber pressure rating and above 1 bar (0 psig) to maintain positive pressure as a rule of thumb. Usually the lower limit is higher, according to the saturation pressure of the absorber. Heat Source Flo w Ratio The limits on the heat source flow ratio arise mostly from limits in pumping capacity and flow meter useful range. The hot water pump produces less than 5.7 lpm (1.5 gpm), and the system pump gets around 1.9 lpm (0.5 gpm) depending on the absorber pressure. A typical heat source flow ratio is around 3.5, and the system was operated near this value usually. The heat source flow ratio was not varied deliberately. System Behavior The system reaches a pseudo steady state typically in 20 30 minutes betwe en trials, and longer when first started or if the experiment borders on component limits such as pump flow rate. As pressures, temperatures and flow rates are closely tied, a slow fluctuation over several minutes of one property causes a predictable, slow fluctuation in other properties. This fluctuation is generally self correcting after the system has been running for a while, but can be sped up by operator intervention. This pseudo steady

PAGE 83

64 fluctuation is taken as acceptable, but is an inconvenience and s hould be reduced with the implementation of control methods. Steady state during operation is determined by monitoring the above system parameters. The indirect source of deviation from steadiness arises from a poor absorber design, in which the cooling oc curs above the liquid pool. The absorber pressure is sensitive to the coolant temperature and flow rate, dropping the absorber to very near its saturation pressure. The absorber should operate slightly above its saturation pressure, so that the pump does n ot cavitate and is able to draw a sufficient flow rate. Helium has been added to the absorber during operation to raise the absolute pressure about 10% above saturation, thus compressing the liquid leaving the absorber. The pump flow rate improves signific antly if pumping liquid that is compressed, and does not work at all if the liquid is near saturation. At times when the pump cavitates to the point of producing no flow rate, either the pressure in the absorber can slowly rise by regulating the coolant fl ow or the absorber can be shocked with excess vapor from the separator. A revision in the absorber design to have the cooling elements in the liquid pool would slightly subcool the fluid steadily, with the same effect to improve the pump flow rate. Elevati ng the absorber would also generate head pressure to compress the liquid at the pump inlet. Another strategy to maintain the absorber at slightly higher than saturation pressure is to permit slow pooling of the liquid solution in the separator. The reduced weak solution return to the absorber slows absorption and raises the absorber pressure. Liquid level indicators on the absorber verify the imbalance in flow rates. After an hour or two, the weak solution can be returned before the absorber liquid level dr ops below the

PAGE 84

65 vapor inlet. If the level is below the vapor inlet, the absorber pressure becomes very sensitive to weak solution flow rate and may rise quickly to above vessel limits. Therefore, the pump flow rate is the direct source of deviation from stea diness. A change in the pump flow rate will change the temperature of the strong solution exiting the boiler, and how much vapor has been formed. The effects compound and the system is shifted from the desired operating conditions. Usually control efforts are localized to the absorber to maintain steady operation. A general control is to vary the coolant flow rate according to the coolant inlet temperature. For changing operating parameters and for other areas of control, Table 4.3 lists the corresponding control measures. These should be used in conjunction with the operating procedures of Appendix D. Table 4.3. Experimental control measures for maintaining operating parameters. System parameter Controlled primarily by boiler pressure vapor flow th rough throttling valve and basic solution flow rate heat source temperature storage tank temperature and mixing valve setting absorber temperature coolant temperature and flow rate to absorber absorber pressure basic solution ammonia mass fraction, coo lant temperature and flow rate, vapor and weak solution flow rates basic solution mass fraction system charging before operation basic solution flow rate absorber pressure, diverting flow from basic solution line back to absorber vapor inlet temperature t o absorber coolant temperature and flow rate to vapor H.E. fluid pooling in separator weak solution flow rate leaving separator Uncertainty of Measurements Data acquisition in the ammonia water side during operation is limited to the pressures, temperat ures, ammonia mass fractions and flow rates taken at locations in the system shown in Fig. 3.12. Properties at all other state points can be extrapolated with

PAGE 85

66 minimal and reasonable assumptions, and estimation of component parameters in Table 4.1. The loss es are expected to reduce the performance of the simulated cycle. Fig. 4.7 shows the first law efficiency for a cycle simulated as ideal and with the typical losses of Table 4.1. Fig. 4.8 shows a similar plot for the second law efficiency. Note that the a ssumption of an ideal heat exchanger causes instabilities in the computations near the lower limit of vapor generation, and rough contours arise in the simulation of the cycle with no losses. Measurement errors are compounded in calculated energy transfers efficiencies and other key results according to the uncertainty analysis discussion in Appendix E. In addition to the error attributed to measurements, the phase of the ammonia water mixture is not well known at each location, which contributes to error in the property evaluation. As property evaluation of liquid is of greater accuracy than that of two phase ammonia water mixtures, data is acquired redundantly in locations where the phase is more certain. The required measurements to obtain key results fr om the ammonia water side of the experiment are reduced, as measurement from the hot water side is also used. The coolant side data is not used, as the response time of the system to coolant flow rate is very slow and thus the coolant flow rate may be in e rror. The coolant flow meters also exhibit a worse rated accuracy than the hot water and ammonia water flow meters. The uncertainty of the key results should not be greater than 5%. In order to minimize the experimental error, the instrumentation should be calibrated carefully, repeated measurements should be taken, and the measurements should be taken only when the system is deemed sufficiently steady.

PAGE 86

67 1 0 1 0 2 0 2 0 3 0 3 0 4 0 4 0 5 0 5 0 6 0 6 0 7 0 7 0 7 0 8 0 8 0 9 0 9 0 1 0 0 1 0 0 1 1 0 1 1 0 1 2 0 1 2 0 1 3 0 1 4 0 1 5 0 h e a t s o u r c e t e m p e r a t u r e ( C ) b o i l e r p r e s s u r e ( b a r ) 5 0 6 0 7 0 8 0 9 0 1 0 0 1 1 0 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 a m b i e n t t e m p e r a t u r e = 2 5 C b a s i c s o l u t i o n m a s s f r a c t i o n = 4 0 % h e a t s o u r c e f l o w r a t i o = 3 5 c y c l e w i t h n o l o s s e s 1 0 1 0 2 0 2 0 3 0 3 0 4 0 4 0 5 0 5 0 6 0 6 0 7 0 8 0 8 0 9 0 h e a t s o u r c e t e m p e r a t u r e ( C ) b o i l e r p r e s s u r e ( b a r ) 5 0 6 0 7 0 8 0 9 0 1 0 0 1 1 0 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 a m b i e n t t e m p e r a t u r e = 2 5 C b a s i c s o l u t i o n m a s s f r a c t i o n = 4 0 % h e a t s o u r c e f l o w r a t i o = 3 5 c y c l e w i t h t y p i c a l l o s s e s Figure 4.7. Expected first law efficiency (%), based on simulation with no losses and with typical los ses.

PAGE 87

68 1 0 1 0 1 0 1 0 1 7 7 1 7 7 2 6 0 2 6 0 3 4 3 3 4 3 4 2 7 4 2 7 5 1 0 5 1 0 5 1 0 5 9 3 5 9 3 5 9 3 6 7 7 6 7 7 6 7 7 6 7 7 7 6 0 7 6 0 7 6 0 h e a t s o u r c e t e m p e r a t u r e ( C ) b o i l e r p r e s s u r e ( b a r ) 5 0 6 0 7 0 8 0 9 0 1 0 0 1 1 0 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 a m b i e n t t e m p e r a t u r e = 2 5 C b a s i c s o l u t i o n m a s s f r a c t i o n = 4 0 % h e a t s o u r c e f l o w r a t i o = 3 5 c y c l e w i t h n o l o s s e s 1 0 1 0 9 3 9 3 1 7 7 1 7 7 2 6 0 2 6 0 3 4 3 3 4 3 3 4 3 4 2 7 4 2 7 h e a t s o u r c e t e m p e r a t u r e ( C ) b o i l e r p r e s s u r e ( b a r ) 5 0 6 0 7 0 8 0 9 0 1 0 0 1 1 0 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 a m b i e n t t e m p e r a t u r e = 2 5 C b a s i c s o l u t i o n m a s s f r a c t i o n = 4 0 % h e a t s o u r c e f l o w r a t i o = 3 5 c y c l e w i t h t y p i c a l l o s s e s Figure 4.8. Expected second law efficiency (%), based on simulation with no losses and with typical losses.

PAGE 88

69 CHAPTER 5 EXPERIMENTAL RESULTS The results show that essential processes of the proposed cycle can be reproduced experimentally. Generation of vapor at high pressures and its expansion back to low pressures verify the potential to produce work from heat. Modeling the turbine and refrigeration unit with an equivalent expansion and temperature change, experiments support the potential for refrigeration at certain operating conditions. Absorption of the vapor back into the basic solution confirms that the cyc le can regenerate itself. As the turbine is simulated by an equivalent expansion, the calculated results that follow are grouped into those that are independent of the expansion process and those that depend on the models used for the turbine and refrigera tion unit. To best convey the key findings of each section, a qualitative interpretation of the results and explanation of relevant phenomena precedes the presentation of the results. Generally the expected trends of simulating the cycle are evident in the experimental results, although quantitative comparisons to simulated results show inconsistencies. Vapor Generation Experiments suggest that there are non equilibrium phase change processes during the generation of vapor from the basic solution. Many of t he results can be attributed to a difference between the wall temperature and the fluid bulk temperature during conditions at or near saturation. The phase change at the wall will not be in proportion to the phase change in the bulk fluid, which is assumed to exist in equilibrium

PAGE 89

70 at the given temperature, pressure and ammonia mass fraction. The conditions near the wall are very sensitive to the wall temperature. Observations The vaporization of the strong solution may initiate in the recovery unit and conti nues in the system boiler. In the boiler and recovery heat exchanger, the wall temperature is higher than the bulk temperature of the strong solution. This causes vapor production of lower ammonia mass fractions as higher temperature boiling can vaporize p roportionally more water, which has a higher boiling point than ammonia. Figure 5.1. Qualitative representation of the change in properties during the boiling of the strong solution. Property measurement locations are shown. However, away from the heat source, the vapor bubbles condense into the relatively cooler bulk fluid and on the cooler tube walls upon leaving the heater section. This absorption lowers the vapor fraction of the fluid, and raises the bulk temperature. Again due to the boiling point o f water, more water will condense than ammonia such

PAGE 90

71 that the remaining vapor will be of a higher ammonia mass fraction and the liquid part will be of a lower ammonia mass fraction. Equilibrium will eventually be reached far downstream from the boiler exit. The bulk temperature when measured too close to the boiler exit, as shown in Fig. 5.1, will not necessarily be the equilibrium temperature. Upon entering the separator, the vapor is extracted from the liquid, with the two streams exiting at saturation in principle. Even with adequate insulation there are external losses, and the vapor continues to condense on the separator vessel walls. This raises the vapor temperature and the ammonia mass fraction in the vapor region. The temperature of the vapor leaving the separator was found to be 0.74 0.16 C higher on average than the temperature of the weak solution exiting the tank. The vapor fraction, defined as the ratio of the mass flow rate of vapor to the mass flow rate of the two phase inlet stream, is also reduced. This reduces the potential for work and cooling output, as the vapor flow rate is reduced. Although water condenses more readily than ammonia from the vapor stream, the condensate is still very high in ammonia concentration. This raises the ammon ia mass fraction in the weak solution as well, which pools in the separator. The weak solution temperature at the separator exit was essentially the same as the bulk fluid temperature at the boiler exit. The incomplete liquid and vapor separation also cont ributes to the greater ammonia content in the weak solution than expected. The results confirm that the weak solution is supersaturated with ammonia as it leaves the separator. Recovered Heat The internal recovery is determined from the heat transfer acros s the recovery heat exchanger as measured from the weak solution side of the unit, since the weak

PAGE 91

72 solution is not expected to undergo a phase change. The certainty in the determination of the phase leads to greater certainty in the properties at that state The maximum possible heat transfer is the energy change of the lower heat capacity fluid in having a maximum temperature change. The lower heat capacity fluid is the weak solution, as it has a lower mass flow rate and should not incur a phase change. The maximum heat it can reject is from cooling to the strong solution recovery inlet temperature. For a real heat exchanger of finite surface area, the maximum heat exchange will not occur and the heat exchanger effectiveness is used to gauge the performance of the unit. 7 6 3 8 8 8 8 8 8 1 0 1 4 1 0 1 4 1 1 3 9 1 1 3 9 1 2 6 5 1 2 6 5 1 3 9 0 1 3 9 0 1 5 1 6 1 5 1 6 1 6 4 2 1 6 4 2 1 7 6 7 1 7 6 7 1 8 9 3 2 0 1 8 2 0 1 8 2 1 4 4 2 5 2 0 h e a t s o u r c e t e m p e r a t u r e ( C ) b o i l e r p r e s s u r e ( b a r ) 5 0 6 0 7 0 8 0 9 0 1 0 0 1 1 0 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 a m b i e n t t e m p e r a t u r e = 2 5 C b a s i c s o l u t i o n m a s s f r a c t i o n = 4 0 % h e a t s o u r c e f l o w r a t i o = 3 5 c y c l e w i t h t y p i c a l l o s s e s Figure 5.2. Expected internal heat recovery (kJ/kg) per unit basic solution flow, based on simulation with typical losses. Simulated results are shown in Fig. 5.2, where the typical losses are those listed in Chapter 4. In particular, 85% e ffectiveness was used for the recovery heat exchanger. In further studies, correlation of the effectiveness with other parameters could yield a better

PAGE 92

73 model. Results suggest that there is more recovery when the weak solution is hotter, as for higher boiler temperatures. There is also more recovery for higher pressures, since there is more weak solution return from the separator as less strong solution is vaporized in the boiler. The upper left region is where the conditions do not allow boiling, and only li quid exits from the separator. The observed behavior follows the expected trends. It is seen from the experimental results that for trials in which boiling initiates in the recovery unit, the effectiveness of the heat exchange improves. This occurs if the system operates at pressures well below the dew point pressure. This low pressure allows the wall temperature of the recovery unit to be high enough to initiate boiling. The initial system averaged 98.1 7.0% effectiveness for when the boiling initiated i n the recovery unit, and otherwise 75.4 8.2% for when boiling did not initiate until the boiler. There was no boiling in the recovery unit for the improved system trials, as higher boiler pressures were evaluated. But with the addition of 286% more heat exchange area, the recovery effectiveness for trials without boiling initiation increased to 90.9 0.9%. Boiler Heat Input The boiler heat input is expected to increase with an increase in vapor production, which occurs at higher boiler temperatures and lower pressures for a given strong solution ammonia mass fraction. The combined experimental results for the initial and improved systems are given in Fig. 5.3, which verify this trend. Using the vapor fraction is a way to combine the effects of various bo iler temperatures, pressures and solution ammonia mass fractions into one parameter. The mass flow rates of the vapor, strong solution and heat source are also normalized in using the specific boiler heat input and the vapor fraction.

PAGE 93

74 v a p o r f r a c t i o n l e a v i n g b o i l e r ( % ) b o i l e r h e a t i n p u t ( k J / k g ) 0 5 1 0 1 5 2 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 Figure 5.3. Boiler heat input per unit basic solution flow, for various vapor fractions leaving the boiler. h e a t s o u r c e f l o w r a t i o b o i l e r e f f e c t i v e n e s s ( % ) 0 1 2 3 4 5 3 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0 i m p r o v e d s y s t e m i m p r o v e d s y s t e m i n i t i a l s y s t e m i n i t i a l s y s t e m Figure 5.4. Boiler heat exchanger effectiveness for the initial and improved systems.

PAGE 94

75 The boiler heat exchanger effectiveness was found decrease with the heat sour ce flow ratio, such that at lower heat source flow rates, the hot water temperature at the boiler exit approaches the strong solution inlet temperature. Thus, the heat source can provide more of the maximum possible boiler input at lower flow rates. Fig 5. 4 supports that the addition of 47% more boiler surface area in the improved system also increased the boiler effectiveness. Although the heat source flow ratio has a significant effect on the boiler effectiveness, it is not the only parameter that does. T he fluctuations in the figure are due to other parameters such as the difference in fluid temperatures, boiler pressure, and solution ammonia mass fraction not being held constant. Vapor Fraction Leaving the Separator A vapor fraction between 0 and 1 is ob tained from operating the boiler between the bubble and dew points. More vaporization will occur for higher temperatures and lower pressures. In the initial system, only 59 3% of possible vapor production was realized on average as shown in Fig. 5.5. The experimental trends are reasonable, although the data do not correlate well with the equilibrium based simulations. With added insulation around the separator and vapor lines, and greater heat exchange area in both the recovery unit and boiler, the improv ed system generated 103 7% of possible vapor production on average for the given boiler conditions. A value over 100% of the maximum is the indirect result of a time lag between the strong solution and vapor flow rates. The vapor flow rate is steady as i t is pressure driven from a large tank. However, the vapor fraction, as a ratio of the vapor and strong solution flow rates, can seemingly increase past the maximum if the pump flow rate temporarily decreases.

PAGE 95

76 b o i l e r p r e s s u r e ( b a r ) v a p o r f r a c t i o n l e a v i n g s e p a r a t o r ( % ) 4 5 6 0 5 1 0 1 5 2 0 2 5 3 0 3 5 4 0 6 0 C s i m u l a t e d 6 0 C d a t a 7 0 C s i m u l a t e d 7 0 C d a t a 8 0 C s i m u l a t e d 8 0 C d a t a Figure 5.5. Vapor fractions in the initial system tests for various boiler temperatures and pressures, with a basic solution ammonia mass fraction of 45.6%. b o i l e r p r e s s u r e ( b a r ) v a p o r f r a c t i o n l e a v i n g s e p a r a t o r ( % ) 4 5 5 5 5 6 6 5 7 7 5 8 8 5 9 9 5 1 0 1 0 5 0 5 1 0 1 5 2 0 7 3 7 C s i m u l a t e d 7 3 7 C d a t a 8 2 9 C s i m u l a t e d 8 2 9 C d a t a 9 7 3 C s i m u l a t e d 9 7 3 C d a t a 1 0 1 1 C s i m u l a t e d 1 0 1 1 C d a t a Figure 5.6. Vapor fractions in the improved system tests for various boiler temperatures and pressures, with a basic solution ammonia mass fraction of 38.3%.

PAGE 96

77 Likewise, if the pump flow rate increases, then the boiler exit temperature will drop and equilibrium calculations will reduce the expected vapor output such that the actual vapor output seems to excel past the calculated limit. For eit her explanation, the results are encouraging, as vapor production has increased due to the system improvements. The vapor fractions of the improved system are shown in Fig. 5.6, where the maximum values are shown as the simulated results. Note that the hig hest vapor fraction seen in the experimentation was 19.8%. Vapor and Weak Solution Ammonia Mass Fractions The weak solution and vapor ammonia mass fractions will not be equal to that in the strong solution, as shown in the bubble and dew point diagram in Fig. B.1. The vapor will be formed at a high ammonia concentration owing to its lower boiling point than that of water, and the corresponding weak solution will have a lower ammonia mass fraction to maintain an overall mass balance in equilibrium. The conc entrations in the system are not those predicted under equilibrium conditions. The weak solution leaves the separator at a 2.0 0.6% (absolute; ie. if a 30% ammonia mass fraction is expected, the fluid leaves at 32%) greater ammonia mass fraction than tha t predicted from equilibrium at the measured boiler exit temperature, pressure and strong solution ammonia mass fraction. This is shown in Fig. 5.7, for trials at various boiler pressures and temperatures. For higher boiler pressures, the trends show that the weak solution is richer, as more ammonia can be sustained in the liquid for the given boiler temperature. For higher boiler temperatures the weak solution is leaner, as more ammonia can vaporize. Water will vaporize more with lower pressure and higher temperature also, but not in proportion to the ammonia vaporization.

PAGE 97

78 b o i l e r p r e s s u r e ( b a r ) m a s s f r a c t i o n o f a m m o n i a i n w e a k s o l u t i o n 4 5 6 0 2 0 2 2 0 2 4 0 2 6 0 2 8 0 3 0 3 2 0 3 4 0 3 6 0 3 8 0 4 0 4 2 0 4 4 6 0 C s i m u l a t e d 6 0 C d a t a 7 0 C s i m u l a t e d 7 0 C d a t a 8 0 C s i m u l a t e d 8 0 C d a t a Figure 5.7. Ammonia mass fraction in the weak solution for various boiler exit temperatures and pressures, with a basic solution ammonia mass fraction of 45.6%. b o i l e r p r e s s u r e ( b a r ) m a s s f r a c t i o n o f a m m o n i a i n v a p o r 4 5 6 0 8 8 0 9 0 9 2 0 9 4 0 9 6 0 9 8 1 6 0 C s i m u l a t e d 6 0 C d a t a 7 0 C s i m u l a t e d 7 0 C d a t a 8 0 C s i m u l a t e d 8 0 C d a t a Figure 5.8. Ammonia mass fraction in the vapor for various boiler exit temperatures and pressures, with a basic solution ammonia mass fraction of 45.6%.

PAGE 98

79 The vapor exits the separator also at a higher concentration, with a 1.9 0.3% greater ammonia mass fraction. Some results are shown in Fig. 5.8 and compared to simulations based on equilibrium. Trend lines for the experimental data are not included, as more data should be taken to produce a reliable fit. The weak solution and vapor ammonia mass fractions are shown on the sam e scale in Fig. 5.9 for one set of trials. The error in the data that is noted is based on measurement uncertainties. Other discrepancy with the simulated results can be attributed to inadequate modeling of losses and unsteady operation. b o i l e r p r e s s u r e ( b a r ) m a s s f r a c t i o n o f a m m o n i a 4 5 6 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 v a p o r s i m u l a t e d v a p o r d a t a w e a k s o l u t i o n s i m u l a t e d w e a k s o l u t i o n d a t a Figure 5.9. Amm onia mass fraction in the weak solution and vapor for a basic solution ammonia mass fraction of 45.6%, for various boiler pressures. The boiler exit temperature is 60 C. The basic solution concentration is measured only while the system is idle and assume d to be in equilibrium. The strong solution does not necessarily remain at the basic solution concentration during operation, and later testing included the measurement of the

PAGE 99

80 operating strong solution ammonia mass fraction. The strong solution tends to be come leaner during operation by 1.5 1.0%, on average, relative to the basic liquid ammonia mass fraction while the system is idle. The increase in ammonia content in the separator should balance the decrease in the absorber. Absorption Experiments sugge st that the mixing and absorption of the vapor stream into the weak solution in the absorber does not occur under equilibrium conditions. Much of the discrepancy can be attributed to the design of the absorber, in which the heat rejection occurs above the liquid pool in the vapor region. The two regions operate at different temperatures, which does not allow for equilibrium to regulate the pressure in the vessel. This has been a major source of instability while operating the system. Observations The vapor is bubbled into the liquid pool of the absorber through small holes to optimize the surface area available for the absorption to take place. Tested with ammonia vapor bubbled into an open water pool before installation into the system, the bubbles did not appear to break the surface. Most of the absorption is expected to take place in the liquid pool. The vapor could also be absorbed above the pool by the cooled weak solution spray, although this will not be as effective. Tests conducted with a pool level b elow the vapor inlet saw the absorber pressure rise quickly. The absorption heat is expelled as the cycle heat rejection to the ambient, which is simulated by the coolant flow. Although the heat is released mostly in the liquid pool, the cooling occurs abo ve the pool. In order to maintain the desired pool temperature, the weak solution spray needs to reach the pool below the desired pool temperature. This is where the problem lies. Essentially the coolant operates below the bubble temperature of

PAGE 100

81 the bulk so lution, which will lower the pressure in the absorber as the fluid condenses on the cooling elements. In order to maintain the desired saturation pressure, excess vapor must be present in the vapor region. Control of the absorber pressure is critical for t he operation of the solution pump. If the fluid at the pump inlet is not a compressed liquid, the pump intake stroke will flash boil the fluid, and the pump flow rate will decrease significantly. Cavitation can be heard and may damage certain pumps. Fluctu ations in pump flow rate will affect other parameters downstream and the system will not be steady. There are three ways to regulate the absorber pressure that have been effective. If less weak solution is returned to the absorber and allowed to instead po ol in the separator, the vapor will build up in the absorber above the pool and the liquid will leave the absorber compressed. This works, but the imbalance of mass flow rates has to be accounted for, and the weak solution needs to be brought back to the a bsorber eventually. Secondly, an inert, insoluble gas can be added to the vapor region in the absorber to increase the pressure by about 10 15%. Helium was used and improved the pump flow rate, but leaked through seals and had to be replenished after a few days. The increase in absorber pressure also limits the expansion possible across the turbine. Finally, careful matching of the coolant inlet temperature and flow rate reduces sway in pressure. Later trials were conducted with the latter two options appli ed. Coolant Flow Rate and Temperature The best coolant temperature was found by trial and error to be about 20 C (36 F) below the desired absorber pool temperature. The chiller maintains the coolant in a 6 C (11 F) range, which is still enough of a flu ctuation to change the absorber pressure by

PAGE 101

82 5 10%, and consequently the pump flow rate by up to 100%, if the coolant flow rate is not regulated. Thus for the low end of coolant inlet temperatures, the coolant flow rate was set low, while for the high end o f the temperature range, the flow rate was set at its maximum. Settings from the improved system tests are given in Fig. 5.10. Although an inconvenience to the operator, careful adjustment of the coolant flow rate kept the absorber pressure well behaved. a b s o r b e r a n d c o o l a n t t e m p e r a t u r e d i f f e r e n c e ( C ) c o o l a n t f l o w r a t e ( l i t e r / m i n ) 1 6 1 8 2 0 2 2 2 4 0 2 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 2 2 Figure 5.10. Coolant flow rates for various temperature differences between the coolant inlet and the absorber pool. Absorber Heat Rejection The part of the boiler heat input that is not converted to work output must be rejected as heat from the absorber to the ambient sink. For both the initial and improved systems, the heat rejection is shown in Fig. 5.11 as a function of the vapor production in the boiler. The heat transfer was calculated from properties of the vapor and weak solution inlet streams and the strong solution exit stream from the absorber. The results

PAGE 102

83 match those given in the boiler heat input plot of Fig. 5.3, with lesser magnitudes. The difference in the values should equal the net work output from the system. v a p o r f r a c t i o n l e a v i n g b o i l e r ( % ) a b s o r b e r h e a t r e j e c t i o n ( k J / k g ) 0 5 1 0 1 5 2 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 Figure 5.11. Absorber hea t rejection per unit basic solution flow, for various vapor fractions leaving the boiler. Potential Work Output, Cooling Capacity and Cycle Efficiencies The experimental evaluation of the cycle is incomplete because the turbine is being modeled as an equiv alent expansion process. Testing has shown which conditions can be produced at the inlet to the expansion device, and which pressure ratios are available for the expansion process, but leave the device characteristics to be a free variable as no turbine is yet installed. Thus, based on the experimentally observed vapor generation and expansion ratios, the work output can be estimated if the turbine efficiency is prescribed. With additional consideration for the heat source and ambient temperature, the cooli ng capacity and cycle efficiencies are also evaluated. The results in

PAGE 103

84 this section will use a semi empirical approach to combine the effects of measured data and component efficiencies. Parametric Dependence The turbine work output exhibits a clear proport ional relationship with the turbine isentropic efficiency, as shown by its definition in Eq. 5.1. More work can be developed for more efficient expansion. Eq. 5.2 shows an inversely proportional relationship of the pump work input with the pump isentropic efficiency, with more pump work required for less efficient pumping. max t t t W W h = (5.1) p p p W W h min = (5.2) The effect of various pump and turbine efficiencies becomes less evident when looking at the first and second law efficiencie s. The results presented in this chapter use the definitions given below. In most cases, cooling was not observed such that the relationships can be simplified somewhat. However, the ratio of pump and turbine work is not consistent for various system tests and a clear correlation is not available, although it should depend on the pressure ratio and vapor production. + + = c p p t t h E W W Q min max 1 1 1 h h h (5.3) + + D = c p p t t hs E W W E min max 2 1 1 h h h (5.4) For typical operating conditions, the simulated effects of the turbine and pump e fficiencies on the cycle first and second law efficiencies are shown in Figs. 5.12 and 5.13.

PAGE 104

85 0 7 0 1 0 5 0 5 0 9 0 9 0 9 1 4 1 4 1 4 1 8 1 8 1 8 2 2 2 2 2 2 2 6 2 6 2 6 3 0 3 0 3 0 3 4 3 4 3 4 3 8 3 8 4 2 4 2 4 6 4 6 5 0 t u r b i n e i s e n t r o p i c e f f i c i e n c y p u m p i s e n t r o p i c e f f i c i e n c y ( % ) 5 0 6 0 7 0 8 0 9 0 1 0 0 5 0 6 0 7 0 8 0 9 0 1 0 0 Figure 5.12. Simulated first law efficiency (%) for various turbine and pump isentropic efficiencies, and with other typical losses. 1 7 0 7 3 1 5 5 5 5 7 9 7 9 7 9 1 0 3 1 0 3 1 0 3 1 0 3 1 2 7 1 2 7 1 2 7 1 5 1 1 5 1 1 5 1 1 7 5 1 7 5 1 7 5 1 9 9 1 9 9 1 9 9 2 2 3 2 2 3 2 4 7 2 4 7 2 7 1 t u r b i n e i s e n t r o p i c e f f i c i e n c y p u m p i s e n t r o p i c e f f i c i e n c y ( % ) 5 0 6 0 7 0 8 0 9 0 1 0 0 5 0 6 0 7 0 8 0 9 0 1 0 0 Figure 5.13. Simulated s econd law efficiency (%) for various turbine and pump isentropic efficiencies, and with other typical losses.

PAGE 105

86 The simulated results above are for a boiler pressure of 8 bar (101 psig), 40% basic solution ammonia mass fraction, ambient at 15 C (59 F), hea t source at 80 C (176 F) and heat source flow ratio of 2.5. Typical losses listed in Chapter 4 are included. The plots verify that the expected performance improves for more efficient pumping and expansion. The trends are similar for both efficiencies, b ut with different magnitudes. The heat source and ambient temperatures are important in considering the cooling capacity, since the expansion can lower the hot vapor temperature only a given amount. For the same operating conditions as above, the expected cooling capacity is given for variation in turbine efficiency and heat source temperature in Fig. 5.14. No cooling is expected for high heat source temperatures, even with an isentropic expansion. The potential for cooling is increased for more efficient e xpansion. 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 2 0 2 0 2 0 3 0 3 0 3 0 4 0 3 h e a t s o u r c e t e m p e r a t u r e ( C ) t u r b i n e i s e n t r o p i c e f f i c i e n c y ( % ) 7 0 7 5 8 0 8 5 8 6 8 8 9 0 9 2 9 4 9 6 9 8 1 0 0 Figure 5.14. Simulated cooling capacity (kJ/kg) per unit basic solution flow, for various turbine isentropic efficiencies and heat source temperatures.

PAGE 106

87 Pump Work Input The pump work is estimated using an assumed pump isentropic efficiency, whi ch was not experimentally measured. Determination of the pump isentropic efficiency would require more accurate measurement of the inlet and exit temperature difference than is available. A typical value of 80% is suggested in the literature (Drbal et al., 1996). Relative to the turbine work output, the pump work input is minor except in cases of low vapor production. The ratio of turbine to pump work is nearly proportional to the vapor production. Fig. 5.15 shows experimental results of pump work intput fo r the pressure rise from the absorber to the boiler. For higher required pressures, the pump must obviously work harder. Results are shown for the improved system, which uses a different pump than the initial system. p u m p p r e s s u r e r i s e ( b a r ) p u m p w o r k i n p u t ( k J / k g ) 1 2 3 4 5 6 7 8 9 0 0 5 1 1 5 2 2 5 3 3 5 4 Figure 5.15. Pump work input (kJ/kg) per unit basic solution flow, for various pressure rises from the absorber to the boiler.

PAGE 107

88 Turbine Work Output The parametric dependence of the turbine work output was discussed previously and a typical turbine isentropic efficiency of 90% is used as sugge sted in the literature (Drbal et al., 1996). Considering the low power output of the current system, however, the efficiency can be lower as friction becomes more significant with micro turbines. Reducing seal friction might be a solution for other applica tions but is not an option for this system as ammonia leaks are not tolerable. Assuming a turbine efficiency of 90%, the largest work output for the conditions tested was found to be 0.47 kW. Even with a perfect turbine, the output would only increase to 0 .52 kW. Higher boiler pressures would increase the power output, but would require higher boiler temperatures. A limit is placed on the boiler temperature if simulating a low temperature heat source, and high boiler temperatures also limit the potential fo r refrigeration. Hence the cycle can be tuned to balance the power and cooling output as required by the application. A factor that can be adjusted for increasing power output is the available pressure ratio. The largest pressure ratio that was tested was 4.4, which is governed by the boiler and absorber pressures. A measurement of losses at the vapor inlet to the absorber yields Fig. 5.16, showing an increase in the pressure drop at the vapor inlet to the absorber for increasing flow rates. Although a high er flow rate of vapor will have the capacity to produce more work and refrigeration, the corresponding increase in inlet pressure loss implies less expansion will be possible as the turbine exit pressure will be higher. Figure 5.17 shows a similar pressure drop across the weak solution spray inlet to the absorber. This pressure loss is not as consequential with respect to the system performance.

PAGE 108

89 v a p o r f l o w r a t e ( l i t e r / s e c ) b u b b l e i n l e t p r e s s u r e l o s s ( b a r ) 0 2 4 6 8 1 0 0 0 1 0 2 0 3 0 4 0 5 0 6 Figure 5.16. Measured pressure loss across the vapor bubble inlet to the absorber. w e a k s o l u t i o n f l o w r a t e ( l i t e r / m i n ) s p r a y n o z z l e p r e s s u r e l o s s ( b a r ) 0 1 2 0 0 5 1 1 5 2 2 5 3 3 5 Figure 5.17. Measured pr essure loss across the weak solution spray inlet to the absorber.

PAGE 109

90 In addition to the expansion ratio, the other major factor found to influence the work output is the specific vapor generation, or vapor fraction. The vapor fraction, expansion ratio, and s pecific turbine work output are all normalized with respect to flow rates and are plotted together in Fig. 5.18. More work is produced for higher vapor flows as well as for greater expansion ratios. A 90% turbine efficiency is assumed. Changing the assumed turbine efficiency would simply change the slopes of the plotted trend lines. v a p o r f r a c t i o n l e a v i n g b o i l e r ( % ) t u r b i n e w o r k o u t p u t ( k J / k g ) 0 5 1 0 1 5 2 0 0 2 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 4 3 4 3 9 2 1 2 6 e x p a n s i o n r a t i o Figure 5.18. Expected work output (kJ/kg) per unit basic solution flow from a 90% efficient turbine, for observed vapor fractions and expansion ratios. Cooling Capacity The p otential for cooling was witnessed only with some high expansion ratio trials and for relatively low heat source temperatures. The magnitude of cooling was about 1/5 th of the power output for the assumed 90% efficient turbine and approach temperature limit s of 5 C (9 F) in the refrigeration unit. The magnitude of cooling is expected to be higher for increased turbine efficiency.

PAGE 110

91 First Law Efficiency The first law or thermal efficiency accounts for both the assumed turbine work output and pump input, and g auges them against the measured heat addition from the source. For cases of low vapor generation, the pump work may outweigh the turbine work and negative efficiencies can result. These results can be avoided experimentally by simply operating under condit ions of sufficient vapor generation. The highest first law efficiency seen in the experimental data for a 90% efficient turbine and 80% efficient pump was 6.6%, for a case with a high expansion ratio but not necessarily a high vapor fraction. Other results for various vapor fractions and expansion ratios are shown in Fig. 5.19, and indicate a greater overall efficiency for when more vapor is generated and a higher expansion potential exists. v a p o r f r a c t i o n l e a v i n g b o i l e r ( % ) f i r s t l a w e f f i c i e n c y ( % ) 0 5 1 0 1 5 2 0 0 2 4 6 8 3 4 3 9 2 1 2 6 e x p a n s i o n r a t i o Figure 5.19. Expected first law efficiency with a 90% efficient turbine, for observed vapor fractions and expansion ratios.

PAGE 111

92 Second Law Efficiency The second law or exergy efficiency gauges the net work output relative to the full potential of the source. It is a measure of how much of the available energy of the heat source is converted by the system to usable output. In a power plant analogy, the second law efficiency could be explained as the electrical output in kWh per ton of coal that is converted. Comparison of the electrical output of multiple power plants per t on of coal would be a good way to compare the power plants, just as the second law efficiency is a good standard to measure this cycle by. The largest second law efficiency seen in the experimental data for a 90% efficient turbine and 90% efficient pump wa s 39.5%, for a case with a high expansion ratio. The results are presented in Fig. 5.20, and are similar to those given for the first law efficiency. v a p o r f r a c t i o n l e a v i n g b o i l e r ( % ) s e c o n d l a w e f f i c i e n c y ( % ) 0 5 1 0 1 5 2 0 0 5 1 0 1 5 2 0 2 5 3 0 3 5 4 0 3 4 3 9 2 1 2 6 e x p a n s i o n r a t i o Figure 5.20. Expected second law efficiency with a 90% efficient turbine, for observed vapor fractions and expansion ratios.

PAGE 112

93 CHAPTER 6 RECOMMENDATIONS FOR FUTURE WORK In operating the experimental system, further areas of improvement in the current setup were determined and are addressed in this chapter. Suggestions are made to orient future experimental work towards a practica l, fully functional and operator friendly system that is ready for demonstration to the industry. Target Areas for Improvement An evaluation of components and processes where improvements in the current experimental system can be made is provided in this s ection. These are areas where concerns have arisen previously and which merit greater attention in further system modifications. Turbine/Generator Set and Refrigeration Unit In order to fully realize the potential for doing work and obtaining cooling, the experimental study should be continued with the addition of a turbine/generator set. The generator may simply be a frictional heat dissipater coupled to a torque arm, from which calculations can provide an estimate of the work output of the expander. A dyn amometer can be used for larger outputs. This will confirm that the cycle can produce work. The expansion to low temperatures will confirm the potential to obtain cooling for certain cases. With the inclusion of a refrigeration unit in line after the turbi ne, the cooling can be utilized. The refrigeration unit can be a finned coil unit cooling the ambient air, or one that uses water as the medium to be cooled. This may be an easier

PAGE 113

94 option to work with, as the flow rate of water can be determined accurately and the temperature measurement at the refrigeration unit inlet and exit will be more uniform. Absorber Design The largest source of instability in the operation of the system has been from an inappropriate absorber design. The cooling provided by the heat exchangers in the vapor region of the absorber encourages more condensation than absorption. The vapor pressure in the absorber is very sensitive to the flow rate and temperature of the coolant, while it should be most sensitive to the absorber pool tempe rature. The condensation in the vapor region serves to lower the absorber pressure to slightly below its saturation pressure, which makes pumping the liquid difficult. With a cooling unit in the liquid pool, the absorber would operate at slightly above its saturation pressure. An immediate improvement can be with the addition of a heat exchanger into the liquid pool of the absorber. Cooling the liquid pool itself will encourage vapor absorption in the pool, where the vapor bubbles into it. The coolant tempe rature also has to be relatively low for the condensation to occur, which is not the best scheme if eventually the ambient is to be used as the heat sink. In a practical unit, the absorption process would be air cooled for small scale applications and wate r cooled for large scale applications. For a more involved improvement, other absorption designs should be considered and tested, which use water or air at ambient temperatures as the heat sink. Boiler Heat Exchanger The observed boiling is not uniform in the strong solution. After the boiler, the mixing of the fluid causes vapor to condense into the cooler bulk solution, thus raising the overall fluid temperature but reducing the final vapor fraction produced by the boiler.

PAGE 114

95 More surface area of boiling wou ld provide an equivalent longer residence time in the boiler and a more uniform fluid leaving the boiler. The surface area was increased in the system modifications, with a corresponding increase in the witnessed heat exchanger effectiveness. This should b e further increased in future improvements. The boiling process is intricate and requires complex modeling. Although a flat plate heat exchanger has been effective, a study is suggested to determine which type of heat exchanger would be most appropriate fo r this type of system. A theoretical approach can be followed with testing an advanced boiler apparatus that can bypass the current boiler in the system. Ammonia Water Pump The pump has been an issue in, but not the cause of, the flow rate instabilities. T he diaphragm pump performs well, provided the liquid leaving the absorber is a compressed liquid. The advantage of this pump is in providing high operating pressures, which do not limit the flow capability of the pump. The pulses of the pump were also an i ssue initially, but have been dampened by the expansion tank at the exit. However, in order to ensure that the pump works well, the liquid leaving the absorber should be compressed either with an improved absorber design as discussed above or with the subc ooling of the liquid before the pump inlet by other means. The absorber can be elevated in future designs to generate head pressure and compress the liquid at the pump inlet. Larger diameter tubing at the pump inlet will also reduce pressure losses, which would otherwise induce cavitation in the pump. Coolant Temperature Fluctuations The chiller cycle provides coolant in roughly a 6 C (11 F) temperature range. As the coolant cycles through the full range, the absorber pressure will vary if the coolant

PAGE 115

96 flo w rate to the absorber is fixed, and the pump performance will not be steady. This is a large factor in the stability of the system. Either a setting on the chiller should be adjusted to have it cycle more frequently over a tighter range of temperatures, o r a mixing scheme can be used to mix the coolant from the supply with some of the return to maintain a constant temperature coolant supply to the absorber. The coolant return to the coolant storage tank can also be sent to the chiller directly, then to the storage, and finally back to the absorber, all in series. This may steady out the absorber coolant supply temperature very effectively. Heat Source Capacity The heat source is capable of providing 5 kW continuously through the heating element in the singl e pass boiler. With thermal storage in the phase change material of the single pass boiler and hot water storage in the hot water tank, higher heat inputs have been produced for system trials but at the expense of heat source temperature. With the heat inp ut typically over 5 kW in the cases studied, the expected work output has at most been around half a kilowatt. Thus in order to obtain a larger work output, the heating capacity of the hot water boiler should be increased by at least 2 or 3 times. Control Methods A study to develop a good control scheme for the system is necessary. Currently, manual adjustment of valves is necessary to obtain pseudo steady behavior. At times the fluctuations are over a long period, but still some level of automated control is welcome. It is certainly in the interest of the long term goals of the project to produce a system that requires only initial settings and no monitoring if it is to be commercially viable. Liquid level The pooling of the liquid into the separator has t o be monitored and occasionally the weak solution needs to be returned. This pooling is not necessarily

PAGE 116

97 proportional by mass of ammonia and water, and the basic solution concentration may change over time. With a better absorber design and proper settings, no pooling would be necessary and the liquid level in the absorber could be maintained with a liquid level controller. Heat source temperature The mixing valve currently employed on the hot water loop is only moderately effective. The cold and hot water inlets are regulated by the thermal expansion of a metal. A thermostatically controlled solenoid valve may be more effective to get the desired heat source temperature. Coolant temperature and flow rate The absorber coolant destabilizes the absorber press ure as it cycles through its range of inlet temperatures. The coolant flow rate is adjusted manually to counter the effect, which can be handled through a control valve. The coolant temperature fluctuations can also be minimized with a tightening of the ch iller supply temperature range, or with a mixing scheme such as that used with the heat source. Flow rates and boiler pressure With initial settings of valves, and provided that the absorber pressure remains steady, the system flow rate and subsequently t he boiler pressure should not fluctuate. The weak and vapor flow rates are set initially with their respective valves and do not change unless the separator pressure changes. The heat source flow rate remains steady also over long periods. The source of th e unsteadiness is usually in the absorber, as addressed before. Operating Limits When applied together, the component limits have reduced the effective range of system operation. The range is not comfortably large, and the system struggles to perform withi n the narrow range. Improvement of the pump flow rates, absorber design, and

PAGE 117

98 heating capacity will increase the operating range, and should be considered if the system is to be scaled up. Insulation The system tubing, vessels and flow meters are well insul ated. Fittings, valves and other components may be insulated as well to reduce heat losses, although this has not been isolated as a major problem with the current setup. Component Modeling in Simulations In the simulations used thus far, the properties ar e determined by assuming a steady system for which thermodynamic equilibrium exists at all the state points. This has been shown not to be the case, and better modeling of component processes should be incorporated into further simulations. Priorities shou ld be given to processes that are critical in the overall simulation, namely the vapor generation, expansion and absorption. Suggested Strategy for System Modifications The improvements listed above should be handled in an order of priority based on time a nd financial considerations. Although a full demonstration unit is the eventual goal, other areas deserve greater attention for immediate improvements. These improvements should use the current setup as a starting point, with localized modifications made a s needed. After sufficient experience and confidence on all the major components is obtained, a scaled up practical demonstration unit should be built from ground up. Short Term Plan The absorber unit should be improved as a priority. With little financial burden and with relative ease, a cooling element can be placed into the liquid pool by changing around some of the plumbing in the current absorber. Concurrently with ongoing tests, a

PAGE 118

99 better design for the absorption process can be developed and construct ed, and eventually installed by bypassing the modified absorber. With these first absorber improvements, the system pump flow rate should be increased to nearly 1 gpm, and will be steady. The higher flow rate should provide on the order of 1 kW of power ou tput, if the heating capacity of the hot water system is increased to around 10 15 kW. This can also be done quickly and inexpensively with the inclusion of resistance heaters into the hot water storage tank, or even with heating tape wrapped around tubing An increase in the hot water pump flow rate should also be considered by installing a higher horsepower motor or with a totally new pump/motor set. Currently the pump produces at most 1.5 gpm through the hot water loop, whereas 3 4 gpm would be a more co mfortable operating range. Note that the hot water loop allows for flow rate control regardless of the pump output, such that a higher possible output would leave more options. The increased flow rate of the basic solution will yield flows near the higher range of strong, weak and vapor flow meter limits. Scaling the system to much higher flow rates will require costly replacement of much of the equipment and should be considered only under long term plans. An expander is currently being developed in lab an d with the cooperation of outside interests. The installation of a working turbine/generator into the system is of immediate interest, to verify that the cycle can indeed produce work. This should be a priority in the system modifications. Although selecti ng the most appropriate turbine will require an involved study, the results from a first expander should be very encouraging and is a significant step.

PAGE 119

100 Long Term Plan A demonstration unit should produce on the order of 5 10 kW of power output, in order to attract serious interest from industry. This will require a scaling up of most of the system, and should proceed only after thorough consideration of component types and sizes. The system configuration should be optimized and its behavior anticipated, as c onstruction of the demonstration unit will be time and cost intensive and should not have to be repeated. However, the unit should also allow flexibility in its design in case further modifications are found necessary. In initial studies, the use of a hot water heater allows for easy simulation of the heat source. For the demonstration unit to be effective, however, the heat source should be water heated by flat plate solar collectors as initially conceived. The heat sink must also be more practical, and sh ould not require external power to produce the cool fluid as it does in the current setup. Ambient air or ground water should be used as the heat sink for the demonstration unit. Finally, an economic analysis should be performed to gauge the viability of t he proposed cycle in comparison with alternative energy conversion systems. Although intuitively the cycle has the potential to be cost effective, this critical analysis has not been performed yet.

PAGE 120

101 CHAPTER 7 CONCLUSIONS An experimental system was designed and constructed based on an ammonia based combined power and cooling cycle that has been proposed by D. Y. Goswami. The heat source of the experimental system operates at temperatures that can be p roduced by low temperature solar and geothermal resources, while the heat sink simulates various ambient conditions. The power and cooling outputs of the system are estimated through simulation of the turbine by an equivalent expansion process and by model ing of the refrigeration heat exchanger. Results verify that vaporization of the working fluid can occur at various boiler temperatures and pressures. The vapor is generated at sufficiently high pressures to produce work if an expander were in place and th e potential for cooling capacity was observed for certain operating parameters. Analysis of losses of the initial system showed where improvements in heat exchange, vapor generation, pump performance and overall stability could be made. Implementation of t hese system changes yielded encouraging results and showed where further improvements could be made. The potential of the power and cooling cycle as an alternative to conventional fossil fuel technologies was acknowledged by the experimental work of this d issertation. However, further studies are necessary to realize this potential and to produce a complete demonstration of the proposed cycle.

PAGE 121

102 APPENDIX A AMMONIA TOXICITY The toxic nature of ammonia is detailed in Table A.1, giving exposure limits and the corresponding responses exhibited by humans. Table A.1. Ammonia exposure limits (Pillis, 1993). Exposure (ppm) Effects 0 5 Smell ha rdly detectable. 5 20 Human nose starts to detect. 25 TLV TWA (Threshold Limit Value Time Weighted Average, 8 h) 35 STEL (Short Term Exposure Limit 15 min) 150 200 Eyes affected to limited extent after about 1 min exposure. Breathing not af fected. 500 IDLH (Immediately Dangerous to Life and Health, per NIOSH) 600 Eyes streaming in about 30 s exposure. 700 Tears to eyes in seconds. Still breathable. 1000 Eyes streamed instantly and vision impaired, but not lost. Breathing intoler able to most participants. Skin irritation to most participants. 1500 Instant reaction is to get out.

PAGE 122

103 APPENDIX B BINARY FLUID PROPERT Y EVALUATION Accurate modeling of the power and cooling cycle relies on the accuracy of property evaluation methods, which has proven difficult for mixtures. This appendix provides a brief overview of property evaluation met hods for mixtures, and discusses in detail the correlations used in the current study for both the ammonia/water working fluid and the ethylene glycol/water coolant. Characteristics of Ammonia and Water The properties of ammonia are well known as it has be en used for over 100 years as a refrigerant (Pillis, 1993). Likewise, the properties of water as a pure fluid are well established. The experimental data and correlations for ammonia and water mixtures are available to a lesser extent, as the use of binary mixtures developed later. Ammonia by itself is termed dry, pure or anhydrous ammonia. The combination of pure ammonia with water into a liquid form produces ammonium hydroxide according to the reaction given in Eq. B.1. If the reaction occurs with an equa l or excess amount of water the resulting solution is termed aqueous ammonia. OH NH O H NH 4 2 3 + (B.1) Ammonia and water liquid mixtures are characterized as non ideal, polar and non azeotropic. The polarity in particular makes the mixture difficult to model owing to intermolecular attractions. The mixture in vapor form is more forgiving to modeling as molecules are further apart.

PAGE 123

104 As the boiling points of ammonia and water are not the same, an isobaric phase change of the mixture will occur over a ran ge of temperatures, as will an isothermal phase change occur over a range of pressures. A constant pressure phase change diagram for a non azeotropic mixture is shown in Fig. B.1. According to Fig. B.1, a mixture that is above the bubble point temperature will exist in equilibrium with vapor that will typically be at a much higher concentration of the more volatile species, here being ammonia. 0 0 2 5 0 5 0 7 5 1 m a s s f r a c t i o n o f a m m o n i a s a t u r a t i o n t e m p e r a t u r e b u b b l e p o i n t l i n e d e w p o i n t l i n e Figure B.1. Bubble and dew point diagram for a non azeotropic mixture. Ideal Models Ideal models for fluid mixtu res are applicable mostly for the vapor phase. For ideal mixing in the liquid phase, the fluids must be inherently similar for the models to be valid.

PAGE 124

105 The partial pressure is proportional to the mole fraction in the vapor phase for ideal mixtures as: P y P i i ~ = (B.2) The ideal mixture properties are found from a linear mass averaging scheme of the single component properties, given generically as: 2 2 1 1 F x F x F avg + = (B.3) For idealized binary vapor systems, Henrys Law or Raoults Law ca n be used (Collier and Thome, 1996). Henrys Law states that the partial pressure is directly proportional to the mole fraction in the liquid phase, according to Eq. B.4. ~ i i x const P = (B.4) This is valid for dilute solutions, where the mole fra ction of component i is small. For high values of the liquid mole fraction, use Raoults Law in Eq. B.5, which states the partial pressure is related to the vapor pressure of pure component i at the same temperature and the liquid mole fraction of componen t i. ~ i g i i x P P = (B.5) Semi Empirical Correlations Most binary mixtures do not behave ideally, and empirical or semi empirical correlations are usually applied. There are a number of published experimental results giving mostly vapor liquid eq uilibrium (VLE) data (Magee and Kagawa, 1998; Polak and Lu, 1975; Rizvi and Heldemann, 1987; Sassen et al., 1990; Tsiklis et al., 1965). Accurate experimental determination of ammonia water thermodynamic properties has been limited by the toxic, corrosive, flammable and volatile nature of the fluid. Modeling and

PAGE 125

106 correlation of the properties has been difficult because of the strong polarity of the two fluids and their mutual chemical affinity (Weber, 1999). The correlations that have been determined are typ ically for lower pressures and temperatures than are used in conventional power cycles, whereas in certain power cycles the critical point is approached or even exceeded. Power cycles using ammonia and water operate as high as 200 bar (2900 psi) and 650 C (1200 F) (Thorin et al., 1998). A summary of about 30 correlations for the determination of ammonia water mixture properties is given by Thorin et al. (1998). Most of the correlations use separate equations for the liquid and vapor phases. The vapor is a ssumed to behave as an ideal mixture while the liquid equations include a corrective term that accounts for mixing effects according to the Gibbs energy. Most of the correlations are based on VLE measurements that have been taken since the early 1900s. Mo st of the data is for low pressures and temperatures. Few studies have given measurements for density, critical pressure and temperature, heat of mixing and heat capacity. Most correlations fall into the following categories (Thorin et al., 1998): 1. Cubic eq uations of state 2. Virial equations of state 3. Gibbs excess energy 4. The law of corresponding states 5. Perturbation theory 6. Group contribution method 7. Polynomial functions

PAGE 126

107 Cubic Equations of State Phase equilibria and all thermodynamic properties of fluid mixtures c an be calculated with equations of state (EOS). The cubic equation of state includes modifications of the ideal gas law with additional terms to adjust for non ideality. Relations are in terms of PVT, such that appropriate constants in the equations can be found experimentally by measuring PVT data. Equations of state are usually cubic in volume. An equation of state for both pure fluids and mixtures would simplify thermodynamic property determination greatly, but has not been achieved. Van der Waals was th e first to produce a cubic equation of state in 1873. Other well known cubic equations of state include Peng Robinson (1976) and Redlich Kwong (1949) (Modell and Reid, 1983; Stryjek and Vera, 1986). Stryjek and Vera (1986) give a modified Peng Robinson EOS Modifications have allowed extension of the EOS to low reduced temperatures. For binary systems, terms are found using conventional mixing rules. Non polar mixtures can be treated with single parameter mixing rules while for polar mixtures two parameters are required. Peng Robinson can also be modified with a temperature dependent interaction parameter to better model liquid phases in the mixture. Enick et al. (1998) looked into the Peng Robinson EOS modified with a two parameter mixing rule, and compared the results with the proprietary correlations used by the Kalina cycle developers with good agreement. Smolen et al. (1991) give a cubic EOS, basically a modified Redlich Kwong EOS with the inclusion of an additional volume term and a term following a den sity dependent mixing rule. Skogestad (1983) reviews another modified Redlich Kwong EOS with greater temperature dependence, but with limited accuracy in ammonia water mixtures.

PAGE 127

108 In general, the cubic EOS works well for non polar mixtures in a limited range For polar systems it can predict phase equilibria well, but cannot predict densities (Thorin et al., 1998; Weber, 1999). Polar mixtures are more difficult to model, and require changes to the typical mixing rules (Vidal, 1983). Virial Equations of State The virial equation of state is a power series in the inverse of molar specific volume, truncated according to approximations to a few terms (Tsonopoulos, 1974). The virial EOS can give the compressibility factor in relation to composition, temperature, an d pressure or density. The virial coefficients in the EOS can be taken from PVT data directly, correlated empirically, or estimated from molecular thermodynamics. Using mixing rules based on statistical mechanics, the intermolecular potential energy is mod eled to obtain the generic coefficients in the virial equation (Hayden and OConnell, 1975). PVT data for pure components is then fitted to the equation to determine mixture specific constants. Duan et al. (1996) presents a 15 parameter virial EOS for ammo nia water mixtures, which accurately provides both phase equilibrium and volumetric properties. There are also methods that are purely theoretical for predicting coefficients of virial equations of state, based on critical properties and molecular characte ristics, but most are not very accurate. Hayden and OConnell (1975) report a method with reasonable accuracy, but poor for polar mixtures such as ammonia and water. Gibbs Excess Energy The Gibbs excess energy method is a convenient means to determine mixt ure properties based on classical thermodynamics and empirical data. Correlations for measured pure component properties are fitted into the definition for Gibbs energy,

PAGE 128

109 which can be manipulated according to Maxwell relations. The result is a set of semi e mpirical pure component relations for desired thermodynamic properties such as molar specific enthalpy, entropy and volume. The pure component properties are weighted and added to determine the mixture properties. Excess properties must also be included, a ccording to the excess Gibbs correlations for the mixture determined from experimental data (El Shaarawi and Al Nimr, 1990; Ibrahim and Klein, 1993). Although pure component and experimental data are needed, the final mixture correlations are typically acc urate and easy to work with. The properties can be determined with few calculations once the final correlations are established, making this method efficient if used in conjunction with computer code that repeatedly calls for the properties of the mixture. Law of Corresponding States Van der Waals developed the law of corresponding states based on classical thermodynamics, stating that a reduced equation of state valid for one fluid is valid for all fluids. This allows for comparison of thermodynamic proper ties of different fluids if the properties at the critical points are known. The law of corresponding states can also be based on statistical mechanics, mapping the potential energy function of molecules in the fluids rather than an equation of state (Mode ll and Reid, 1983; Prausnitz et al., 1999). The results are comparatively accurate, even when the behavior of one of the fluids is sometimes poorly known. The method involves mapping the thermodynamic properties of a mixed fluid to that of a pure fluid. Th e pure fluid is arbitrary and only used as a reference fluid. The difficultly in the methods lies in proper choice of the reference fluid, as the mapping may transform the phase to a point close to the phase boundary, where convergence of the numerical sol ution may not occur. Nowarski and

PAGE 129

110 Friend (1998) found that either ammonia or water served well as the reference fluid, with few problems with the transformation to the binary mixture of them. Nowarski and Friend (1998) map a surface for Helmholtz energy, w hich involves knowing how temperature and density are scaled in the mapping. Scaling factors are solved for simultaneously with empirically determined fluid interaction parameters, in an iterative manner. The mixed fluid can have more than two components, although Nowarski and Friend discuss the analysis for the binary mixture of ammonia and water. The binary interaction parameters in this case were dependent on both temperature and mixture composition. Weber (1999) also applied the basic method for corresp onding states, with mixture specific modifications made for ammonia water with the addition of three mixing parameters to account for chemical interactions between unlike molecules. The method has few parameters that need to be adjusted, and it can be read ily extrapolated through the range of available data. The method predicted density and phase equilibrium data over a wide range of temperatures. Perturbation Theory Perturbation theory provides corrections for the imperfect behavior of real fluids relative to an assumed ideal fluid. The properties of the ideal fluid will be known, and correlations are established relating the properties of the real fluid to those of the ideal fluid. For gases, the ideal fluid is taken simply as an ideal gas and the relation ship includes an expansion in density about the compressibility factor. For dense fluids, it becomes more difficult to select an appropriate reference fluid (Prausnitz et al., 1999).

PAGE 130

111 Group Contribution Method The group contribution method is a molecular th ermodynamic approach that predicts VLE data for systems of multiple molecules based on the behavior of the components of these molecules. The prediction is independent of the molecules in the system, and relies completely on the atomic groups that the mole cules are composed of. The interaction of the molecular species is determined from the weighted interaction of the functional groups. Each group is assumed to behave independently of the molecular species in which it appears, so the behavior of the group c an be determined from any species containing the group for which there is experimental data. The group contribution method is useful for predicting VLE data for systems in which data is partial or not available at all. However, the predictions are not very accurate and should only be taken as first order approximations (Prausnitz et al., 1999). Property Correlations Used in the Current Study There are three fluids that require property estimation in the proposed cycle. The Gibbs energy correlation for pure water as the heating fluid is well established. Ammonia and water mixture properties are determined by combining pure fluid correlations and excess Gibbs energy for the mixture. A similar Gibbs energy approach is used for the ethylene glycol and water cool ant. Ammonia and Water Mixtures A convenient semi empirical scheme is used here that combines the Gibbs energy method for mixtures and bubble and dew point temperature correlations for phase equilibrium. The calculated results have been compared to experim ental mixture properties in the literature with good agreement (Xu and Goswami, 1999). The

PAGE 131

112 correlations are valid in the 0.2 to 110 bar ( 12 to 1580 psig) pressure range and 43 to 327 C ( 46 to 620 F) temperature range (Ibrahim and Klein, 1993). Gibbs e nergy in integral form can be defined as Eq. B.6. The liquid phase volume can be assumed to be first order in pressure and second order in temperature, according to Eq. B.7. Similarly, the liquid heat capacity at constant pressure can be assumed to be seco nd order in temperature according to Eq. B.8 (Ziegler and Trepp, 1984). + + = T T T T P P P P dT T C T VdP dT C TS H G 0 0 0 0 0 (B.6) 2 4 3 2 1 T a T a P a a V L + + + = (B.7) 2 3 2 1 T b T b b C L P + + = (B.8) The corresponding gas relations are empirically determined to be of the form given in Eqs. B.9 and B.10 (Ziegler and Trepp, 1984). 11 2 4 11 3 3 2 1 T P c T c T c c P RT V g + + + + = (B.9) 2 3 2 1 T d T d d C g P + + = (B.10) Integration of Eq. B.6 yields a liquid correlation for the Gibbs energy of a pure fluid, as given in Eq. B.11 in reduced form. ( ) ( ) ( ) 3 0 3 3 2 0 2 2 0 1 0 0 3 2 r r r r r r L r r L r L r T T B T T B T T B s T h g + + + = ( ) ( ) 2 0 2 3 0 2 0 1 2 ln r r r r r r r r r T T T B T T T B T T T B ( ) ( ) ( ) 2 0 2 2 0 2 4 3 1 2 r r r r r r P P A P P T A T A A + + + + (B.11)

PAGE 132

113 Likewise, integration of Eq. B.6 with the assumed forms of heat capacity and volume of Eqs. B.9 and B.10 yields Eq. B.12, in reduced form. Eq. B.12 is the reduced Gibbs energy relation for a pure comp onent vapor. ( ) ( ) ( ) 3 0 3 3 2 0 2 2 0 1 0 0 3 2 r r r r r r g r r g r g r T T D T T D T T D s T h g + + + = ( ) ( ) 2 0 2 3 0 2 0 1 2 ln r r r r r r r r r T T T D T T T D T T T D ( ) + + + + 4 0 0 3 0 0 3 2 0 1 0 3 4 ln r r r r r r r r r r r r T T P T P T P C P P C P P T + + + + 12 0 3 0 11 0 3 0 11 3 4 12 0 0 11 0 0 11 3 11 12 3 11 12 r r r r r r r r r r r r r r T T P T P T P C T T P T P T P C (B.12) The reduced properties are defined by Eqs. B.13 to B.18, where the reference values are T B = 100 K, P B = 10 bar, and the gas constant R = 8.314 kJ/kmolK. The resulting reduced parameters are dimensionless. B r T T T = (B.13) B r P P P = (B.14) B r RT g g = (B.15) B r RT h h = (B.16) R s s r = (B.17) B B r RT P n n = ( B.18)

PAGE 133

114 Table B.1. Coefficients for pure water and pure ammonia (Ziegler and Trepp, 1984). Coefficient Ammonia Water A 1 3.971423 10 2 2.748796 10 2 A 2 1.790557 10 5 1.016665 10 5 A 3 1.308905 10 2 4.452025 10 3 A 4 3.752836 10 3 8.389246 10 4 B 1 1.6 34519 10 +1 1.214557 10 +1 B 2 6.508119 1.898065 B 3 1.448937 2.911966 10 1 C 1 1.049377 10 2 2.136131 10 2 C 2 8.288224 3.169291 10 +1 C 3 6.647257 10 +2 4.634611 10 +4 C 4 3.04532 10 +3 0.0 D 1 3.673647 4.019170 D 2 9.989629 10 2 5.175550 10 2 D 3 3.6 17622 10 2 1.951939 10 2 h r,0 (L) 4.878576 21.821141 h r,0 (g) 26.468873 60.965058 s r,0 (L) 1.644773 5.733498 s r,0 (g) 8.339026 13.453430 T r,0 3.2252 5.0705 P r,0 2.0000 3.0000 Fitting experimental data for ammonia and water yields the constants of Ta ble B.1, such that there are now four correlations: vapor and liquid Gibbs energy equations for pure water and pure ammonia. With this information, other thermodynamic properties can be found by differentiation of the reduced Gibbs energy equation accordin g to the Maxwell relations given in Eqs. B.19 to B.21, in reduced form. r P r r r r B T g T T RT h = 2 (B.19) r P r r T g R s = (B.20)

PAGE 134

115 r T r r B B P g P RT = n (B.21) The equations above are for pure fluids. In determining the properties of mixtures it is easier to first look at vapor mixtures, which are assumed here to behave ideally. The pure component properties are proportionally added based on the mole fraction of the more volatile species, as given in Eqs. B.22 to B.24. Note that in the ideal mixing of fluids, there is an entropic mixing term, which is given by Eq. B.25. g w g a g m h x h x h ) 1 ( ~ ~ + = (B.22) mix g w g a g m s s x s x s + + = ) 1 ( ~ ~ (B.23) g w g a g m v x v x v ) 1 ( ~ ~ + = (B.24) + = ) 1 ln( ) 1 ( ) ln( ~ ~ ~ ~ x x x x R s mix (B.25) In real liquid mixtures there are excess thermodynamic terms that are included in calculating mixture properties, which account for deviation from ideal solution behavior. A correlation for the excess Gibbs energy of liquid mixtures was proposed by Redlich and Kester, according to the three term relation in Eq B.26 (Ziegler and Trepp, 1984). + + = 2 ~ 3 ~ 2 1 ~ ) 1 2 ( ) 1 2 ( ) 1 ( x F x F F x g E r (B.26) The functions, F i are given as Eqs. B.27 to B.29 as a function of reduced pressures and temperatures. The constants are listed in Table B.2. ( ) 2 6 5 4 3 2 1 1 r r r r r T E T E T P E E P E E F + + + + + = (B.27)

PAGE 135

116 ( ) 2 12 11 10 9 8 7 2 r r r r r T E T E T P E E P E E F + + + + + = (B.28) 2 16 15 14 13 3 r r r T E T E P E E F + + + = (B.29) Table B.2. Coefficients for the ammonia water Gibbs energy functions (Ibrahim and Klein, 1993). E 1 41.733398 E 9 0.387983 E 2 0.02414 E 10 0.004772 E 3 6.702285 E 11 4.648107 E 4 0.011475 E 12 0.836376 E 5 63.608967 E 13 3.553627 E 6 62.490768 E 14 0.000904 E 7 1.761064 E 15 24.361723 E 8 0.008626 E 16 20.736547 Eq. B.26 can be then differentiated according to the previous Maxwell relations in reduced form, to determine the excess molar specific enthalpy, e ntropy and volume. Therefore, liquid mixture properties will be determined according to Eqs. B.30 to B.32. E L w L a L m h h x h x h + + = ) 1 ( ~ ~ (B.30) mix E L w L a L m s s s x s x s + + + = ) 1 ( ~ ~ (B.31) E L w L a L m x x n n n n + + = ) 1 ( ~ ~ (B.32) The state of the fluid as liquid, vapor or a liquid v apor mixture must be known in order to use the above equations proposed by Ibrahim and Klein. This can be done if the bubble and dew points of the mixture are known. Empirical correlations in Eqs. B.33 and B.34 are given by El Sayed and Tribus (1985) for a mmonia and water mixtures to determine bubble and dew point temperatures of the mixture. The constants are in Table B.3.

PAGE 136

117 Table B.3. Coefficients for determining bubble and dew point temperatures, critical temperatures, and critical pressures of ammonia wa ter mixtures (El Sayed and Tribus, 1985).* a i a[1] = 205.8889 a[2] = 280.930556 a[3] = 317.0138889 a[4] = 263.194444 A i Ai[1] = 153.17055346 Ai[2] = 11.7705687461 Ai[3] = 1.78126355957 Ai[4] = .647385455059 Ai[5] = .0719950751898 Ai[6] = .002854239 50786 A ij Aij[1][1] = 194.793913463 Aij[1][2] = 74.236124188 Aij[1][3] = 9.84103819552 Aij[1][4] = .436843852745 Aij[2][1] = 74.3508283362 Aij[2][2] = 33.2941879809 Aij[2][3] = 4.78866918581 Aij[2][4] = .225416733476 Aij[3][1] = 13.0175447367 Aij[3 ][2] = 6.1586564117 Aij[3][3] = .789740337141 Aij[3][4] = .0321510834958 Aij[4][1] = .90857587517 Aij[4][2] = .356752691147 Aij[4][3] = .0238067275502 Aij[4][4] = .00495593933952 Aij[5][1] = .00071863574153 Aij[5][2] = .0251026383533 Aij[5][3] = .0191664613304 Aij[5][4] = .0017014253867 Aij[6][1] = .00195441702983 Aij[6][2] = .00280533348937 Aij[6][3] = .0013899436563 Aij[6][4] = .000116422611616 b i b[1] = .368105523897 b[2] = 3.6679548875 b[3] = 46.6000470809 b[4] = 262.921061996 b[5] = 7 32.99536936 b[6] = 1076.0613489 b[7] = 797.948078048 b[8] = 235.903904222 C i Ci[1] = 153.634521459 Ci[2] = 13.0305543892 Ci[3] = 1.14845282991 Ci[4] = .550358094447 Ci[5] = .0753450148427 Ci[6] = .0048111666267 Ci[7] = .000120433757177 C ij Cij[1] [1] = 462.460321366 Cij[1][2] = 23739.9986309 Cij[1][3] = 194504.35292 Cij[1][4] = 639383.528867 Cij[1][5] = 523748.057636 Cij[1][6] = 2328271.47551 Cij[1][7] = 7562418.53499 Cij[1][8] = 9668295.89504 Cij[1][9] = 5922081.87086 Cij[1][10] = 143240 5.52125 Cij[2][1] = 421.443122208 Cij[2][2] = 14560.354925 Cij[2][3] = 53051.4495633 Cij[2][4] = 382763.793582 Cij[2][5] = 3583589.86875 Cij[2][6] = 12243265.3815 Cij[2][7] = 22307970.0156 Cij[2][8] = 22896656.8499 Cij[2][9] = 12483324.8091 Cij[2] [10] = 2813311.71633 Cij[3][1] = 248.783804168 Cij[3][2] = 4807.07241098 Cij[3][3] = 13565.1003309 Cij[3][4] = 466407.780832 Cij[3][5] = 2827083.44764 Cij[3][6] = 8469715.15799 Cij[3][7] = 14459588.8962 Cij[3][8] = 14281087.5331 Cij[3][9] = 759640 3.59678 Cij[3][10] = 1684002.64482 Cij[4][1] = 126.965580728 Cij[4][2] = 2090.45270574 Cij[4][3] = 1993.17101166 Cij[4][4] = 100706.510396

PAGE 137

118 Table B.3. Continued. C ij Cij[4][5] = 687388.808612 Cij[4][6] = 2132412.46959 Cij[4][7] = 3699199.65914 Cij[ 4][8] = 3688365.22546 Cij[4][9] = 1975122.39296 Cij[4][10] = 440201.446068 Cij[5][1] = 33.5343446156 Cij[5][2] = 601.878586689 Cij[5][3] = 3064.82070658 Cij[5][4] = 71.7954752052 Cij[5][5] = 51780.666659 Cij[5][6] = 209714.899856 Cij[5][7] = 40501 1.985355 Cij[5][8] = 428310.461566 Cij[5][9] = 238153.698326 Cij[5][10] = 54497.0973336 Cij[6][1] = 3.97454953787 Cij[6][2] = 77.026846469 Cij[6][3] = 541.19105807 Cij[6][4] = 1696.60270972 Cij[6][5] = 1713.45942707 Cij[6][6] = 4019.01019872 Cij[6 ][7] = 14844.7928004 Cij[6][8] = 19481.0094551 Cij[6][9] = 12107.0794501 Cij[6][10] = 2966.92804386 Cij[7][1] = .170806170177 Cij[7][2] = 3.48182859299 Cij[7][3] = 27.7957587743 Cij[7][4] = 113.762064546 Cij[7][5] = 258.750496922 Cij[7][6] = 311.0 02585218 Cij[7][7] = 123.917993454 Cij[7][8] = 123.480627492 Cij[7][9] = 154.375042114 Cij[7][10] = 48.508382870 Note that these constants in conjunction with the parent equations in chapter 2 yield bubble, dew and critical temperatures in R and cr itical pressures in psia. = = + = 7 1 10 1 ln i i c j j ij i c b P P x C C T T (B.33) ( ) [ ] = = + = 6 1 4 1 ln 0001 1 ln i i c j j ij i c d P P x A A T T (B.34) The critical temperatures and pressures of the mixture are found from the correlations in Eqs. B.35 and B.36, with the constants given in Table B.3. T cw is the criti cal temperature of water in R and P cw is the critical pressure of water in psia. = = 4 1 i i i cw c x a T T (B.35) = = 8 1 exp i i i cw c x b P P (B.36)

PAGE 138

119 In gas liquid absorption, the diffusion coefficient of the solute in the solvent and the viscosity strongly in fluence the mass transfer process. From an experimental study by Frank et al. (1996), the following correlations were found for the diffusion coefficient and liquid viscosity for ammonia and water mixtures. Ammonia is very soluble in water and so the diffu sion coefficient and viscosity can be expressed as functions of the ammonia concentration. At low ammonia mass fractions, there is large dissociation of ammonia and so the correlations do not hold for pure water. Temperature is in Kelvin, diffusion coeffic ients in m 2 /s, and dynamic viscosities in Pa s. Equations B.37 and B.38 correlate data measured between 20 and 60 C (68 and 140 F). ( ) RT e x D 15000 ~ 6 74 1 96 0 10 + = (B.37) ( ) RT e x 13300 ~ 6 64 0 52 2 10 = m (B.38) The above correlation for viscosity is not necessary for e valuation of the cycles performance as a whole, but rather for modeling the individual component processes with greater accuracy. This will eventually reflect on a better overall model for the cycle. The boiling and absorption processes require correlatio ns for viscosity, surface tension and thermal conductivity for modeling the heat transfer coefficients of the binary mixture. Correlations for these transport properties in ammonia and water mixtures are still being investigated. Some transport properties are correlated by Celata et al. (1994) and Chai et al. (1998). Ethylene Glycol and Water Mixtures The properties of ethylene glycol and water mixtures can be calculated in much the same way as ammonia and water properties, using the Gibbs excess energy met hod.

PAGE 139

120 Water properties can be taken from the pure component correlations for water presented above. Correlations for excess properties of ethylene glycol and water mixtures are determined from isothermal P x data at 60 C (140 F) and heat of mixing data at 50 C (122 F) (Villamaan, et al., 1984). The data is fit to the four parameter Margules equation. The excess enthalpy and Gibbs energy become, in general form: ~ 2 ~ 1 ~ 2 12 ~ 1 21 ~ 2 12 ~ 1 21 ~ 2 ~ 1 x x x C x C x A x A RT x x H E + + = (B.39) ~ 2 ~ 1 ~ 2 12 ~ 1 21 ~ 2 12 ~ 1 21 ~ 2 ~ 1 x x x C x C x A x A RT x x G E + + = (B.40) The constants are temperature dependent. As the Gibbs excess energy is defined as Eq. B.41, the excess entropy can be written as Eq. B.42. G H TS E E E = (B.41) ( ) ( ) ( ) ( ) ~ 2 ~ 1 ~ 2 12 12 ~ 1 21 21 ~ 2 21 12 ~ 1 21 21 ~ 2 ~ 1 x x x C C x C C x A A x A A R x x S E + + = (B.42) The Gibbs Helmholtz equation relates excess enthalpy to Gibbs energy for consta nt mole fraction according to Eq. B.43. = RT x x G dT d T RT x x H E E ~ 2 ~ 1 ~ 2 ~ 1 (B.43) From the above, the generic parameters are related as Eq. B.44. M T dM dT ij ij = (B.44)

PAGE 140

121 Assuming that the excess enthalpy is linear in temperature, the temperature dependent parameters are given as Eqs. B.45 and B.46, with the constants given in Table B.4. M M T M M T ij ij ij ij = + 0 1 2 ln (B.45) M M T M ij ij ij = + 0 2 (B.46) Table B.4. Empirical constants in ethylene glycol (2) and water (1) mixture property evaluation (Vill amanan et al., 1984). M ij 0 M ij 1 M ij 2 A 210 2059 209 = A 211 31 24356 = A 212 4 33806 = A 120 258 827 = A 121 9 50026 = A 122 1 47594 = C 210 4535 562 = C 211 82 06843 = C 212 11 80013 = C 120 1994 695 = C 121 36 25889 = C 122 5 23041 =

PAGE 141

122 APPENDIX C EXPERIMENTAL COMPONE NT LIST The major components used in the experimental investigation are listed with technical specifications as supplemental detail to the description of the system given in Chapter 3. Minor plumbing and electrical component s are not listed. Data Acquisition and Electrical DaqBook/200 Description: motherboard in data acquisition Manufacturer: Iotech, Cleveland, OH DaqView Description: software for data acquisition Manufacturer: Iotech, Cleveland, OH Specifications: vers ion 7.11.08 DBK 10 Description: chassis for DA expansion cards Manufacturer: Iotech, Cleveland, OH Specifications: 3 slots for expansion cards DBK 15 Description: DA expansion card for transducer signals Specifications: 16 channel capability; 4 20 mA input DBK 82 Description: DA expansion card for thermocouple signals Manufacturer: Iotech, Cleveland, OH Specifications: 14 channel capability Power Supply, transducers Description: DC power supply for transducers Manufacturer: Mamac Systems, Minneapo lis, MN Specifications: 115 Vac input; 24 Vdc output Thermocouple Display Description: 6 channel panel mount temperature display Manufacturer: Omega Engineering, Stamford, CT Quantity: 4 Specifications: model DP 462; accuracy 0.5 C; 1 /year shift Ther mocouple Wire Description: thermocouple signal transmission locally Manufacturer: Omega Engineering, Stamford, CT Specifications: Copper Constantin; various gauge and sheath types

PAGE 142

123 Thermocouple Wire Bundle Description: thermocouple signal transmission f rom system to DA Manufacturer: Omega Engineering, Stamford, CT Specifications: 12 channels; Copper Constantin Transducer Wire Manufacturer: Newark Electronics Specifications: various types Transducer Wire Bundle Manufacturer: Newark Electronics Speci fications: 6 channels Fluids Ammonia, anhydrous Description: primary working fluid component Manufacturer: Matheson Tri Gas, GA Ammonium Hydroxide Description: GC calibration standard Manufacturer: Spectrum Chemicals, New Brunswick, NJ Specifications : 28.76% assay Ethylene glycol Description: coolant working fluid component Supplier: Lewis Oil Co., Gainesville, FL Quantity: 55 gal Helium Description: added to absorber to boost pump performance; carrier gas in GC operation Manufacturer: BOC Gas es, NJ Specifications: Grade 5.0 purity Water Description: primary working fluid component; heat source working fluid; coolant working fluid component Specifications: tap water Heat Exchange Electrical Heater Description: electrical heating element in single pass heater Manufacturer: Watlow, Hannibal, MO Specifications: 5 kW Finned Coil Heat Exchanger Description: car evaporator for absorber coolant heat exchange Quantity: 5 Specifications: aluminum; part EA 172205 Insulation Description: Rubat ex tube and pipe insulation Plate type Heat Exchanger

PAGE 143

124 Description: recovery, boiler and vapor HE units Manufacturer: WTT America, Inc.; Bohemia, NY Quantity: 4 Specifications: nickel brazed stainless steel; operating 16 bar; 195 C; (2) model NP1 10 R; 10 plates; (1) model NP1 20R; 20 plates; (1) NP22 20R; 20 plates Instrumentation Column Description: separation of components in GC Manufacturer: HayesSep Supplier: SRI, Cleveland, OH Specifications: HayesSep T packed column; 3 x 1/8 Flow Mete r, absorber coolant Description: liquid flow rate measurement; variable area; magnetically coupled indicator Manufacturer: Brooks Instrument, Hatfield, PA Specifications: model 3809; 10.1 gpm; accuracy 5%; calibrated to s.g. 1.086, visc 15.2 cS; 1 500 psi max Flow Meter, hot water Description: liquid flow rate measurement; variable area type; brass and polysulfone Manufacturer: Meter Equipment Manufacturing, Willoughby, OH Specifications: 5 gpm; accuracy 2%; calibrated to s.g. 1.0; 300 psi max ; 250 psi at 250 F; 300 F max Flow Meter, strong solution Description: liquid flow rate measurement; variable area type Manufacturer: Brooks Instrument, Hatfield, PA Specifications: model 1110; 300 psig max; accuracy 1%; calibrated to s.g. 0.833, visc 1.08 cP; 1.11 gpm Flow Meter, vapor coolant Description: liquid flow rate measurement; variable area type Manufacturer: (1) Cole Parmer, Vernon Hills, IL Quantity: 2 Specifications: (1) 1 gpm; accuracy 2%; model F45375; (1) 5 gpm; Cole Parmer; accuracy 2%; model HLIT205PL; 3500 psi max; operating 240 F Flow Meter, vapor Description: vapor flow rate measurement; variable area type Manufacturer: Brooks Instrument, Hatfield, PA Specifications: model 1110; 100 psig max; accuracy 1%; calibrated to NH 3 at 45 F, 100 psig; 66.6 scfm NH 3 Flow Meter, weak solution Description: liquid flow rate measurement; variable area type Manufacturer: Brooks Instrument, Hatfield, PA Specifications: model 1110; 300 psig max; accuracy 1%; calibrated to s.g. 0.88

PAGE 144

125 visc 1.08 cTK; 0.84 gpm Gas Chromatograph Description: analysis of liquid and vapor samples for concentration Manufacturer: SRI Liquid Flow Indicator Description: liquid flow indication between absorber and storage tank Supplier: McMaster Carr, At lanta, GA Specifications: 20 to 446 F Liquid Level Indicator Description: liquid level indication on separator and absorber Manufacturer: Conbraco Industries, Inc., Mathews, NC Supplier: McMaster Carr, Atlanta, GA Specifications: EPDM packing for sig ht glass; 305 psig max; stainless steel; 425 F Needle, liquid sampling Description: liquid sampling with syringe Manufacturer: Hamilton Company, Reno, NV Supplier: Fisher Scientific, Pittsburgh, PA Quantity: 2 types Description: removable needle; 22 g auge; small dead volume; needle point 2; large hub for SampleLock syringe; (1) Hamilton # 80728; (22s/2/2)L; 2 length for normal use; (1) Hamilton # 160821; (22s/6/2)L; 6 length for absorber port Needle, vapor sampling Description: vapor samp ling with syringe Manufacturer: Hamilton Company, Reno, NV Supplier: Fisher Scientific, Pittsburgh, PA Description: removable needle; 22 gauge; regular dead volume; needle point 2; large hub for Gastight syringe; Hamilton # 80725; (22/2/2)L; 2 l ength Pressure Gauge Description: measurement and display of pressure Supplier: (5) Cole Parmer, Vernon Hills, IL; (4) McMaster Carr, Atlanta, GA Quantity: 9 Manufacturer: (5) Ashcroft; (4) Noshok Specifications: stainless steel; glycerine filled; (2) Ashcroft; model 68022; 0 to 200 psig; accuracy 1%; (1) Ashcroft; model 68022; 0 to 100 psig; accuracy 1%; (1) Ashcroft; model 68022; 0 to 30 psig; accuracy 1%; (1) Ashcroft; model 68022; 30 to 30 psig; accuracy 1%; (2) Noshok; model 3902K1 1; 30 to 60 psig; accuracy 1.5%; (1) Noshok; model 3902K11; 30 to 100 psig; accuracy 1.5%; (1) Noshok; model 3902K11; 30 to 200 psig; accuracy 1.5% Sampling Port Description: septum port for syringe sampling Manufacturer: Flow Design, Dallas, TX Quantity: 4

PAGE 145

126 Specifications: SuperSeal; NPT (m); operating 500 psi at 275 F; max 1000 psi at 140 F; (2) EPDM septum; 1 C to 135 C (30 F to 275 F); (2) neoprene septum; 40 C to 93 C ( 40 F to 200 F) Syringe, gas Description: gas sampling from system and insertion into GC Manufacturer: Hamilton Company, Reno, NV Supplier: Fisher Scientific, Pittsburgh, PA Specifications: model Gastight; 5 mL; plunger lock; removable needle type; 200 psig max; 10 115 C Syringe, liquid Description: liquid s ampling from system and insertion into GC Manufacturer: Hamilton Company, Reno, NV Supplier: Fisher Scientific, Pittsburgh, PA Specifications: model SampleLock; 50 m L; removable needle type Thermal Conductivity Detector Description: gas composition an alyzer used in GC Manufacturer: SRI Specification: 4 gold filaments; part 8690 0007T Thermocouple Description: temperature measurement Manufacturer: Omega Engineering, Stamford, CT Quantity: 19 Specifications: T type copper constantan; accuracy 1 C; mini quick connect Transducer Description: pressure measurement Manufacturer: Cole Parmer, Vernon Hills, IL Quantity: 5 Specifications: input 24 Vdc; output 4 20 mA; (2) model 68073; compound 30 to 60 psig; accuracy 0.13%; (1) model 68073; compound 30 to 100 psig; accuracy 0.13%; (1) model 07356; compound 30 to 60 psig; accuracy 0.4%; (1) model 68075; 0 to 250 psig; accuracy 0.25% Pumping Diaphragm Pump Description: rotary vane oil pump driven diaphragm, ammonia water pump; taken from a Servel system Manufacturer: (motor) Emerson Supplier: pump taken from Robur system Specifications: hp motor; tested in lab up to 1 gpm and 170 psig Expansion Tank Description: dampening of pump pulses with air filled bladder Manufacturer: Amtrol, I nc., West Warwick, RI Specifications: Therm X Trol Thermal Expansion Absorber; ST5C 300WP; 300 psig max; 200 F max; stainless steel; heavy duty butyl diaphragm; polypropylene liner

PAGE 146

127 Centrifugal Pump, chiller coolant Description: coolant circulation to chiller Manufacturer: (motor) Franklin Electric, Bluffton, IN; (pump) AMT, Royetsford, PA Specifications: 1/3 hp motor; 20 to 210 F Centrifugal Pump, absorber coolant Description: coolant circulation to absorber Manufacturer: (pump) Sherwood, Hy pro Corp. Specifications: hp motor; pump 201J Centrifugal Pump, vapor coolant Description: coolant circulation to vapor HE Manufacturer: (motor) Franklin Electric, Bluffton, IN; (pump) AMT, Royetsford, PA Specifications: 1/3 hp motor; 20 to 210 F Centrifugal Pump, hot water Description: hot water circulation Manufacturer: Finish Thompson, Inc., Erie, PA Supplier: McMaster Carr, Atlanta, GA Specifications: 300 F max; hp motor; 316 stainless steel; Viton O ring Rotary Vane Pump Head Descript ion: used in initial system trials Manufacturer: Berns Corp, Gurnee, IL Supplier: Cole Parmer, Vernon Hills, IL Specifications: carbon graphite vanes and ring; stainless steel; 240 psig max; 170 F max; 34 gph avg Safety Eyewash Station Description: Emergency eyewash station Manufacturer: Bradley Corp, Menomonee Falls, WI Goggles Description chemical spill resistant goggles Manufacturer: American AllSafe Co. Supplier: Fisher Scientific, Pittsburgh, PA Specifications: 211 style Respirator Descri ption: full face respirator for emergencies and maintenance Manufacturer: 3M Supplier: Fisher Scientific, Pittsburgh, PA Specifications: 600 series Respirator Cartridges Description: Ammonia filtration used with respirator Manufacturer: 3M Quantity: 2 Specifications: Combination Cartridge/P100 Filter

PAGE 147

128 Valves Ball Valve, system side Manufacturer: Swagelok Supplier: Jax Valve and Fitting, Jacksonville, FL Quantity: 2 Specifications: 3/8 pressure fit connections; 316 stainless steel Check valve De scription: for ammonia charging to the storage tank Manufacturer: Swagelok Supplier: Jax Valve and Fitting, Jacksonville, FL Specifications: 3/8 pressure fit connections; 316 stainless steel Mixing Valve Description: regulate the source temperature to the system boiler Manufacturer: Leonard Valve Co., Cranston, RI Needle Valve, system side Manufacturer: Swagelok Supplier: Jax Valve and Fitting, Jacksonville, FL Quantity: 6 Specifications: 3/8 pressure fit connections; 316 stainless steel Plug val ve, hot water side Manufacturer: Swagelok Supplier: Jax Valve and Fitting, Jacksonville, FL Quantity: 2 Specifications: pressure fit connections; silicone lubricated neoprene O rings; 316 stainless steel Plug Valve, system side Manufacturer: Swa gelok Supplier: Jax Valve and Fitting, Jacksonville, FL Quantity: 25 Specifications: Krytox grease lubricated Kalrez O rings; 316 stainless steel; (24) 3/8 pressure fit connections; (1) pressure fit connections Pressure Regulator, Ammonia Manufa cturer: Matheson Tri Gas, GA Specifications: Model 3141 705; low 60 psi, high 3000 psi Pressure Regulator, Helium Manufacturer: Controls Corporation of America, Virginia Beach, VA Specifications: 400 series; low 200 psi, high 4000 psi; 375 psi relief Re lief Valve, absorber Description: pressure relief valve on absorber tank Manufacturer: Swagelok Supplier: Jax Valve and Fitting, Jacksonville, FL Specifications: 10 225 psi range; 40 psi setting; SS RL3M4S4 NE; stainless with neoprene seal Relief Va lve, hot water Description: pressure relief valve on hot water storage tank

PAGE 148

129 Manufacturer: Cash Acme; Cullman, AB Specifications: set for 150 psig; 50 150 psi range; max 800 psi Relief Valve, system high Description: pressure relief valve at high pressu re location in system Manufacturer: Swagelok Supplier: Jax Valve and Fitting, Jacksonville, FL Description: set for 220 psi; 50 350 psi range; SS 4R3A1 NE; stainless with neoprene seal Vessels Gasket, absorber Description: gasket for flange on abso rber Manufacturer: Blaylock Gasket and Packing Co., Haltom, TX Specifications: Thermoseal Tank, absorber Description: Mild steel flanged top absorber tank Manufacturer: Dal Worth Fabrication, Inc., Grand Prairie, TX Specifications: 30 psig max at 120 F; 20 F min at 30 psig; mild steel Tank, emergency Specifications: 200 psig max; carbon steel; 80 gal Tank, hot water Description: domestic hot water heater used as hot water storage tank Manufacturer: US/Craftmaster Water Heater Co., Johnson City, TN Specifications: 400 F max; 300 psig max; 150 psig operating Tank, separator & storage Description: storage tank and separator Specifications: 200 psig max; carbon steel; 30 gal

PAGE 149

130 APPENDIX D EXPERIMENTAL PROCEDU RES The information in this appendix is a detailed account of the laboratory procedures taken before, during and after experimentation, given as a supplemental section to the experimental chapters. This summary documents the methods used in calibrating the equipment, provides suggested operating procedures for the system, and lists practical precautions in dealing with ammonia water mixtures and the system in general. Calibration of Instrumentation Although most instruments a re either factory calibrated or come with a rated uncertainty, it is suggested that the operator calibrates instruments manually. This has the dual purpose of verifying the rated accuracy of the manufacturer, and to reduce the error in the instrument by hi gher order calibration. Thermocouple Calibration The change in thermoelectric voltage produced at the junction of a thermocouple is expected to be linear with temperature change. Therefore, only two temperatures are needed in the calibration of a thermocou ple. For the T type thermocouples used in the experiment, the freezing and boiling points of water at atmospheric pressure were used as they are well known and easily achieved in the lab. The voltage signal of a thermocouple can be affected across the long distance of the wire, so the calibration was performed end to end to calibrate the entire pathway from the thermocouple to the data acquisition board. An ice bath was used for the low

PAGE 150

131 temperature in the calibration, making sure the thermocouples did not r est against the walls of the Dewer flask. Boiling water in a conventional hot water beverage heater was used for the high temperature in the calibration. The thermocouples in the boiling water were maintained away from the heater surface. After a minute or two, the temperatures obtained by the data acquisition system were recorded at a rate of 25 Hz for 10 seconds, and averaged for each thermocouple. The correlation obtained for transforming measured values to actual values is given in Eq. D.1. T H,c and T L, c are the high and low standard temperatures used, here 100 C (212 F) and 0 C (32 F), respectively. T H,m and T L,m are the measured temperatures while exposed to the standard temperatures. These are specific to each thermocouple. ( ) c L m L m m L m H c L c H T T T T T T T T , , , + = (D.1) Pressure Transducer Calibration The pressure transducers used here yield a 4 20 mA output over the range of operation, whi ch varies per transducer. The data acquisition system converts the output with 250 resistors, for a corresponding 1 5 V response. The full range voltage response corresponds to the full range of pressures for each transducer, according to Eq. D.2, which correlates the voltage read from the data acquisition system to the pressure it measures. ( ) min min min max min max P V V V V P P P m + = (D.2) As the transducer response is linear to the pressure change, the calibration should be done at two points as with the thermocouples. Ho wever, if a pressure measurement with higher accuracy than that of the transducers is not available for the high pressure

PAGE 151

132 standard, then a one point calibration method is selected. Atmospheric pressure is used as the single calibration standard, and the tr ansducer readings are simply shifted in calibrating them. ( ) m L c L m P P P P , + = (D.3) The one point calibration method was used in this work although the two point calibration method is outlined for when a higher accuracy pressure measurement device bec omes available. The transducers are then calibrated over the full length of the signal, end to end, from transducer to data acquisition board. The low pressure tested is 1.0 bar (0 psig), with the transducers open to the atmosphere. The high pressure is se lected near the upper limit of most of the transducers. The high pressure can be obtained with pressurized air and verified using the high accuracy pressure gauge. Note that the system as a whole cannot withstand pressures over 3.1 bar (30 psig) die to the absorber pressure rating, which would limit the calibration range, so a portable pressure vessel is used at each transducer to obtain the high calibration pressure. The correlation between the measured pressure and the actual pressure is then given by Eq. D.4 which includes transducer specific calibration constants (P H,m and P L,m ) found at the high and low calibration pressures (P H,c and P L,c ). These are averaged values of readings taken at 25 Hz for 10 seconds once the pressure at the transducer is steady As the data acquisition system gives the transducer output in volts, combining Eq. D.2 and D.4 yields a correlation between the voltage response and the actual calibrated pressure, as given in Eq. D.5. ( ) c L m L m m L m H c L c H P P P P P P P P , , , + = (D.4)

PAGE 152

133 ( ) c L m L m m L m H c L c H P P P V V V V P P P P P P P , min min min max min max , , + + = (D.5) Gas Chromatograph Calibration The gas chromatograph (GC) uses a heated packed column to vaporize and separate the ammonia and water constituents of the sample that is inserted into it. The components pass through the column with different residen ce times, and are analyzed by a thermal conductivity detector (TCD). The TCD measures the thermal conductivity change of the helium carrier gas that is mixed with the sample relative to the carrier gas alone. The signal is given as a change in voltage due to a change in resistance of the TCD filaments cooled by the gas. Plotting voltage over time will yield peaks under which the area is proportional to the amount of the sample gas component. Provided the sample components have thermal conductivities very di fferent from the carrier gas, and separate out well in the column, distinct peaks will be shown for each component. For the ammonia and water mixture, the ratio of the area under the ammonia peak to the total area under all of the peaks is given in Eq. D.6 as a percentage. O H NH NH NH A A A A 2 3 3 3 100 % + = (D.6) The area of each component is proportional to the mass fraction of that species in the sample, according to Eqs. D.7 and D.8, and scales with the sample size, a. The constants on the right hand side of the e quations, A, can be found for a unit sample size of pure water and pure ammonia. x A a A x NH NH 1 3 3 = = (D.7) ( ) x A a A x O H O H = = 1 0 2 2 (D.8)

PAGE 153

134 Using the above equations as calibration curves is inconvenient, requiring exact measurement of every sample size and two calibration points to determine the constants. Eq. D.6 can be written by combining it with Eqs. D.7 and D.8, into the form given by Eq. D.9 as the general calibration curve. The constant, C, is evaluated at a single calibration point, using a know n sample of mass fraction x c The constant is given as Eq. D.10. ( ) 1 1 100 % 3 + = x C Cx A NH (D.9) = = = = = = c c x x NH x x NH c c x O H x NH A A x x A A C 3 3 2 3 % 100 % 1 0 1 (D.10) The standard mixture was ordered from a chemical supply company, with an assay sheet describing its precise composition. The sta ndard mixture was refrigerated to prevent volatilization from the container. A sample was drawn through a septum in the container lid with a cooled L syringe, according to the syringe sampling procedures given in this appendix. The GC results of several s amples of the standard mixture were averaged to determine the calibration constant in Eq. D.9. The obvious advantage of having an area percentage in the calibration curve rather than the area itself is to take sample size out of the correlation provided th e sample size is within limits of the column and TCD performance. With the calibration constant found from Eq. D.10, the ammonia mass fraction of any sample can be determined from the NH 3 area percentage given by the GC analysis according to Eq. D.9. Flow Meter Calibration The flow meters are all variable area type, calibrated by the manufacturers for a fluid with a known specific gravity. The specific gravity of the calibration fluid is not necessarily that of the working fluid, which has to be accounted f or.

PAGE 154

135 If the fluid is not highly viscous, the force balance on the float in the flow meter is approximated as Eq. D.11, which gives the flow rate as a function of float and working fluid densities. The forces on the float are buoyancy, gravity and fluid pres sure. ( ) 2 1 2 1 = r r r f f f gV A Q (D.11) For gases, the float density is much greater than the working fluid density, such that the flow rate is proportional to the square root of the inverse of the working fluid density. A comparison can be made between the measured flow rate and the actual flow rate, which accounts for the calibration of the flow meter. Eq. D.12 gives this correlation, written in terms of pressures and temperatures instead of densities according to the ideal gas law. In Eq. D.12, the calibra tion temperatures are those used for the vapor flow meter, listed on the vapor flow meter as 7.2 C (45 F) and 7.9 bar (100 psig). Ammonia is not a vapor at these conditions although the calibration fluid used by the manufacturer is. Note that absolute un its must be used in Eq. D.12. 2 1 2 1 = c c m P P T T Q Q (D.12) Often the vapor flow meter gives readings in scfm, which must be converted to actual cfm as the working fluid temperature and pressure are not necessarily at standard conditions. The vapor flow meter used here gives readings in scfm, so a correction is applied to Eq. D.12 to account for the difference in flow rate, as given in Eq. D.13. The standard conditions are 21.1 C (70 F) and 1.0 bar (0 psig). Note that absolute units must be used in Eq. D.13.

PAGE 155

136 2 1 2 1 = c c s s m P P T T P P T T Q Q (D.13) For liquid flow meters, the corrections to account for different fluids are given in terms of densities. Writing Eq. D.11 for the calibration fluid and for the actual fluid and dividing the two relationships yields Eq D.14. 2 1 2 1 1 1 = f c f c m Q Q r r r r r r (D.14) The ratio of the actual fluid to float density is approximately the same as the ratio of the calibration fluid to float density, so the relation reduces to Eq. D.15 for determining the actual flow rate in a liquid fl ow meter. The calibration fluid density is usually given on the flow meter as a specific gravity. 2 1 = r r c m Q Q (D.15) Syringe Sampling Techniques The GC can determine ammonia mass fractions of both liquid and vapor samples that are inserted int o its injection port via syringe. The liquid sample is vaporized while it passes through the packed column, such that the sample is a vapor as it reaches the TCD for analysis. Because the liquid sample specific volume is much smaller than that for a vapor sample, the liquid sample size is correspondingly smaller. Experimentally it has been determined that a liquid sample size of about 2 5 L and a vapor sample size of 2 3 mL are sufficient to detect peaks. If the sample size is much larger, the ammonia and water peaks may not separate out well.

PAGE 156

137 The temperature of the syringe is important for two reasons. A false sample may be drawn if the needle tip is below the dew point in vapor sampling or above the bubble point in liquid sampling. This unwanted phase change at the needle tip will influence the composition of the sample being drawn. Secondly, it is desirable to avoid a liquid vapor mixtu re in the sample, as during insertion not all of the liquid or vapor will be inserted. Since the liquid and vapor will not contain the same ammonia mass fraction, the sample that is inserted may not correctly reflect the sample that was drawn. For liquid samples, the syringe should thus be cooled prior to use to ensure the sample remains cooled below the bubble point during drawing. Note under normal operation, both liquid sampling ports in the system are at subcooled fluid regions. The syringe should be p umped a few times during drawing, to replace any dead volume and water that was used for cleaning it with the desired sample. Take care not to allow the system pressure to force the plunger out of the syringe, or the system will leak through the syringe. T he syringe is then locked, removed from the sampling port, and transported to the GC. After complete insertion into the GC, the syringe is unlocked and the contents injected quickly. The syringe is removed and cleaned with tap water to prevent corrosion, t hen cooled for the next use. For vapor samples, the syringe is heated prior to use under a hot air gun or in an oven, to ensure the syringe remains above the dew point of the sample and so that all excess ammonia and water has been vaporized from the syrin ge. Note that the vapor sampling port in the system is at a superheated vapor region. Removing the plunger from the syringe during the heating will allow for the contaminant vapors to escape. A thermocouple placed next to the syringe can estimate its tempe rature, provided the

PAGE 157

138 syringe has been heated for a few minutes. Typically with this system the syringe is heated to 60 80 C (140 176 F), and it will be hot to the touch. When ready to sample, the syringe is inserted into the sampling port while hot and s teadily draws the vapor sample. The syringe is not pumped as with liquid sampling, as the dead volume of air in the syringe is not significant. Under normal operation, the system pressure will fill the syringe so no external pulling is required on the pist on. The sample will become more superheated as the syringe pressure will be less than the system pressure. Do not compress the sample too much, as vapor may condense on the tube walls and this will have to be vaporized by heating it prior to GC injection o r a new sample will need to be drawn. Lock the syringe for transport, and unlock it after the needle is inserted completely into the GC injection port. The contents should be injected steadily, preventing condensation on the syringe tube walls. The needle is then removed, opened and heated for the next use. Observations on the Behavior of Ammonia Water Mixtures The behavior of a fluid that boils over a range of temperatures is counterintuitive, and especially if this range crosses over room temperature. Thi s becomes a consideration in most of the experimental procedures, and is stressed to promote safety and proper experimentation. The emphasis of this section is to expect certain behavior, and to be prepared to best handle the unexpected. A key observation that is physically simple but not immediately obvious is that ammonia will want to be in vapor form, over the entire temperature range in the system design. The vapor formation is controlled with appropriate pressures, temperatures and dilution. As the tem perature rises, the pressure in a closed region will rise, owing to the

PAGE 158

139 vaporization of more fluid. This should be considered when pressure limitations or pressure driven flow are a concern. Likewise, if the pressure drops then more of the fluid will boil. Thus if an external valve is opened or a leak occurs, the ammonia water mixture will flash boil, releasing a high concentration of toxic ammonia vapor. It is suggested to consult a bubble and dew point diagram, or properties program before any experimenta l procedures are performed. Normal Operating Procedures Listed here is a guide for warming up the system, operation and testing, and shut down. Preparation for each step and anticipation of expected results will develop safe and successful experimentation. Before proceeding, it is suggested that the design of the experiment to be run is completed with the data analysis and design program. Startup The hot water heating is the most time consuming of the startup procedures, taking several hours to fully heat t he phase change material in the single pass boiler and heat the water in the hot water tank. It is recommended that the single pass boiler be preheated already a day in advance according to the following guidelines: 1. If the hot water side of the system has been drained in light of a freeze warning, fill the tanks and components by cracking valves and fittings to expel all air. The hot water pump must be flooded to prevent damaging it. Maintain the hot water side pressure by leaving the water supply open to i t. This will allow for higher temperatures to be reached before the boiling point. 2. Set the single pass boiler thermostat to about 107 C (225 F). The actual temperature will be 6 11 C (10 20 F) higher as the dial is offset. 3. Allow the phase change materi al to heat for several hours or overnight. 4. Circulate the hot water through the single pass boiler with the hot water circulation pump, making sure the valves are set for the desired loop.

PAGE 159

140 5. Monitor the hot water storage tank temperature with the thermocouple after the hot water circulation pump. 6. Do not exceed the hot water circulation pump temperature limit of 149 C (300 F). 7. Do not operate the hot water pump at temperatures over 116 C (240 F) for excessive periods. It will make a slight grinding sound at this point, possibly from thermal expansion. 8. Turn the hot water circulation pump off when the desired hot water storage tank temperature is reached. The heating of the hot water in the storage tank should take no more than 1 hour once the single pass boile r is preheated. 9. Close the valves for the hot water circulation loop and open the valves for the hot water supply loop. 10. The single pass boiler heater can remain on indefinitely. Note the hot water side has a pressure relief valve set for 11.4 bar (150 psig) which will not be achieved unless the water is heated to 185 C (365 F). In the event of the relief valve cracking, turn off the heater and allow the cold water supply to fill the hot water supply and cool the components. The pressure will drop and the relief valve will reset itself. The coolant side of the system can be cooled within an hour, according to the following guidelines: 1. All valves should be set correctly as they are not adjusted in normal operation, except for the absorber and vapor HE coolan t supply valves. However, make sure all coolant flow loops are open. 2. Remove the chiller control box cover while the power is off and set the thermostat to the desired setting, typically around 10 21 C (50 70 F). The temperature should remain within 6 C (11 F) of this. The suggested temperature of the coolant is 15 20 C (27 36 F) cooler than the absorber temperature, for the current absorber design. Replace the cover. 3. Turn on the chiller circulation pump. 4. Turn on the chiller main power. Make sure that the circulation pump is on first.

PAGE 160

141 Occasionally the screen at the system pump inlet needs to be purged of rust and other particles. Provided there is positive pressure in the system, the valve can be cracked quickly to purge the collected debris. Wearing th e respirator is recommended, and allows for viewing of the purged spray to see if there were particles in it. Standing away from the line of the spray is suggested also, or behind a shield to prevent ammonia from getting on footwear and clothing. The spray ed area should then be drenched with water from a hose and the surrounding air diluted with the overhead showers before removing the respirator. The GC needs about 1 1 hours to warm up, such that the baseline no longer drifts. The GC startup is as follows : 1. Open the oven to check that the correct column is installed. 2. The helium supply is opened and allowed to go through the GC. The pressure should be preset at 4.5 bar (50 psig) and the flow rate through the GC at 200, corresponding to 6.4 mL/min. Note the p ressure leaving the regulator is at 5.1 bar (60 psig) for this. The head pressure at the column inlet is around 0.3 bar (4 psi) for this flow rate. Always make sure there is helium flowing through the GC and TCD while the heater is on. 3. The computer that in terfaces with the GC is turned on, and a default control file is loaded in the PeakSimple program, which boots automatically. The control file cycle.con is used for the cycle, and must be loaded manually once PeakSimple is started. 4. The GC main power is t urned on. 5. The filter bake switch is turned on. 6. The TCD current switch is set to high, and the attenuation knob to 1. 7. The rest should set itself according to the cycle.con file. It is recommended the set points and actual readings be checked to prevent burn out. The TCD should be set to 95 C (203 F), and the oven to 80 C (176 F). The sampling time is set at 25 minutes, although after 20 minutes both components should have passed through the GC. 8. The baseline is set to zero and monitored for drift.

PAGE 161

142 On the d ata acquisition side, make all connections and boot the DaqView program according to the data acquisition discussion above. Operation and Testing As the behavior of ammonia and water mixtures is counterintuitive, the system operation should be approached cautiously, especially if it is unfamiliar to the operator. Basic warnings are to stay within the pressure limits of components, increasing pressures gradually. Before changing a valve setting it should be made clear that the valve does not open to the atm osphere or block the pump circulation. Although adjustments are made constantly for the greatest control, the operator should be ready for unexpected behavior and know how to deal with it. 1. The system operation should begin with the mixture in the absorber within the level of the sight glass. If it is not, use the system pump or pressure driven flow to return the mixture to the absorber. 2. The mixture temperature in the absorber should begin at or below the desired absorber temperature. If it is above it, cool it first by circulation over the absorber cooling coils. 3. If the pressure in the absorber is below 1.4 bar (5 psig), it may be necessary to raise the system pressure by circulating hot water through the system boiler and boiling some of the mixture. The sy stem pump will not work well below 1.4 bar (5 psig). 4. Assuming that all is set, the absorber and vapor HE coolant pumps are turned on and set to approximate flow rates. If the coolant is at its coldest, as in right after the chiller shuts off, then the cool ant flow rate to the absorber should be set at a low range value near 1 gpm. Otherwise, a middle range flow rate value of 3 gpm is a good starting point. 5. Check to make sure all the valves are set correctly in the system and hot water side. 6. Turn on the hot water supply pump. 7. Restrict the separator vapor and weak solution exit lines to little or no flow, to build up pressure in the separator.

PAGE 162

143 8. Turn on the system pump. 9. Adjust the vapor and weak line valves to set the desired system high pressure. 10. Adjust the str ong solution absorber return valve to control the flow rate of the strong solution, to control the temperature of the mixture leaving the system boiler and the heat source flow ratio. 11. Adjust the coolant supply valves to control the amount of cooling to the vapor HE and the absorber. Set a high flow rate to the absorber if the coolant temperature is at its highest, and a low flow rate if the temperature is at its lowest, according to the chiller cycle. 12. Adjust the weak solution valve to maintain the liquid le vel in the absorber, once some liquid is allowed to build in the separator tank bottom. 13. Adjust the temperature setting on the mixing valve to obtain the desired source temperature. 14. Maintain this for 30 60 minutes, allowing the system components to heat up such that there is minimal condensation in the vapor lines and none in the vapor flow meter. The system should then be relatively steady. Provided that the system becomes steady for the desired operating conditions, the data can then be recorded. Note that for additional trials, the time to steady state is reduced as the system is already warmed up. 1. The vapor line is sampled and the sample is injected into the GC for analysis, according to the outlined syringe sampling procedures. The weak solution line can be sampled and the sample placed into refrigeration while locked in the syringe, or this sampling can occur 20 minutes later when the GC is ready for another sample if the system is deemed steady for those 20 minutes. The strong solution can be sampled an d stored likewise before analysis. 2. The flow rates are recorded manually with repetition, along with any observations. 3. The temperatures and pressures are recorded into a data file from the DaqView window. After the weak and strong solution samples are also injected for analysis, the system parameters can be changed for the next configuration. The system should be steady by the end of the GC analysis of the previous sample or shortly afterwards.

PAGE 163

144 Shutdown The ammonia water system is shut down first, as it req uires the greatest monitoring otherwise. The hot water and coolant elements are shut down next, followed by the GC. The following procedures are used: 1. Turn off the system pump. 2. Close the strong solution line leaving the pump to prevent possible back flow t hrough the pump that could raise the absorber pressure. This back flow has not been witnessed, but is an added precaution. 3. Leave the vapor and weak solution lines cracked. This is to have the separator pressure drive most fluid back into the absorber acros s the cooling elements, which will lower the temperature and pressure of the absorber. The vapor line needs to remain cracked also, or fluid will build up in the vapor flow meter when it later condenses after cooling. Note that if the weak solution line is closed, a siphoning effect will occur through the vapor line as the absorber pressure increases past that in the separator perhaps during the next warm day. 4. Turn off the hot water supply pump to the system. 5. Leave the single pass boiler on. 6. After the fluid has all been returned to the absorber, turn off the coolant supply pumps. 7. Turn off the main power supply to the chiller. 8. Turn off the chiller circulation pump. 9. Turn off the thermocouple displays and unplug them as prevention against lightning damage. 10. All pumps should now be off and only the single pass boiler will be on. 11. Exit the DaqView program and unplug all connections to the data acquisition equipment as prevention against lightning damage. 12. On the GC, turn off the TCD current and the filter bake switc h. 13. Quit the PeakSimple program and turn off the GC power and the computer. 14. Close the helium cylinder valve and the valve leaving the regulator.

PAGE 164

145 15. Rinse the liquid syringes in tap water and place into the refrigerator. Pull out the plunger in the vapor syrin ge and set aside to dry. System Maintenance and Initial Setup Procedures for the initial setup and subsequent maintenance of the system are outlined below. Observations are noted based on work that has been performed to facilitate further similar work. The se procedures are not preformed while the system is in normal operation. Data Acquisition Setup The data acquisition system hardware consists of a DaqBook/100 motherboard and analog inputs from a DBK19 thermocouple expansion card and a DBK15 universal v oltage/current expansion card. In later operation, a DaqBook/200 replaced the 100 series and a DBK82 with greater noise reduction replaced the DBK19 The thermocouples are wired to the DBK82 card with no external power supply. A 15 V dc supply is provi ded to the pressure transducers, which are connected to the DBK15 card. The jumpers on the motherboard are maintained in their factory default settings, as described in the literature. The DBK15 card requires a short to be set up for the RA resistors and the inclusion of 250 resistors in the RB positions. This will provide a 1 5 V signal from the 4 20 mA output of the transducers. Care must be taken to observe the polarities of thermocouples and transducers connected to the cards. The equipment should b e surge protected as lightning can destroy expensive component circuitry. With the hardware having been set up, it needs to be recognized by the PC by adjusting certain settings. Procedures for setting up the DaqView software are also listed below.

PAGE 165

146 1. Instal l the DaqBook support software from the supplied disk. The latest version of DaqView installed is version 7.11.08. 2. From the PC Settings, enter the Control Panel window and select the Daq Configuration icon. Choose Add Device and include the DaqBo ok under the named device, DaqBook0. With the DaqBook powered up and connected to the PC, choose to Test Hardware. 3. Start DaqView Note that the DaqBook must be powered up and connected. 4. In the Device menu, go to the Select Device window. Choose the appropriate device, DaqBook/0. Make sure it is set for the correct DaqBook. 5. In the Device menu, go to the Hardware Configuration window. Assign the correct expansion cards to their channels. The channels should correspond to the jumper settings on the expansion cards. 6. In the Channel Setup window, set the thermocouple type as T and the transducer gain as x1 if 250 resistors were used on the DBK15 Set the polarity as Bipolar for both thermocouples and transducers, and label the channel s as desired. 7. For conversion of voltages to pressures, adjust the mx+b setting under Units. Note that calibration values are also applied here for temperatures and pressures such that calibrated values are sent to file. 8. Turn off the channels not in us e. Leave on the CJC cold and hot reference junctions for the thermocouple card. 9. The DBK19 card comes with a calibration file for it, which can be referenced through the Preferences menu. This does not provide calibration for individual thermocouples, but for the card itself. The DBK82 has internal calibration. 10. In the Acquisition Setup window, enter preferences. In this work the pre trigger scans were set at 0, such that recording would start immediately upon triggering. The trigger event was set at Manual Trigger, and the stop event at Number of Scans, which was set at 250. The scan rate was set at 25 Hz, with 10 scans used for each average shown. 11. In the Data Destination window, select the destination and file type for recording. The data analy sis program that is used requires a text file to be produced from DaqView The file should be named MMDDYYL, where the first six spaces denote the date and the last an alphabetical letter starting with a to sequence the files recorded on that date. Ch oose to delete the source file after recording and to ask the user before doing so.

PAGE 166

147 12. Save all the DaqView settings as a *.daq file. The cycle configuration is saved as cycle.daq. The current configuration is always saved as DaqView.daq, which boots u p when DaqView is started. The acquisition is very simple and the real time display is convenient. Real time signals are smoothed by displaying the averages of several readings. Operation of the data acquisition system consists of the following steps. 1. Con nect and turn on the power to the DaqBook, and connect the power supply to the transducers. Connect the transducer and thermocouple quick connects, and the parallel port cable to the PC. Note that all of these connections are undone after operation to pre vent possible lightning damage. 2. Boot the DaqView program. 3. Enter a filename and type to save data to in the Data Destination window. Data recording can then be triggered. 4. To record to another file on the next trial, be sure to rename the file destination or previous data may be lost. Note that the data acquisition system does not record flow rates or ammonia mass fractions from the system. These values must be recorded by hand and can later be inputted to the data analysis program manually. System Drainin g and Cleaning For quick repairs or system modifications, the working fluid in the system does not have to be displaced from the work area. The pressure in the system need only be lowered through cooling to slightly above ambient. The slight positive press ure is best if too low for monitoring by gauges but can be determined by placing a fingertip at a vent and opening the valve. Any higher pressure will cause too much loss of working fluid. The positive pressure will allow some escape while repairs are made which if done carefully will not be much. A respirator should be worn at this time.

PAGE 167

148 For more involved repairs or system modifications, the working fluid needs to be drained. If the work can be isolated, the fluid need only be pumped or pressure driven to the absorber or the separator. If the entire system needs to be drained, the fluid is stored in the storage tank. Any fluid that is not drained is typically high concentration ammonia vapor and will be discarded. To minimize wasting ammonia, the absorber can be chilled to liquefy as much vapor as possible before draining. The liquid can then be pumped to the separator or storage tank as needed, or pressure driven if the potential allows. Without chilling, the pressure in the system rests at 1.7 2.4 bar (10 20 psig) for the approximate 40% ammonia mass fraction and atmospheric temperatures. With chilling, this pressure reduces to 1.0 1.4 bar (0 5 psig). Note that even a few psig of ammonia vapor will be extremely potent if released into the air. The vapor mu st be purged by absorption into a larger body of water such as the solar pond, and allowed to vaporize over time into the air. If the entire system is to be purged then the fluid is moved to the storage tank and isolated by way of valves. With all valves t o the ambient still closed, a pressurized water hose is attached to a valve near the separator. A drain hose to the solar pond is attached to the absorber drain valve. The hose is attached to the high pressure end of the system for safety as the water pres sure at 5.8 6.5 bar (70 80 psig) can exceed the absorber pressure rating. The water supply valve is opened, and allowed to begin filling the system. Note the water will very quickly absorb the mostly ammonia vapor, dropping the system pressure well below a tmospheric. Close valves to transducers or gauges that have lower limits that might be compromised. Also keep the valve closed to the drain hose, or risk drawing dirty solar pond water into the system.

PAGE 168

149 As the water reaches the absorber, the absorber sight glass will indicate the water level. By this time there is enough water in the system such that the overall system concentration is very low and the absorber vent can be opened. The additional head from the absorber water level will allow for the drain val ve to be opened and the system to drain to the solar pond. Be very careful not to allow the water to fill up all of the system while no vent is open, as the pressure may rise very fast to the water supply pressure. Although the system is flooded, the effl uent to the solar pond will have a strong ammonia odor. The system must be flooded for a while, directing the water stream through all parts of the system. Even then, pockets of ammonia will still be trapped in parts of the system such as in the pump or tr ansducers, so caution must be used. After the system is cleaned, rusting becomes a concern. If the system is filled completely with water and no oxygen, this will inhibit rusting. Care must be taken while filling the system if there is no open drain as the water pressure will rise very fast and can exceed component limits. Better yet is to dry the system with pressurized air as much as possible, then to evacuate it with a vacuum pump. This is the recommended procedure, especially if the system will not be u sed or charged for a while. System Charging For the initial charging of the system, there is no ammonia or water in either the system or the storage tank. The system and storage tank are assumed to be cleaned and under a vacuum. The mixture is produced in the storage tank first from the ammonia and water components, then pumped or pressure driven to the system. Isolating the storage tank from the system, the water is first added in the quantity desired to produce the desired ammonia mass fraction. Approxima tely 30 38 liters (8 10 gallons) of the mixture is desired, in order to fill the absorber to within the sight glass

PAGE 169

150 range. It will be important to have the level visible in the absorber during normal operation to ensure the pump is indeed pumping liquid an d the vapor is bubbling into the liquid pool. The water is added through the water holding tank above the storage tank. Water is added to the holding tank and drained to the ambient until the water level is at the marked position. The desired amount of wat er to charge the system is added into the water holding tank, with the ambient drain closed. The feed to the storage tank is opened, noting that the liquid will drain very fast as the storage tank is under vacuum. The water is drained until the water level reaches the marked position, making sure that it does not fall below it far enough to allow air into the storage tank, as this will corrupt the charge. The water feed to the storage tank is then closed, and the desired amount of water should be in the sto rage tank. The pressure of the storage tank should be at the saturation pressure for water at the water temperature. Next the ammonia is to be added in the desired amount, noting the approximate mass transfer by placing the ammonia cylinder on a scale duri ng the process. As the cylinder is stored outside, the ammonia is at saturation pressure at the ambient temperature while in the cylinder, at around 5.1 13.4 bar (60 180 psig) depending on the day. The bulk of the ammonia is liquid in the cylinder, and wil l vaporize as it is fed into the storage tank. The vaporization in the cylinder will extract a tremendous amount of heat from the ammonia that is liquid, cooling the cylinder to well below 0 C (32 F). The cylinder saturation pressure drops along with thi s, reducing the pressure driven flow to the storage tank. The cylinder should therefore be heated, either with hot air guns or by submersion in a heated bath.

PAGE 170

151 The ammonia will be absorbed into the water as it enters the storage tank. This process is made m ore efficient by circulation of the water through the system pump and in from the storage tank upper inlet. If not absorbed, the pressure in the storage tank will quickly rise. A byproduct of the absorption is the heating of the storage tank. This heat nee ds to be removed, to allow for quicker absorption. This can be accomplished by pouring tap water over the storage tank or by circulation of the storage tank fluid across the absorber cooling coils. Note that the storage tank will not be in danger as its pr essure limit is higher than the saturation pressures in the ammonia cylinder. However, the pressure in the cylinder must be maintained higher than in the storage tank, to prevent any back flow or diffusion of water into the cylinder, which could contaminat e it. Idling in Cold Weather Cold weather precautions must be taken with the system only a few days out of the year, in the event of freeze warnings. The 40% ammonia and water mixture will not freeze with a FL winter, but the hot water system might while i dling. To prevent this, the hot water system should be thoroughly drained or circulated during low temperature periods, with the pressurized supply lines drained. Idling in Hot Weather Hot weather will raise the temperature of the ammonia and water mixture in the system while idling, and can therefore raise the saturation pressure beyond the limits of the absorber tank. This can easily occur with the ammonia mass fraction over 40%. In the event of hot weather, the mixture can be cooled across the absorber c ooling coils, which will take a while to heat up again in the insulated absorber. Otherwise the liquid can be cooled to absorb as much vapor as possible and then isolated to the separator, which has a

PAGE 171

152 higher pressure rating. The absorber will be left with vapor that will not rise above 1.7 bar (10 psig) in the hot weather. Emergency Procedures Safety should be the primary concern in the case of an emergency. If a problem occurs, the system and hot water pumps should be shut down. This will minimize any pres sure buildup of and heat addition to the working fluid. There is an emergency power cutoff switch on the wall next to the experiment. Goggles should be worn at all times and a respirator is available when needed. If there is a leak, the overhead sprinklers should be turned on to dilute the air. If the leak is major, it is best to first leave the area and then to return with the respirator to turn the power off and to dilute the area. If the leak is minor, the power can be cut and the sprinklers can be turne d on first. The biggest concern is for the lungs and eyes. Chemical spill goggles will prevent direct injury to the eyes although fumes will get through without a respirator. The lungs are irritated by the ammonia and will burn if the fumes are concentrate d. Ammonia gas on the skin may irritate and will feel warm, as it is absorbed into moisture on the skin, releasing heat. Aqueous ammonia on the skin will simply feel slimy, but will not hurt unless an open wound is exposed to it in which case it burns. The best solution to skin and eye contact is dilution, and medical attention if the exposure is more serious.

PAGE 172

153 APPENDIX E UNCERTAINTY ANALYSIS The methods for determination of uncertainties of measured and calculated quantities are presented to establish confidence in the experimental results of the power and cooling cycle. Following some introductory theory, the analysis is applied to the measurements of temperature, pressure, ammonia mass fraction and flow rate, and to the calculation of other results based on these measurements. Theoretical Background of Analysis Method The background of the uncertainty analysis method is established in this section, discussing the error of primary measurements and then of the calculations based on these measurements. Uncertainty of Measured Values The uncertainty of a measured quantity is the error of the measurement that can be expected within a level of confidence. This error is the deviation of the measurement from the actual value. Knowing the uncertainty of the measurements will allow the determination of uncertainties of quantities calculated from these measurements later o n. For repeated measurements of the same quantity using the same instrument, the average of the measurements is given as Eq. E.1. = = N i i x N x 1 1 (E.1)

PAGE 173

154 The individual measurements will deviate from the mean, and the standard deviation is expressed as Eq. E.2. Note for a finite set of measurements, the denominator is N 1 and not simply N. 2 1 1 2 1 ) ( = = N x x N i i s (E.2) For a large set of measurements, the deviation from the mean will resemble a Gaussian normal error distribution. If the set has a sm all standard deviation, the data points are centered near the mean. For a larger standard deviation, the data set is more spread out. If an infinite number of measurements are taken, then the mean of these measurements will be the true mean of the measurem ents. For a finite set, the mean will vary from the true mean. The standard deviation of the mean is found according to Eq. E.3. Note that as N approaches infinity, the mean of the set approaches the true mean. That is, the standard deviation of the mean a pproaches zero. s s N m 1 = (E.3) When taking a set of measurements, it is possible that some readings have been taken in error despite care taken by the measurer. The basis to reject certain measurements should be exact and not done on whim. Chauvenets criterion for rejection of data points states that a reading may be eliminated from the set if the probability of attaining its deviation is less than 1/2N (Holman, 1966). In other words, the point can be rejected if the ratio of the deviation of that point to the standard deviation of the set is greater than an acceptable maximum, R, as given in Eq. E.4.

PAGE 174

155 reject x i if R x x i > s (E.4) The values of R are given in Table E.1, and depend on the number of points in the data set. For a larger set, a point must deviate more from the mean to be rejected. Note that Chauvenets criterion can be applied to the data set only once, and then a new standard deviation is determined based on the modified set. This standard deviation will be small er than for the original set, as erroneous points have been eliminated. Table E.1. Chauvenets criterion for elimination of data points (Holman, 1966). Number of readings Ratio of maximum deviation to std. deviation 2 1.15 3 1.38 4 1.54 5 1.65 6 1.73 7 1.80 10 1. 96 15 2.13 25 2.33 50 2.57 100 2.81 300 3.14 500 3.29 1000 3.48 The uncertainty of a measurement can be determined from the standard deviation of previous measurements. This uncertainty is a statistical approach that attempts to predict how much the measurement deviates from the true mean. No prediction can be 100% accurate, and so the uncertainty is made within a level of confidence. Table E.2 gives some values of uncertainty within confidence intervals.

PAGE 175

156 Table E.2. Uncertainties in single measurements within confidence inte rvals (Holman, 1966; Wilson, 1952). Confidence interval Uncertainty of the measurement 50% 0.674 s 68% s 80% 1.282 s 90% 1.645 s 95% 1.960 s 99% 2.576 s 99.9% 3.291 s From Table E.2 it is evident that one can be more confident that the measureme nt lies further from the mean than close to it. Every measurement should have an associated uncertainty with it. The level of confidence of the uncertainty is based on tolerances of the work and individual preference. Measurement uncertainties of the power /cooling cycle will be within a 90% confidence interval. For measurements that are repeated, the uncertainty is expected to get smaller, as the mean of these measurements will approach the true mean. The uncertainty of a measurement repeated K times, for a 90% confidence interval is then expressed as Eq. E.5. The standard deviation in Eq. E.5 is that for the population as found in Eq. 2, and not that for the smaller measurement set repeated K times. K 1 645 1 s w = within 90% confidence (E.5) Unc ertainty of Calculated Values The uncertainty of a calculated quantity is dependent on the uncertainties of all the primary measured quantities used in its determination. Express the quantity of interest as F, given in Eq. E.6, as a function of N variables ) ,..., ( 2 1 N F F F F F = (E.6)

PAGE 176

157 The uncertainty of F is then found from Eq. E.7, as a function of the uncertainties of the primary measurements (Holman, 1966; Wilson, 1952). Note that these uncertainties are not simply added, but are weighted according to the sensitivity of the function F to changes in each variable. 2 1 2 2 2 2 1 ... 2 1 + + + = N F N F F F F F F F F F w w w w (E.7) The partial derivatives in Eq. E.7 can be found with an analytical or numerical approach. Uncertainty of State Point Measurements The primary quantities to ident ify each state point are temperature, pressure, ammonia mass fraction and mass flow rate. Strictly speaking, these quantities are not measured, but are rather calculated from more elementary direct readings modified according to instrument specific calibra tion and correction terms. Determination of the uncertainties of temperature, pressure, ammonia mass fraction and mass flow rate at each state point must account for uncertainties of all the basic measurements of which they are composed. Temperature Measur ements A two point calibration curve for each thermocouple can be derived according to Eq. E.8. Within the range of operation, the high and low calibration temperatures were that of the boiling and freezing points of water. ( ) c L m L m m L m H c L c H T T T T T T T T , , , + = (E.8)

PAGE 177

158 T o determine the uncertainty of the temperature measurement, the uncertainties of all of the variables in Eq. E.8 must be accounted for according to Eq. E.7. However, it is assumed that both the high and low temperatures of the calibration fluid were well e stablished during the calibration procedure, and so their uncertainties are negligible. The uncertainty of temperature is then given as Eq. E.9. 2 1 2 2 2 , , i T m T m L T m H i T m m L m H T T T T T T + + = w w w w (E.9) The uncertainty of each measured temperature is the same for a given thermocoupl e, as the same thermocouple was used for all three measurements. Thus, i T i T i T m L m H m , , w w w = = (E.10) Differentiating Eq. E.8 and substituting, Eq. E.9 can be written as Eq. E.11. ( ) ( ) ( ) ( ) ( ) ( ) 2 1 , 2 , 2 2 , 2 , , 1 2 i m L m H m L m m L m H m L m m L m H c L c H i T i T T T T T T T T T T T T T m + = w w (E.11) The uncertainty was determined for each the rmocouple by taking 1000 samples at 25 Hz of the same temperature, which was assumed constant over the 40 second duration. The standard deviations were then found. If typically 250 samples are taken and averaged to get the mean, then according to Eq. E.5 t he uncertainties of the 14 thermocouples were found to vary from 0.077 to 0.094 C, within 90% certainty. Solving Eq. E.11, it is found that over the range of measured temperatures to be observed in this work, and for all the measured temperatures at high and low calibration, the uncertainty of calibrated temperature for each thermocouple is approximately the same.

PAGE 178

159 C i T T = = 1 0 w w within 90% confidence (E.12) The thermocouple uncertainties have not changed in the full period of system operation. The thermocouples have been periodically recalibrated, with the calibration results applied in DaqView such that the recorded readings are correct. The uncertainty depends on calibration values, but very slightly. As the calibration values used in the unc ertainty determination in the program are current values, applying them to old readings to determine their uncertainties is not exact but approximate. The uncertainties, however, will not change much and are all about 0.1 C. Note that the uncertainty code used in the analysis program is on the assumption that 250 samples are taken in DaqView The user may take a different number of readings, as long as the number of readings is inputted as a multiple of 250 for correct uncertainty analysis. Pressure Measu rements A two point calibration curve can be derived for each pressure transducer according to Eq. E.13. Although the transducers send a current response to the data acquisition system, the signals are converted to an equivalent pressure with no error assu med in the conversion. The low pressure for calibration is taken as atmospheric pressure, while the high pressure is a suitably high value and is measured with an accurate pressure gauge. ( ) c L m L m m L m H c L c H P P P P P P P P , , , + = (E.13) Assuming no uncertainty of the low c alibration pressure, the uncertainty of the calibrated pressure from each transducer is given as Eq. E.14.

PAGE 179

160 2 1 2 2 2 2 , , i P m P m L P m H P c H i P m m L m H c H P P P P P P P P + + + = w w w w w (E.14) The uncertainty of each pressure measured by the same transducer is the same, as in Eq. E.15. The high calibration pr essure is measured by a gauge with an uncertainty that may be different. i P i P i P m L m H m , , w w w = = (E.15) Differentiating Eq. E.13 and rewriting Eq. E.14 yields Eq. E.16. The uncertainty of each pressure thus depends on uncertainty of the transducers and th e uncertainty of the gauge used for calibration. ( ) ( ) = 2 , 2 2 , m L m H m L m P i P P P P P c H w w ( ) ( ) ( ) ( ) ( ) ( ) 2 1 , 2 , 2 2 , 2 , 2 1 2 i m L m H m L m m L m H m L m m L m H c L c H P P P P P P P P P P P P P m + + w (E.16) The gauge used in the high pressure calibration has an uncertainty of 2% full scale in the region of operation. The uncertainty of each transducer w as found similarly to the uncertainty of each thermocouple. If multiple samples are used in each measurement, then according to Eq. E.5 the uncertainty of each transducer reading is approximately 2 orders of magnitude less than that of the gauge reading. E q. E.16 reduces to Eq. E.17. i m L m H m L m P i P P P P P c H , , w w within 90% confidence (E.17) The uncertainty of the pressure will vary depending on the region of operation. To reduce the uncertainty of the pressure, the uncertainty of the calibration pressure gaug e

PAGE 180

161 can be reduced by taking multiple gauge readings. Likewise, the high pressure used in the calibration can be increased. However, as the transducers exhibit a rated accuracy that is already an order of magnitude better than the pressure gauge available fo r calibration, no additional calibration beyond that performed by the manufacturer is needed. However, if the calibration of the manufacturer is suspected to be in error, or the transducer readings have shifted, then a one point calibration can be used acc ording to Eq. E.18. The calibration pressure to use is atmospheric pressure, which does not vary much. ( ) m L c L m P P P P , + = (E.18) The measured pressures in Eq. E.18 have the same uncertainty, and the uncertainty in the pressure reduces to Eq. E.19. Th e uncertainty in the atmospheric pressure is taken as negligible. i P i P m , 2 w w = (E.19) The pressure readings have been performed with mainly transducers, but with gauges while certain transducers were at fault. The error in the devices is differ ent, with the error of the gauges being an order of magnitude greater than that of the transducers. Therefore, a time history of which devices were used with which experiments is necessary for correct determination of uncertainties. Table E.3 is in terms o f the chronological file names storing the data and results. The integer in the file name is the date in MMDDYY format, and the letter, if any, sequences the tests done on that day. The uncertainties of pressure in the program are calculated based on 250 s amples of the transducers, or one reading of the gauge. A different number of readings may be used but must be noted when data analysis begins.

PAGE 181

162 Table E.3. Device history of pressure measurement in the system. Dates P Device full scale range error %fs in itial 060901a boiler transducer 0 250 psig 0.25 vaporflo transducer 30 60 psig 0.13 vaporin transducer 30 60 psig 0.40 weakin transducer 30 60 psig 0.13 absexit transducer 30 60 psig 0.40 060901c 123101 boiler transducer 0 250 psig 0.25 vaporflo transducer 30 60 psig 0.13 vaporin transducer 30 60 psig 0.40 weakin gauge 30 100 psig 1.5 absexit transducer 30 60 psig 0.40 010102 100102 boiler transducer 0 250 psig 0.25 vaporflo transducer 30 60 psig 0.13 vaporin transducer 30 60 psig 0.40 weakin gauge 30 100 psig 1.5 absexit transducer 30 60 psig 0.13 100202 102502 boiler gauge 0 200 psig 1 vaporflo transducer 30 60 psig 0.13 vaporin transducer 30 60 psig 0.40 weakin gauge 30 100 psig 1.5 absexit transducer 30 60 psig 0.13 102602 present boiler gauge 0 200 psig 1 vaporflo transducer 30 60 psig 0.13 vaporin transducer 30 60 psig 0.40 weakin transducer 30 100 psig 0.13 absexit transducer 30 60 psig 0. 13 Ammonia Mass Fraction Measurements The ammonia mass fraction is determined according to Eq. E.20, as a function of the output from the gas chromatograph (GC) and its calibration. The output of the GC is a voltage response over time, for which peaks de velop corresponding to the sample constituents. The GC integrates the area under the voltage time plot and yields a percentage of the area under the ammonia peak of the total area. ( ) 3 3 3 % % 100 % NH NH NH A C A A x + = (E.20)

PAGE 182

163 The constant in Eq. E.20 is given as Eq. E .21, where the ammonia mass fraction of the standard sample was determined by the supplier. = c NH c NH c c A A x x C , 3 3 % 100 % 1 (E.21) The uncertainty of the calibration constant depends on the uncertainty of the standard sample and that of the GC reading, as given i n Eq. E.22. 2 1 2 2 % , 3 3 % + = c c NH x c A c NH C x C A C w w w (E.22) It is assumed that the uncertainty of the standard sample is very small and can be neglected in comparison to that of the reading. Eq. E.22 can be rewritten as Eq. E.23, dependent on the uncertainty of the GC rea ding. ( ) c NH A c NH c c C A x x 3 3 % 2 % 100 100 1 w w = (E.23) In order to determine the uncertainty of the GC reading, 50 measurements of the same liquid sample were taken. The uncertainty of the liquid and vapor readings is not necessarily the same and should be determined indep endently. For the liquid readings, the 50 measurements yielded a standard deviation of 0.416 in the area percentage, with one reading having been eliminated with Chauvenets criterion. For a single sample, the uncertainty of liquid GC readings according to Eq. E.5 is given as Eq. E.24a. It was sufficient to use fewer readings for determination of the vapor GC uncertainty. 15 readings of the same sample yielded a standard deviation of 0.187 in the area percentage, with one reading having been eliminated with Chauvenets criterion. For a single sample, the uncertainty of vapor GC readings according to Eq. E.5 is given as Eq. E.24b.

PAGE 183

164 684 0 % 3 = L A NH w for liquid, within 90% confidence (E.24a) 307 0 % 3 = g A NH w for vapor, within 90% confidence ( E.24b) The GC measurement of the area percentage of the standard sample of liquid was taken as an average of 10 readings, and according to Eq. E.5, the uncertainty of the standard sample is given as Eq. E.25. One reading was rejected. 228 0 3 % = c NH A w for standard, within 90% confidence (E.25) The uncertainty of the calibration constant can then be found from Eq. E.23 as Eq. E.26. The calibration constant was found to be 1.068 according to Eq. E.21, such that the error is approximately 1.1%. 012 0 = C w within 90% confidence (E.26) From Eq. E.20, the error of the mass fraction is then given as Eq. E.27. 2 1 2 % 2 3 3 % + = NH A NH C x A x C x w w w (E.27) Upon differentiation, the uncertainty of the mass fraction becomes Eq. E.28 for both liquid and vapo r readings. ( ) ( ) ( ) [ ] 2 1 4 2 % 2 2 2 2 3 3 3 3 3 % % 100 100 % 100 % + + = NH NH A C NH NH x A C A C A A NH w w w (E.28) Over a range of GC readings from 0 to 100 area percent, the liquid mass fraction uncertainty varies from 0.005 to 0.009. The corresponding error of a single typical liquid mass fraction measurement is approximate ly 2%. The vapor mass fraction uncertainty varies from 0.002 to 0.004. The corresponding error of a single typical vapor mass

PAGE 184

165 fraction measurement is approximately 0.5 %, less than that of the liquid mass fraction uncertainty. The uncertainty in the mass f raction, as determined from the GC, is based on the uncertainty of the GC reading and the calibration values. Raw, uncalibrated GC values are recorded, such that any recalibration will be applied to new and old results accordingly when results files are ca lled up. Any changes in the uncertainties will be applied as well. Note that the uncertainties in the program are for 1 sample. Multiple sampling with the GC will reduce the uncertainty. Flow Rate Measurements The flow rate of vapor is given by Eq. E.29, a ccounting for the actual fluid temperature and pressure not being that those in the calibration that was performed by the manufacturer. The measured flow rate is also given in scfm, and is adjusted to account for the actual temperature and pressure of the fluid. 2 1 2 1 = c c s s g m g P P T T P P T T Q Q (E.29) The uncertainty of the temperature and pressure in factory calibration is assumed to be negligible. The uncertainty of the vapor flow rate is then given as Eq. E.30. 2 1 2 2 2 vapor P T Q m g Q P Q T Q Q Q m + + = w w w w (E.30) Differentiation and substitution yield Eq. E.31 for vapor flow rate. 2 1 2 3 2 2 2 2 1 1 4 1 vapor P T m Q c c s s g Q P T TP Q P T P T T P m + + = w w w w (E.31)

PAGE 185

166 The uncertainty of the vapor flow meter reading is 1% full scale. The uncertainties in the temperature and pressure contribute little to the overall uncertainty, and Eq. E. 31 can reduce to Eq. E.32 with less than 5% error of the approximation. g Q c c s s g Q m P T P T T P 2 1 2 1 w w (E.32) The uncertainty depends on the temperature and pressure, but is typically 0.05 g/s or roughly 2 5% error, with less error for higher flow rates. The flo w rate of liquid is given by Eq. E.33, and is dependent on the flow meter reading and the variation in fluid density from the calibration fluid density. 2 1 = r r c L m L Q Q (E.33) Assuming that the calibration fluid density is well known, the uncertai nty of the liquid flow rate is given as Eq. E.34. The uncertainty is flow meter dependent. 2 1 2 2 , i liquid Q m i L Q Q Q Q m + = r w r w w (E.34) Differentiation of Eq. E.33 and substitution into Eq. E.34 yields Eq. E.35. 2 1 2 3 2 2 , 4 1 i liquid c m Q c i L Q Q m + = r w r r w r r w (E.35) The density of the liquid depends on the temperature, pressure and mass fraction of the liquid, such that Eq. E.35 can be written as Eq. E.36. 2 1 2 2 2 3 2 2 , 4 1 i liquid x P T c m Q c i L Q x P T Q m + + + = w r w r w r r r w r r w (E.36)

PAGE 186

167 For a liquid, the partial derivatives of density with respect to temperature, pressure and mass fraction o ver a small range away from the saturation curve do not vary much and can be assumed small. Eq. E.36 then reduces to Eq. E.37. 2 1 , i liquid c Q i L Q m r r w w (E.37) As the flow meters were calibrated with fluids that have densities near the actual fluid densitie s, the square root of the ratio of the two is nearly unity. Eq. E.37 reduces further to Eq. E.38. i L Q i L Q m , , w w (E.38) The uncertainties of the liquid flow meters are listed in Table E.4. Note that the uncertainties that are used in the analysis program are for 1 reading. Multiple readings will reduce the uncertainty. Table E.4. Liquid flow meter uncertainties. flow meter full scale (gpm) uncertainty (%fs) uncertainty (gpm) Strong 1.11 1.0 0.011 weak 0.84 1.0 0.0084 hot water 5.0 2.0 0.1 abs orber coolant 10.1 5.0 0.51 vapor coolant 1.0 2.0 0.02 Uncertainty of Energy Transfers and Efficiencies From the temperature, pressure, mass fraction and flow rates, properties are determined at all state points, which are used to calculate the energy t ransfers over boundaries of the cycle. This approach includes a binary properties routine, which cannot be easily written into a differentiable form as Eq. E.6 for the uncertainty analysis. The option is to use a numerical approximation for the derivatives in Eq. E.7. A central

PAGE 187

168 difference scheme is used as shown in Eq. E.39. The increment used should be no smaller than the uncertainty of that variable. ( ) ( ) * 2 * i j i F i i i i i F i F F F F F F F F F D D D + = (E.39) Table E.5. The uncertainties of calculated values and their dependence on measurements for a typical test of the experimental system, within 90% confidence. % error due to individual contribution Q h Q a Q c Q rec W p W t vf HE r HE h 1 st 2 nd T Absorber 2.56 .20 .12 Strong absorber exit .20 1.11 .24 .00 .09 Strong boiler in / rec ex .16 Vapor turbine inlet .00 1.16 .06 .00 .03 Weak recovery inlet .39 .14 Weak absorber inlet .17 .39 .40 Hot boiler inlet .46 .12 .41 .56 Hot boiler exit .39 .4 0 .41 .32 P Absorber .00 .00 .00 .00 .00 System high .00 7.21 .01 1.11 .75 .02 .40 1.02 1.03 Vapor absorber inlet .01 9.07 .90 1.43 1.35 Weak absorber inlet .00 .00 .00 Hot .00 .00 .00 .03 X St rong .61 .37 .00 1.96 .00 .06 Vapor .36 75.12 .39 2.85 2.94 Weak .93 .17 .02 m Strong 2.14 4.07 4.03 .41 .35 Vapor 4.72 8.37 8.32 8.36 8.96 9.02 Weak .38 4.05 .00 Hot 9.52 1.06 9.57 9.60 TOTAL % error 9.54 5.33 76.13 4.09 4.39 8.41 9.26 .47 2.31 13.65 13.63 Table E.5 shows the dependence of the error of the final values on each measurement. The total % error is the uncertainty on the final results, F, based on all contributions, F i The % err or contribution from each measurement, if acting alone, is given in the rows. This can be written as Eq. E.40. Note that these individual % error

PAGE 188

169 contributions will not add up to the total % error, as they do not add linearly, but rather according to Eq. E .7. 100 % F F F error i i F i F w = (E.40) In determining which of the uncertainties are significant it should be noted that the squares of the individual uncertainties are added. Thus, if one uncertainty is an order of magnitude greater that another, then the sm aller uncertainty is not very significant. As a complete study of the uncertainties over all tests has not been concluded, all uncertainties, except those that are more than 2 orders of magnitude smaller than the largest ones, are included for the current analysis. Boiler Heat Input The boiler heat input, as determined from the hot water side of the boiler, is dependent on the enthalpy change of the hot water across the boiler. Likewise, it is dependent on the specific enthalpy change and the mass flow rate Looking at primary measurements, the specific enthalpy depends on temperature and pressure of the hot water. Thus, the boiler heat input can be written as Eq. E.41. From continuity, the mass flow rates at the inlet and exit are equal. The measured pressu re is only at the inlet, with the exit pressure not expected to be significantly lower. ( ) hot in boiler ex boiler in boiler h h m P T T Q Q , , , = (E.41) As the enthalpy change in liquid is not sensitive to changes in pressure in the compressed liquid region, the boiler heat input will not vary much with variation in hot water pressure. Table E.5 confirms this small dependence on pressure. The uncertainty of

PAGE 189

170 the boiler heat input can then be written as Eq. E.42. The uncertainties of the temperatures are about the same. 2 1 2 2 2 2 2 hot m h T ex boiler h in boiler h Q m Q T Q T Q h + + w w w (E.42) The uncertainty of the flow rate is 6.2 g/s from Table E.5, and the uncertainty of the temperature is 0.1 C from Eq. E.12. The partial derivatives are then evaluated with the numerical approximation of Eq. E.39 for incremental changes in temper ature and mass flow rate while other parameters are held constant. Table E.5 shows the largest dependence on the mass flow rate, as the error of the flow rate measurement is high compared to the error of the temperature measurement. This error can be reduc ed in further testing according to Eq. E.5 by taking multiple readings of the flow rate. Absorber Heat Rejection The absorber heat rejection is calculated from the vapor, weak, and strong stream properties at the absorber boundaries. The dependence of the heat rejection is given as Eq. E.43. The vapor inlet temperature is not measured, but rather depends on whether or not there is refrigeration, which in turn depends on the turbine conditions. Thus, other parameters are cited which are measured. = weak vap str weak vap str in abs weak in abs vap high sys abs in abs weak in tur vap ex abs str a a m m m x x x P P P P T T T Q Q , , , , , , , , , , , (E.43) Table E.5 shows again the error of the mass flow rates to be significant. As the heat rejection is largely the latent heat of absorption, the mass fractions of the streams are significant as they dictate how much is to be absorbed. Var iation of weak inlet and strong

PAGE 190

171 exit pressures are expected to minimally affect the liquid properties, and can be ignored. Eq. E.44 shows the uncertainty of the heat rejection. 2 2 2 2 , 2 , 2 , high sys a P high sys a T in abs weak a in tur vap a ex abs str a Q P Q T Q T Q T Q w w w + + + 2 2 2 2 2 2 2 2 , , L x weak a g x vap a L x str a P in abs vap a x Q x Q x Q P Q in abs vap w w w w + + + + 2 1 2 2 2 2 2 2 + + + weak vap str m weak a m vap a m str a m Q m Q m Q w w w (E.4 4) Again, the error due to mass flow rate measurements can be reduced by taking multiple readings. Cooling Capacity The cooling capacity is determined from properties of the vapor at the inlet and exit of the refrigeration heat exchanger, as given in Eq. E .45. ( ) vap vap ex abs vap high sys in tur vap abs c c m x P P T T Q Q , , , , , = (E.45) From continuity, the mass flow rates and mass fractions are the same at the inlet and exit. However, the refrigeration heat exchanger is simulated. The inlet pressure is not known, but is estimated based on the measured refrigeration unit exit pressure (at the absorber inlet) and an assumed pressure loss across the unit. The exit temperature is assumed to approach the ambient within a limit, where the ambient is assumed to be a certain amount lower than the measured temp erature in the absorber. The inlet temperature is estimated based on the measured inlet conditions to the simulated turbine, an assumed turbine efficiency, and the refrigeration heat exchanger inlet pressure as found before.

PAGE 191

172 Table E.5 shows a large uncerta inty of the cooling capacity. Amongst the sources of error, the cooling capacity appears very sensitive to changes in the vapor mass fraction. This is especially true at high vapor mass fractions, which is why the original cycle design includes a rectifier to purify the vapor stream. The case shown in Table E.5 has over 99% ammonia in the vapor stream. The uncertainty of the cooling capacity is then written as Eq. E.46, where the different pressures are measured with transducers, which may have distinct unc ertainties. The uncertainties of the temperatures are the same, however. + + 2 2 2 2 2 high sys c P high sys c T in tur c abs c Q P Q T Q T Q w w w 2 1 2 2 2 2 2 2 , vapor m c g x c P in abs c m Q x Q P Q in abs + + + w w w (E.46) Internal Heat Recovery The internal heat recovery is determined from the weak solution side as it is expected to remain in one phase throughout the heat exchanger as a compressed liquid. With the same reasoning as before, the recovery is a function of temperature, pressure, mass fraction and mass flow rate according to Eq. E.47. ( ) weak weak in abs weak high sys in abs weak in rec weak rec rec m x P P T T Q Q , , , , , , = (E.47) From continuity, th e mass flow rate and the mass fraction remain the same across the heat exchanger. As the weak solution is compressed liquid, the specific enthalpy will not be sensitive to incremental changes in pressure. The uncertainty of the recovered heat is written as Eq. E.48, where the temperature uncertainty is the same at the inlet and exit. Note the additional term to account for changes in recovered heat with changes in mass

PAGE 192

173 fraction. This is included as the specific heat capacities and thus the enthalpies of amm onia and water are not necessarily equal. The exit temperature of the weak solution from the recovery unit is not measured, but is rather calculated from the measured weak solution inlet to the absorber. The temperature at the recovery exit is then found b y assuming an isenthalpic throttle. + + 2 2 2 2 2 , high sys rec P high sys rec T in abs rec in rec rec Q P Q T Q T Q w w w 2 1 2 2 2 2 2 2 , weak m rec L x rec P in abs rec m Q x Q P Q in abs + + + w w w (E.48) The largest error of the recovered heat is due to the error of the mass flow rate measurement, which can be reduced by taking multiple flow rate readings. Pump Wor k Input The pump work is determined from the inlet and exit states, according to Eq. E.49. The exit pressure is not measured at the pumps immediate exit but rather further downstream at the boiler exit. Pressure losses are assumed for the boiler and recov ery heat exchanger. The exit temperature is not measured either, but is found from the measured inlet temperature and an assumed pump isentropic efficiency. This temperature does not vary much across the pump. ( ) str str high sys abs ex abs str P P m x P P T W W , , , = (E.49) The uncertaint y of the pump work is then written as Eq. E.50, applying continuity and with the above assumptions. The pump work is relatively small, so a small error of temperature, pressure, and mass fraction measurements can be significant. As from Table E.5, the larg est error is again due to mass flow rate.

PAGE 193

174 + 2 2 2 2 abs P P abs P T ex abs P W P W T W w w w 2 1 2 2 2 2 2 2 , str m P L x P P high sys P m W x W P W high sys + + + w w w (E.50) Turbine Work Output The turbine work output depends on the inlet and exit states as shown in Eq. E.51. The exit pressure is measured further downstream at th e absorber inlet. As the turbine is simulated by an equivalent expansion, the exit temperature is calculated based on the pressure ratio and isentropic efficiency of the turbine. ( ) vap vap vap in abs vap high sys in tur vap T T m x m P P T W W , , , , , = (E.51) The uncertainty of the turbine work output th en becomes Eq. E.52, applying continuity and the above assumptions. + 2 2 2 2 , high sys T P high sys T T in tur T W P W T W w w w 2 1 2 2 2 2 2 2 , vap m T g x T P in abs T m W x W P W in abs + + + w w w (E.52) The turbine work output also depends highly on the pressure ratio, and is sensitive to the error of the pressure measurements. Howe ver, from Table E.5, the turbine work output is most sensitive to mass flow rate measurement. Vapor Fraction The uncertainty of the vapor fraction can be found without a numerical approximation as Eq. E.53. From Table E.5, the uncertainty of the vapor frac tion depends equally on both the mass flow rates of the strong solution and vapor streams.

PAGE 194

175 2 1 2 2 2 2 2 1 + = vap str m str m str vap vf m m m w w w (E.53) Boiler Heat Exchanger Effectiveness The boiler heat exchanger effectiveness is the ratio of the actual boiler heat input to the maxim um possible heat input. The maximum is determined from certain conditions on the strong solution side of the heat exchanger, in addition to the actual heat transfer as found from the hot water side. The parameters are given in Eq. E.54. ( ) hot str in boiler hot high sys ex boiler hot in boiler hot in boiler str b b m x P P T T T HE HE , , , , , , , = (E.54) The source pressure contributes little error, as shown in Table E.5. The uncertainty of the boiler heat exchanger effectiveness reduces to Eq. E.55. The largest error is due to the mass fraction and mass flow rate measurements. The strong solu tion mass fraction plays a role as it affects the pinch point in the maximum heat transfer calculations. + + 2 2 , 2 , 2 , T ex boiler hot b in boiler hot b in boiler str b HE T HE T HE T HE b w w 2 1 2 2 2 2 2 2 , + + + hot high sys m hot b L x str b P high sys b m HE x HE P HE w w w (E.55) The total error of the boiler heat exchanger effectiveness is relatively small. Recovery Heat Exchanger Effectiveness The recovery heat exchanger effectiveness is the ratio of the actual recovered heat to the maximum possible recovery. The maximum is determined from certain conditions on the strong solution side of the heat exchanger, in addition t o the actual recovery as found from the weak solution side. The parameters are given in Eq. E.56.

PAGE 195

176 = weak weak str in abs weak high sys abs in abs weak in rec weak ex abs str rec rec m x x P P P T T T HE HE , , , , , , , , (E.56) The uncertainty of the recovery heat exchanger effectives is very small over the trials observed. The pressures of the liquid streams do not contribute much to the overall error, and are neglected. The dependence, in general, becomes that of Eq. E.57. For the typical case of Table E.5, the boiling is not initiated in the recovery heat exchanger, and thus the strong solution mass fraction is not needed for determination of the pinch point and would drop out from Eq. E.57. The maximum recovery is rather assumed to occur for a weak solution temperature drop to near the strong solution inlet. The dependence of the overall error shifts towards the accurate determination of these temperatures. + + 2 2 , 2 , 2 , T in abs weak rec in rec weak rec ex abs str rec HE T HE T HE T HE rec w w 2 1 2 2 2 2 2 2 2 , + + + + weak high sys m weak rec L x weak rec str rec P high sys rec m HE x HE x HE P HE w w w (E.57) In general, the error of the recovery heat exchanger effectiveness is small when there is no boiling in the recovery heat exchanger. First La w Efficiency The first law efficiency accounts for the turbine and pump work, the boiler heat input and the cooling capacity, according to Eq. E.58. ( ) h c P T Q Q W W , 1 1 h h = (E.58) Looking at more primary measurements, the first law efficiency depends on s everal temperature, pressure, mass fraction and mass flow rate readings, as given in Eq. E.59.

PAGE 196

177 + + + + 2 2 1 2 1 2 , 1 2 , 1 2 1 1 T ex hot in hot in tur vap ex abs str abs T T T T T w h h h h h w h 2 2 1 2 2 , 1 2 2 1 , L x str P in abs vap P high sys x P P in abs vap high sys w h w h w h + + + 2 1 2 2 1 2 2 1 2 2 1 2 2 1 + + + + hot vap str m hot m vap m str g x vap m m m x w h w h w h w h (E.59) It is seen from Table E.5 that the first law efficiency is most dependent on th e mass flow rates owing to the large error of the measurements. However, it is also sensitive to the pressure ratio, which determines the extent of the expansion, and the vapor mass fraction, which determines the amount of cooling capacity. Second Law Effi ciency The uncertainty of the second law efficiency will depend on the pump and turbine work, the cooling capacity, and the exergy change of the solar heat source. ( ) hs c P T E Q W W D = , 2 2 h h (E.60) Using primary measurements, the uncertainty of the second la w efficiency will have the same measurement dependence as that of the first law efficiency, as shown in Eq. E.61. However, the uncertainty of the second law efficiency will not necessarily equal that of the first law efficiency. + + + + 2 2 2 2 2 2 , 2 2 , 2 2 2 2 T ex hot in hot in tur vap ex abs str abs T T T T T w h h h h h w h 2 2 2 2 2 , 2 2 2 2 , L x str P in abs vap P high sys x P P in abs vap high sys w h w h w h + + + 2 1 2 2 2 2 2 2 2 2 2 2 2 2 + + + + hot vap str m hot m vap m str g x vap m m m x w h w h w h w h (E.61)

PAGE 197

178 It is emphasized that as the turbine and recovery heat exchanger are simulated, the work output, cooling capacity and cycle efficiencies are speculations based on the vapor that was experimentally gene rated and certain assumptions. Therefore, the corresponding uncertainties are also speculations into what could be witnessed were a turbine and refrigeration unit in place. Conclusions of the Uncertainty Analysis A method of uncertainty analysis for experi mental data was presented and applied to a typical test performed on the power and cooling cycle. The errors in calculated energy transfers and efficiencies are mostly on the order of 5 10%, with larger error of the estimated refrigeration. The largest sou rce of error is in the mass flow rate determination. By taking multiple readings or with careful calibration, this error has been reduced in subsequent trials. The goal is to have no more than 5% error of the calculated results.

PAGE 198

179 APPENDIX F COMPUTER PROGRAM FOR DESIGN AND DATA ANA LYSIS A computer program was written to assist in the design, execution and analysis of experimental trials, making the cycle evaluation more efficient and informative. There are several functions in the program that were developed purely for convenience, but the critical functions that motivated the effort are presented below, along with key parts of the computer source code. Program Features The program was written initially to assist in the evaluation o f thermodynamic properties and the analysis of the energy transfers and efficiencies of the cycle. More features were subsequently developed to expedite processing experimental data, increase the tools for design and results analysis, and to improve the ov erall understanding of the cycle concept. Property Evaluation The property evaluation section of the program allows for evaluation of ammonia water mixture and pure component properties. The available inputs are enthalpy (h), pressure (P), entropy (s), tem perature (T), specific volume (v) and mass fraction (x). For the pure components, saturation properties are found for the user input of either pressure or temperature. Compressed liquid and superheated vapor properties can be found for inputs of P T or T v For mixtures, saturation properties can be found for the user inputs of P T, P x, or T x. Other properties can be found with three property inputs, including h P x, P s x, P T x, P T v, P v x, and T v x.

PAGE 199

180 Properties over a range of input conditions can be determined and sent to file to observe trends in the pure fluid or mixture properties. This function was used, for instance, to generate the bubble and dew point diagram in Appendix B. Design and Simulation Simulations can be produced with the program for the user specified 5 key fluid parameters and 12 component parameters that are discussed in Chapter 4. The simulated cycle can be used to show expected temperature, pressure, flow rate and mass fractions if designing an experiment. The results can be anal yzed and sent to file. An option is also available to simulate multiple cycles with variation in key fluid parameters and component parameters and send the results to file, for comparison and observation of trends. The plots in Chapter 4 were generated thi s way. Experimental Data Processing The files provided by the data acquisition include all of the raw measured pressure and temperature values in the chosen units in DaqView The units are set for C and psig, with the typical data files having 250 readin gs from 14 thermocouples and 5 transducers. Provided the raw files are in the same directory as the dataprocess.cpp source code, and labeled in the MMDDYYL format, the program can retrieve any specified raw data file and average the values. However, th e values from the raw data file are not enough to describe all the state points in the cycle. The user is thus prompted for more information in a step by step completion of the missing state point properties, including flow rate and GC values as given by t he instruments, the number of readings taken, and loss estimates.

PAGE 200

181 The cycle state points are now complete with measured properties and determined properties such as enthalpy, entropy and specific volume. Mass fractions and vapor qualities of mixture stream s are also found. An analysis of the results can proceed with the program. Displayed results include exergy losses, energy transfers, efficiencies, boiler and recovery heat exchanger effectiveness, and vapor fraction leaving the boiler. For experimental da ta, a full uncertainty analysis included for the results. The complete set of data, assumptions and loss estimates for the trial can then be saved into a results text file with the designation MMDDYYLr Calculated results for the experimental trial are al so saved. These results files can be later called up to retrieve the complete experimental data, and can be analyzed differently based on modified assumptions or preferences. Experimental Data Organization and Comparison Comments for the trials can be incl uded and saved to the file filelist.txt, which is a list of all the raw experimental files along with key operating parameters. It is a convenient list that can be called up and viewed from the main program. For quick comparison, backing up and printing purposes, all of the measured data from all of the tests can be sent to one file. This data includes the date and file name, averaged temperatures and pressures, flow rates, GC mass fraction readings, number of readings for each measurement, and comments. The results from the experimental data analysis of all files can all be compiled and sent to one file, along with the results of the corresponding simulated cycles with the same key operating parameters. The results that are sent can be sorted by a number of parameters, and limited by certain criteria selected by the user. This allows for quick comparison, grouping and analysis of multiple trials.

PAGE 201

182 Uncertainty Analysis The uncertainty analysis applies to experimental data, as the simulations have no uncertai nty. If an experimental data file is active, the uncertainty of all the measurements can be displayed as well as those in the calculated values. The individual contribution of each of the measurements is also displayed, in order to assess the sensitivity o f calculated results to primary measurements. This addresses where measurement error should be reduced for more accurate final results. Units Preferences The raw data sent to the text files by the data acquisition system are in units of C and psig. The wo rking units in the program are by user preference, and can be changed at any time. Options for temperature are C, F and K. Options for pressure are psig and bar. Flow rates are in the instrument readings (gpm and scfm) or in units of g/s. Mass fraction u nits are always in kg NH 3 / kg total. Calculated values are always in SI units of kJ/kg, kJ/kgK, m 3 /kg, and kW. Calculation Preferences and Calibration Constants The program allows for the user to enter preferences for the calculation of final results, and input of calibration constants, which are updated periodically for the measurement of temperatures, pressures, and mass fractions. The uncertainty of liquid and vapor sampling can be modified, as well as the rated uncertainty of components if they are rep laced. These values are always saved to the tempvalues.txt file upon exiting and are retrieved when the program boots up. Program Structure There are basically two major elements of the computer code, written in C ++ The functions for the calculation of mixture properties are in the dataprocess.h header file,

PAGE 202

183 with approximately 1300 lines of code. The functions for thermodynamic calculations of the cycle are in the dataprocess.cpp main file, with over 9000 lines of code. This file also contains the ba sic structure of the program, along with all other practical functions of the program. The abbreviated C ++ source code for both files is given below for the functions, omitting several secondary functions not critical to the program. The two dataprocess files also call standard C ++ header files. In starting the dataprocess.exe program, it searches for a file called tempvalues.txt in the same directory, which is of a specific format. This file stores the last operating values of the program when the pr ogram is exited, and retrieves them upon startup. These include calibration constants, analysis preferences, state points of the last shown cycle whether simulated or experimental, loss specifications, and instrument uncertainties. If the tempvalues.txt file is in error, then the program will have inaccurate information for subsequent data analysis and simulation. A typical tempvalues.txt file is shown at the end of this appendix. Functional commands of the program can be listed by typing help at the user prompt on each page while running the program. Some notes from the programmer can also be read by accessing the help sections. Property Evaluation Functions The functions for the determination of thermodynamic properties of the ammonia water mixtures are given below. The evaluation methods and constants are noted in Appendix B. The units for the property evaluation functions for both input and return values are given with each function.

PAGE 203

184 Critical Temperature and Pressure of the Mixture // output: criti cal temperature (K) of the mixture // input: mass fraction double crittemp(double xx) { Tc=0; i=1; sum1=0; while (i<=4) { sum1=sum1+(a[i]*pow(xx,i)); i++; } Tc=(Tcw sum1)/1.8; return Tc; } // output: critical pressure (bar) of the mixture // input: mass fraction double critpress(double xx) { Pc=0; i=1; sum1=0; while (i<=8) { sum1=sum1+(b[i]*pow(xx,i)); i++; } Pc=Pcw*exp(sum1)/14.51135; return Pc; } Bubble and Dew Point Temperature, Pressure and Mass Fraction // output: bubble point temperature (K) of the mixture // input: pressure (bar), mass fraction double bubbleT(double PP,double xx) { Tc=crittemp(xx); Pc=critpress(xx); i=1; sum2=0; while (i<=7) { j=1; sum1=0; while (j<=10) { sum1=sum1+Cij[i][j]*pow(xx,j); j++; } sum2=sum2+(Ci[i ]+sum1)*pow(log(Pc/PP),i); i++; } Tb=Tc sum2/1.8; return Tb; } // output: dew point temperature (K) of the mixture // input: pressure (bar), mass fraction double dewT(double PP,double xx) { Tc=crittemp(xx); Pc=critpress(xx); i=1; sum2=0; while (i<=6) { j=1; sum1=0; while (j<=4) { sum1=sum1+Aij[i][j]*pow(log(1.0001 xx),j); j++; } sum2=sum2+(ai[i]+sum1)*pow(log(Pc/PP),i); i++; } Td=Tc sum2/1.8; return Td; } // output: bubble point pressure (bar) of the mixture

PAGE 204

185 // input: temperature (K), mas s fraction double bubbleP(double TT,double xx) { Pb=critpress(xx); Tb=1000; incr=10; while(fabs(TT Tb)>.0001) { while(TT Tb<0) { Pb=Pb incr; while(Pb<=incr) { Pb=Pb+incr; incr=incr/10; } Tb=bubbleT(Pb,xx); } Pb=Pb+incr; Tb=bubbleT(P b,xx); incr=incr/10; } return Pb; } // output: dew point pressure (bar) of the mixture // input: temperature (K), mass fraction double dewP(double TT,double xx) { Pd=critpress(xx); Td=1000; incr=10; while(fabs(TT Td)>.0001) { while(TT Td<0) { Pd= Pd incr; while(Pd<=incr) { Pd=Pd+incr; incr=incr/10; } Td=dewT(Pd,xx); } Pd=Pd+incr; Td=dewT(Pd,xx); incr=incr/10; } return Pd; } // output: bubble point ammonia mass fraction of the mixture // input: temperature (K), pressure (bar) dou ble bubblex(double TT,double PP) { xb=0; Tc=crittemp(xb); Pc=critpress(xb); Tb=10000; incr=.1; while(fabs(TT Tb)>.0001) { while(TT Tb<0) { xb=xb+incr; if(xb>1) { xb=2; goto done; } Tc=crittemp(xb); Pc=critpress(xb); Tb=bubbleT(PP,xb); } xb=xb incr; if(xb<0) { xb= 1; goto done; } Tc=crittemp(xb); Pc=critpress(xb); Tb=bubbleT(PP,xb); incr=incr/10; } done: return xb; } // output: dew point ammonia mass fraction of the mixture

PAGE 205

186 // input: temperature (K), pressure (bar) double dewx(double TT,double PP) { xd=0; Tc=crittemp(xd); Pc=critpress(xd); Td=10000; incr=.1; while(fabs(TT Td)>.0001) { while(TT Td<0) { xd=xd+incr; if(xd>1) { xd=2; goto done; } Tc=crittemp(xd); Pc=critpress(xd); Td=dewT(PP,xd); } xd=xd incr; if(xd<0) { xd= 1; goto done; } Tc=crittemp(xd); Pc=critpress(xd); Td=dewT(PP,xd); incr=incr/10; } done: return xd; } Enthalpy of Pure Components and the Mixture // output: enthalpy (kJ/kmol water) of pure water in liquid phase // input: reduced temperature, reduced pressure double enthalpylw(double Tr,double Pr) { hLw= R*TB*( hLrow+B1w*(Trow Tr)+B2w/2*(Trow*Trow Tr*Tr)+B3w/3*(pow(Trow,3) pow(Tr,3)) (A1w+A4w*Tr*Tr)*(Pr Prow) A2w/2*(Pr*Pr Prow*Prow)); return hLw; } // output: enthalpy (k J/kmol ammonia) of pure ammonia in liquid phase // input: reduced temperature, reduced pressure double enthalpyla(double Tr,double Pr) { hLa= R*TB*( hLroa+B1a*(Troa Tr)+B2a/2*(Troa*Troa Tr*Tr)+B3a/3*(pow(Troa,3) pow(Tr,3)) (A1a+A4a*Tr*Tr)*(Pr Proa) A2a/2 *(Pr*Pr Proa*Proa)); return hLa; } // output: enthalpy (kJ/kmol water) of pure water in gaseous phase // input: reduced temperature, reduced pressure double enthalpygw(double Tr,double Pr) { hgw= R*TB*( hgrow+D1w*(Trow Tr)+D2w/2*(Trow*Trow Tr*Tr)+D3w/3*( pow(Trow,3) pow(Tr,3)) C1w*(Pr Prow) 4*C2w*(Pr*pow(Tr, 3) Prow*pow(Trow, 3)) 12*C3w*(Pr*pow(Tr, 11) Prow*pow(Trow, 11)) 4*C4w*(pow(Pr,3)*pow(Tr, 11) pow(Prow,3)*pow(Trow, 11))); return hgw; } // output: enthalpy (kJ/kmol ammonia) of pure ammonia in gas eous phase // input: reduced temperature, reduced pressure double enthalpyga(double Tr,double Pr) { hga= R*TB*( hgroa+D1a*(Troa Tr)+D2a/2*(Troa*Troa Tr*Tr)+D3a/3*(pow(Troa,3) pow(Tr,3)) C1a*(Pr Proa) 4*C2a*(Pr*pow(Tr, 3) Proa*pow(Troa, 3)) 12*C3a*(Pr*po w(Tr, 11) Proa*pow(Troa, 11)) 4*C4a*(pow(Pr,3)*pow(Tr, 11) pow(Proa,3)*pow(Troa, 11))); return hga; } // output: excess enthalpy (kJ/kmol mixture) of the mixture

PAGE 206

187 // input: reduced temperature, reduced pressure, mole fraction double enthalpyE(double Tr,do uble Pr,double y) { hE= R*TB*y*(1 y)*( E1 E2*Pr 2*E5/Tr 3*E6*pow(Tr, 2)+(2*y 1)*( E7 E8*Pr 2*E11/Tr 3*E12*pow(Tr, 2))+pow((2*y 1),2)*( E13 E14*Pr 2*E15/Tr 3*E16*pow(Tr, 2))); return hE; } // output: enthalpy (kJ/kmol mixture) of the mixture in liquid p hase // input: enthalpy (kJ/kmol ammonia), enthalpy (kJ/kmol water), excess enthalpy (kJ/kmol mixture), // mole fraction double enthalpyL(double hLa,double hLw,double hE,double y) { hLm=y*hLa+(1 y)*hLw+hE; return hLm; } // output: enthalpy (kJ/kmol mixtu re) of mixture in gaseous phase // input: enthalpy (kJ/kmol ammoia), enthalpy (kJ/kmol water), mole fraction double enthalpyg(double hga,double hgw,double y) { hgm=y*hga+(1 y)*hgw; return hgm; } Entropy of Pure Components and the Mixture // output: entro py (kJ/kmol K water) of pure water in liquid phase // input: reduced temperature, reduced pressure double entropylw(double Tr,double Pr) { sLw= R*( sLrow B1w*log(Tr/Trow)+B2w*(Trow Tr)+B3w/2*(Trow*Trow Tr*Tr)+(Pr Prow)*(A3w+2*A4w*Tr)); return sLw; } // output: entropy (kJ/kmol K ammonia) of pure ammonia in liquid phase // input: reduced temperature, reduced pressure double entropyla(double Tr,double Pr) { sLa= R*( sLroa B1a*log(Tr/Troa)+B2a*(Troa Tr)+B3a/2*(Troa*Troa Tr*Tr)+(Pr Proa)*(A3a+2*A4a*Tr)); r eturn sLa; } // output: entropy (kJ/kmol K water) of pure water in gaseous phase // input: reduced temperature, reduced pressure double entropygw(double Tr,double Pr) { if(Pr==0) sgw=0; else sgw= R*( sgrow D1w*log(Tr/Trow)+D2w*(Trow Tr)+D3w/2*(Trow* Trow Tr*Tr)+ log(Pr/Prow) 3*C2w*(Pr*pow(Tr, 4) Prow*pow(Trow, 4)) 11*C3w* (Pr*pow(Tr, 12) Prow*pow(Trow, 12)) 11/3*C4w*(pow(Pr,3)* pow(Tr, 12) pow(Prow,3)*pow(Trow, 12))); return sgw; } // output: entropy (kJ/kmol K ammonia) of pure ammonia in ga seous phase // input: reduced temperature, reduced pressure double entropyga(double Tr,double Pr) { if(Pr==0) sga=0; else

PAGE 207

188 sga= R*( sgroa D1a*log(Tr/Troa)+D2a*(Troa Tr)+D3a/2*(Troa*Troa Tr*Tr)+ log(Pr/Proa) 3*C2a*(Pr*pow(Tr, 4) Proa*pow(Troa, 4)) 1 1*C3a* (Pr*pow(Tr, 12) Proa*pow(Troa, 12)) 11/3*C4a*(pow(Pr,3)* pow(Tr, 12) pow(Proa,3)*pow(Troa, 12))); return sga; } // output: excess entropy (kJ/kmol K mixture) of the mixture // input: reduced temperature, reduced pressure, mole fraction doubl e entropyE(double Tr,double Pr,double y) { sE= R*(1 y)*y*(E3+E4*Pr E5*pow(Tr, 2) 2*E6*pow(Tr, 3)+(2*y 1)*(E9+E10*Pr E11*pow(Tr, 2) 2*E12*pow(Tr, 3))+pow((2*y 1),2)*( E15*pow(Tr, 2) 2*E16*pow(Tr, 3))); return sE; } // output: entropy (kJ/kmol K mixture) of mixing // input: mole fraction double entropymixing(double y) { if(y==0||y==1) smix=0; else smix= R*(y*log(y)+(1 y)*log(1 y)); return smix; } // oputput: entropy (kJ/kmol K mixture) of the mixture in liquid phase // input: entropy (kJ/kmol K am monia), entropy (kJ/kmol K water), excess entropy (kJ/kmol K mixture), // entropy of mixing (kJ/kmol K mixture), mole fraction double entropyL(double sLa,double sLw,double sE,double smix,double y) { sLm=y*sLa+(1 y)*sLw+sE+smix; return sLm; } // output: e ntropy (kJ/kmol K mixture) of mixture in gaseous phase // input: entropy (kJ/kmol K ammonia), entropy (kJ/kmol K water), entropy of mixing (kJ/kmol K mixture), // mole fraction double entropyg(double sga,double sgw,double smix,double y) { sgm=y*sga+(1 y)* sgw+smix; return sgm; } Specific Volume of Pure Components and the Mixture // output: specific volume (m3/kmol water) of water in liquid phase // input: reduced temperature, reduced pressure double specvollw(double Tr,double Pr) { vLw=R*TB/PB*(A1w+A3w*Tr +A4w*Tr*Tr+A2w*Pr); vLw=vLw/100; return vLw; } // output: specific volume (m3/kmol ammonia) of ammonia in liquid phase // input: reduced temperature, reduced pressure double specvolla(double Tr,double Pr) { vLa=R*TB/PB*(A1a+A3a*Tr+A4a*Tr*Tr+A2a*Pr); vLa= vLa/100; return vLa; } // output: specific volume (m3/kmol water) of pure water in gaseous phase // input: reduced temperature, reduced pressure

PAGE 208

189 double specvolgw(double Tr,double Pr) { if(Pr==0) vgw=0; else vgw=R*TB/PB*(Tr/Pr+C1w+C2w*pow(Tr, 3)+C3w* pow(Tr, 11)+ C4w*pow(Pr,2)*pow(Tr, 11)); vgw=vgw/100; return vgw; } // output: specific volume (m3/kmol ammonia) of pure ammonia in gaseous phase // input: reduced temperature, reduced pressure double specvolga(double Tr,double Pr) { if(Pr==0) vga=0; else vga=R*TB/PB*(Tr/Pr+C1a+C2a*pow(Tr, 3)+C3a*pow(Tr, 11)+ C4a*pow(Pr,2)*pow(Tr, 11)); vga=vga/100; return vga; } // output: excess specific volume (m3/kmol mixture) of the mixture // input: reduced temperature, reduced pressure, mole fraction dou ble specvolE(double Tr,double Pr,double y) { vE=R*TB/PB*y*(1 y)*(E2+E4*Tr+(2*y 1)*(E8+E10*Tr)+pow((2*y 1),2)*E14); vE=vE/100; return vE; } // output: specfic volume (m3/kmol mixture) of the mixture in liquid phase // input: specific volume (m3/kmol ammo nia), specific volume (m3/kmol water), excess specific // volume (m3/kmol mixture), mole fraction double specvolL(double vLa,double vLw,double vE,double y) { vLm=y*vLa+(1 y)*vLw+vE; return vLm; } // output: specific volume (m3/kmol mixture) of mixture in gaseous phase // input: specific volume (m3/kmol ammonia), specific volume (m3/kmol water), mole fraction double specvolg(double vga,double vgw,double y) { vgm=y*vga+(1 y)*vgw; return vgm; } Determination of All Properties at a State Point The following functions are used to determine all properties (enthalpy in kJ/kg, entropy in kJ/kgK, specific volume in m 3 /kg, pressure in bar, temperature in K, and mass fraction) at a state point, given 3 of these properties (2 at saturation). The input units are the same. // output: non saturation properties // input: enthalpy (kJ/kg), pressure (bar), mass fraction void hPx (double temph,double PP,double xx) { TT=400; incr2=10; PTx(PP,TT,xx);

PAGE 209

190 while (fabs(temph hm)>.01) { while(hm>temph) { TT=TT incr2; PTx(PP,TT ,xx); } incr2=incr2/10; while(hm.001) { while(sm>temps) { TT=TT incr2; PTx(PP,TT,xx); } incr2=incr2/10; while(smlimit) { n=1; while(vm>tempv) { if(n>=11) { incr2=incr2*10; n=1; } xx=xx incr2; if(xx<0) { printf("mixture not possible \ n"); getch(); goto done; } PTx(PP,TT,xx); n=n+1; } incr2=incr2/10; n=1; while(vm=11) { incr2=incr2*10; n=1; } xx=xx+incr2; if(xx>1) { printf("mixt ure not possible \ n"); getch(); goto done; } PTx(PP,TT,xx); n=n+1; } incr2=incr2/10;

PAGE 210

191 } done: ; } // output: non saturation properties // input: pressure (bar), temperature (K), mass fraction void PTx (double PP,double TT,double xx) { Tr=TT/TB; Pr=PP/PB; M=18.015*17.031/((1 xx)*17.031+xx*18.015); y=xx*18.015/(xx*18.015+(1 xx)*17.031); Tb=bubbleT(PP,xx); Td=dewT(PP,xx); // for compressed liquid if(TTTd) { istate=3; h m=enthalpyg(enthalpyga(Tr,Pr),enthalpygw(Tr,Pr),y)/M; sm=entropyg(entropyga(Tr,Pr),entropygw(Tr,Pr),entropymixing(y),y)/M; vm=specvolg(specvolga(Tr,Pr),specvolgw(Tr,Pr),y)/M; xxNH3l=xxH2Ol=0; xxNH3v=xx; xxH2Ov=1 xx; } // for liquid vapor mixture else { istate=2; xb=bubblex(TT,PP); xd=dewx(TT,PP); qm=(xx xb)/(xd xb); /* quality of mixture */ yb=xb*18.015/(xb*18.015+(1 xb)*17.031); yd=xd*18.015/(xd*18.015+(1 xd)*17.031); Mb=18.015*17.031/((1 xb)*17.031+xb*18.015); Md=18.015*17.031/((1 xd)*17.031+xd* 18.015); hm=(1 qm)/Mb*enthalpyL(enthalpyla(Tr,Pr),enthalpylw(Tr,Pr),enthalpyE(Tr,Pr,yb),yb)+ qm/Md*enthalpyg(enthalpyga(Tr,Pr),enthalpygw(Tr,Pr),yd); //kJ/kg(mix at current state) sm=(1 qm)/Mb*entropyL(entropyla(Tr,Pr),entropylw(Tr,Pr),entropyE(Tr,Pr,yb ), entropymixing(yb),yb)+qm/Md*entropyg(entropyga(Tr,Pr),entropygw(Tr,Pr), entropymixing(yd),yd); vm=(1 qm)/Mb*specvolL(specvolla(Tr,Pr),specvollw(Tr,Pr),specvolE(Tr,Pr,yb),yb)+ qm/Md*specvolg(specvolga(Tr,Pr),specvolgw(Tr,Pr),yd); xxNH3v=(xx xb)/(xd xb) *xd; xxH2Ov=(xx xb)/(xd xb)*(1 xd); //mass fractions xxNH3l=(1 (xx xb)/(xd xb))*xb; xxH2Ol=(1 (xx xb)/(xd xb))*(1 xb); //mass fractions } } // output: non saturation properties // input: pressure (bar), specific volume (m3/kg), mass fraction void Pvx (do uble PP,double tempv,double xx) { TT=300; incr2=1; limit=.000001; PTx(PP,TT,xx); while (fabs(tempv vm)>limit) { n=1; while(vm>tempv) { if(n>=11) { incr2=incr2*10; n=1; } TT=TT incr2; PTx(PP,TT,xx); n=n+1;

PAGE 211

192 } incr2=incr2/10; n=1; whi le(vm=11) { incr2=incr2*10; n=1; } TT=TT+incr2; PTx(PP,TT,xx); n=n+1; } incr2=incr2/10; } } // output: non saturation properties // input: temperature (K), specific volume (m3/kg), mass fraction void Tvx (double TT,double tempv,double xx) { PP=1; incr2=.01; limit=.000001; PTx(PP,TT,xx); while (fabs(tempv vm)>limit) { while(vmtempv) { if(n>=11) { incr2=incr2*10; n=1; } PP=PP+incr2 ; PTx(PP,TT,xx); n=n+1; } incr2=incr2/10; } } // output: saturation properties // input: pressure (bar), mass fraction void Px (double PP,double xx) { Pr=PP/PB; y=xx*18.015/(xx*18.015+(1 xx)*17.031); Tb=bubbleT(PP,xx); Td=dewT(PP,xx); M=18.015*17.0 31/((1 xx)*17.031+xx*18.015); TT=Tb; hm=hb=enthalpyL(enthalpyla(Tb/TB,Pr),enthalpylw(Tb/TB,Pr),enthalpyE(Tb/TB,Pr,y),y)/M; hd=enthalpyg(enthalpyga(Td/TB,Pr),enthalpygw(Td/TB,Pr),y)/M; sm=sb=entropyL(entropyla(Tb/TB,Pr),entropylw(Tb/TB,Pr), entropyE(Tb/TB ,Pr,y),entropymixing(y),y)/M; sd=entropyg(entropyga(Td/PB,Pr),entropygw(Td/TB,Pr),entropymixing(y),y)/M; vm=vb=specvolL(specvolla(Tb/TB,Pr),specvollw(Tb/TB,Pr),specvolE(Tb/TB,Pr,y),y)/M; vd=specvolg(specvolga(Td/TB,Pr),specvolgw(Td/TB,Pr),y)/M; } // sa turation properties // input: pressure (bar), tempeature (K) void PT (double PP,double TT) { Tr=TT/TB; Pr=PP/PB; xb=bubblex(TT,PP); if(xb<0) { hm=hb=hd=sm=sb=sd=vm=vb=vd=xb=0; goto done; } xx=xb; y=xx*18.015/(xx*18.015+(1 xx)*17.031); M=18.015*17.031 /((1 xx)*17.031+xx*18.015); xd=dewx(TT,PP); if(xd>1) { hm=hb=hd=sm=sb=sd=vm=vb=vd=0; xd=1; goto done;

PAGE 212

193 } Md=18.015*17.031/((1 xd)*17.031+xd*18.015); yd=xd*18.015/(xd*18.015+(1 xd)*17.031); hm=hb=enthalpyL(enthalpyla(Tr,Pr),enthalpylw(Tr,Pr),enthalpyE( Tr,Pr,y),y)/M; hd=enthalpyg(enthalpyga(Tr,Pr),enthalpygw(Tr,Pr),yd)/Md; sm=sb=entropyL(entropyla(Tr,Pr),entropylw(Tr,Pr),entropyE(Tr,Pr,y),entropymixing(y),y)/M; sd=entropyg(entropyga(Tr,Pr),entropygw(Tr,Pr),entropymixing(yd),yd)/Md; vm=vb=specvolL(spe cvolla(Tr,Pr),specvollw(Tr,Pr),specvolE(Tr,Pr,y),y)/M; vd=specvolg(specvolga(Tr,Pr),specvolgw(Tr,Pr),yd)/Md; done: ; } // output: saturation properties // input: temperature (K), mass fraction void Tx (double TT,double xx) { Tr=TT/TB; y=xx*18.015/(xx*1 8.015+(1 xx)*17.031); Pb=bubbleP(TT,xx); Pd=dewP(TT,xx); M=18.015*17.031/((1 xx)*17.031+xx*18.015); PP=Pb; hm=hb=enthalpyL(enthalpyla(Tr,Pb/PB),enthalpylw(Tr,Pb/PB),enthalpyE(Tr,Pb/PB,y),y)/M; hd=enthalpyg(enthalpyga(Tr,Pd/PB),enthalpygw(Tr,Pd/PB),y)/M; sm=sb=entropyL(entropyla(Tr,Pb/PB),entropylw(Tr,Pb/PB), entropyE(Tr,Pb/PB,y),entropymixing(y),y)/M; sd=entropyg(entropyga(Tr,Pd/PB),entropygw(Tr,Pd/PB),entropymixing(y),y)/M; vm=vb=specvolL(specvolla(Tr,Pb/PB),specvollw(Tr,Pb/PB),specvolE(Tr,Pb/PB,y),y) /M; vd=specvolg(specvolga(Tr,Pd/PB),specvolgw(Tr,Pd/PB),y)/M; } Thermodynamic Analysis Functions The thermodynamic analysis functions are used to determine mixture properties at the state points, and calculate cycle energy transfers and efficiencies. The se routines are based on the discussions in Chapter 2 and Appendix E, with simple mass and energy balances applied. These functions rely on the property functions of the previous section. Determination of All State Points in the Cycle This function determi nes the state points in the cycle based on primary operating parameters in the cycle. The cycle must first be checked to verify that boiling can occur. // input: heat source temperature (K), boiler pressure (bar), ambient temperature (K), absorber pressure // (bar), basic solution mass fraction, heat source flow ratio, turbine isentropic efficiency, pump // isentropic efficiency, recovery HE effectiveness, boiler HE effectiveness // (note that other losses are also applied) void statepoints(double bufhst,do uble bufbp,double bufat,double bufap,double bufbsmf,double bufhsfr,double buftie,double bufpie,double bufrecHE,double bufboilerHE) { // assign the dead state to equal the ambient temperature calcT[0]=bufat; // assign the boiler pressure calcP[4]=calcP[5 ]=calcP[8]=bufbp; // determine the absorber temperature from the ambient and the approach limit

PAGE 213

194 calcT[1]=bufat+lossT[1]; // assign the basic solution mass fraction calcx[1]=calcx[2]=calcx[3]=calcx[4]=bufbsmf; // assign the basic solution mass flow rate calcm[2]=calcm[3]=calcm[4]=calcm[1]=m[1]; // assign the absorber pressure calcP[1]=bufap; // assign the coolant properties for(i=11;i<=14;i++) { calcT[i]=T[i]; calcP[i]=P[i]; calcm[i]=m[i]; calcx[i]=x[i]; } // assign the heat source temperature calcT [15]=bufhst; // tack on pressure losses calcP[7]=calcP[1]/(1 lossP[1]/100); calcP[10]=calcP[1]/(1 lossP[2]/100); calcP[3]=calcP[4]/(1 lossP[5]/100); calcP[2]=calcP[3]/(1 lossP[3]/100); calcP[9]=calcP[8]*(1 lossP[4]/100); calcP[6]=calcP[7]/(1 lossP[6]/10 0); //assign the heat source flow rate calcm[15]=calcm[16]=bufhsfr*calcm[1]; // finding the pump exit temperature, given the exit pressure, mass fraction, and pump efficiency PTx(calcP[1],calcT[1],calcx[1]); h[1]=hm;s[1]=sm; Psx(calcP[2],s[1],calcx[2]); h[2]=(hm h[1])/bufpie+h[1]; hPx(h[2],calcP[2],calcx[2]); calcT[2]=TT; // DETERMINATION OF THE TEMPERATURES ABOUT THE HEAT EXCHANGERS // this is an iterative process, which is done for 3 cases of no vapor, all vapor, or partial boiling // normally, we are interested in the partial boiling case // initial guess as the heat source temperature calcT[4]=calcT[8]=calcT[5]=bufhst; count2=0; iterate: guess=calcT[4]; calcT[8]=calcT[5]=calcT[4]; // The boiler temperature is not independent, but is found through i teration and by assuming the boiler and // recovery units each have an effectiveness and temperature approach limits // Finding the mass fractions and flow rates based on separator conditions and basic x and m PTx(calcP[4],calcT[4],calcx[4]); h[4]=hm; x NH3l[4]=xxNH3l;xNH3v[4]=xxNH3v; xH2Ol[4]=xxH2Ol;xH2Ov[4]=xxH2Ov; calcx[5]=calcx[6]=calcx[7]=xNH3v[4]/(xNH3v[4]+xH2Ov[4]); calcx[8]=calcx[9]=calcx[10]=xNH3l[4]/(xNH3l[4]+xH2Ol[4]); calcm[5]=calcm[6]=calcm[7]=(xNH3v[4]+xH2Ov[4])*calcm[4]; calcm[8]=calcm[ 9]=calcm[10]=(xNH3l[4]+xH2Ol[4])*calcm[4]; // FIRST look at case of no vaporization if(calcm[5]<.0001) { cycletype=3; // for no vaporization calcm[8]=calcm[9]=calcm[10]=calcm[1]; // weak flow = strong flow calcx[8]=calcx[9]=calcx[10]=calcx[1]; / / weak mf= strong mf calcm[5]=calcm[6]=calcm[7]=0; // no vapor flow calcx[5]=calcx[6]=calcx[7]=1; //assign highest mass fraction to vapor // DETERMINING strong recovery exit T[3], weak recovery exit T[9] // 1.assume heat capacity C weak < C strong, s uch that T[9] drops to pump exit T[2] for max recovery // (100% effectiveness: I define as max recovery within approach limits, // with the approach limits usually taken as 0 for maximum heat theoretical recovery) calcT[9]=calcT[2]; PTx(calcP[9],calcT[9], calcx[9]); h[9]=hm; // 2.determine T[3] from control volume energy balance PTx(calcP[8],calcT[8],calcx[8]); h[8]=hm; h[3]=calcm[8]/calcm[3]*(h[8] h[9])+h[2]; hPx(h[3],calcP[3],calcx[3]); calcT[3]=TT; // 3.determine max recovery with the C weak < C strong assumption temp4=m[2]*(h[3] h[2]);

PAGE 214

195 // 4.assume C weak > C strong, such that T[3] rises to near T[8] for max rec calcT[3]=calcT[8]; PTx(calcP[3],calcT[3],calcx[3]); h[3]=hm; // 5.determine T[9] from control volume energy balance PTx(calcP[3],calcT[3],ca lcx[3]); h[3]=hm; h[9]=h[8] calcm[3]/calcm[9]*(h[3] h[2]); hPx(h[9],calcP[9],calcx[9]); calcT[9]=TT; // 6.determine max recovery with the C weak > C strong assumption temp5=m[2]*(h[3] h[2]); // 7.from the two max rec found, take the smallest one as the a ctual max if(temp5>temp4) Qrecmax=temp4; else Qrecmax=temp5; // 8.impose an effectiveness to the maximum, and compute again T[9], T[3] h[3]=h[2]+Qrecmax*bufrecHE/calcm[2]; // evaluate new h[3] hPx(h[3],calcP[3],calcx[3]); calcT[3]=TT; // to get T [3] h[9]=h[8] Qrecmax*bufrecHE/calcm[8]; // evaluate new h[9] hPx(h[9],calcP[9],calcx[9]); calcT[9]=TT; // to get T[9] // DETERMINING strong boiler exit T[4], source boiler exit T[16] // 1.assume C source < C strong, so T[16] drops to near T[3] for max heat transfer // (100% effectiveness I define as max ht within approach limits, // with the approach limits usually taken as 0 for maximum theoretical heat transfer) calcT[16]=calcT[3]; PTx(calcP[16],calcT[16],calcx[16]); h[16]=hm; // 2.determine T[4] fr om control volume energy balance PTx(calcP[15],calcT[15],calcx[15]); h[15]=hm; h[4]=calcm[15]/calcm[3]*(h[15] h[16])+h[3]; hPx(h[4],calcP[4],calcx[4]); calcT[4]=TT; // 3.determine max recovery with the C source < C strong assumption temp4=m[3]*(h[4] h[3 ]); // 4.assume C source > C strong, such that T[4] rises to T[15] for max ht calcT[4]=calcT[15]; PTx(calcP[4],calcT[4],calcx[4]); h[4]=hm; // 5.determine T[16] from control volume energy balance PTx(calcP[4],calcT[4],calcx[4]); h[4]=hm; h[16]=h[15] cal cm[3]/calcm[16]*(h[4] h[3]); hPx(h[16],calcP[16],calcx[16]); calcT[16]=TT; // 6.determine max recovery with the C source > C strong assumption temp5=m[3]*(h[4] h[3]); // 7.from the two max rec found, take the smallest one as the actual max if(temp5>temp4 ) Qboilermax=temp4; else Qboilermax=temp5; // 8.impose an effectiveness to the maximum, and recompute T[16], T[4] h[4]=h[3]+Qboilermax*bufboilerHE/calcm[3]; // evaluate new h[4] hPx(h[4],calcP[4],calcx[4]); calcT[4]=TT; // to get T[4] h[16]=h[15 ] Qboilermax*bufboilerHE/calcm[15]; // evaluate new h[16] hPx(h[16],calcP[16],calcx[16]); calcT[16]=TT; // to get T[16] // usually it only takes 5 iterations to converge. It will not converge if the // guessed T4 yields the calc T4 which again yields th e guessed T4. First try // to throw force the values out of oscillation by adding a small amount to // the guessed T4. If the oscillation continues, average the guessed and // evaluated T4. They should only differ slightly. if(count2==15) { count2++; ca lcT[4]=calcT[4]+.05; goto iterate; } else if(count2==30) { calcT[4]=.5*(calcT[4]+guess); goto done; } count2++;

PAGE 215

196 if(fabs(guess calcT[4])>.01) { goto iterate; } } // SECOND look at case of complete vaporization else if(calcm[8]<.0001) { cycletyp e=2; calcm[8]=calcm[9]=calcm[10]=0; // no weak flow calcm[5]=calcm[6]=calcm[7]=calcm[1]; // vapor flow = strong flow calcx[5]=calcx[6]=calcx[7]=calcx[1]; // vapor mf = strong mf calcx[8]=calcx[9]=calcx[10]=0; // assign lowest mf to weak calcT[3] =calcT[2]; h[3]=h[2]; //recovery unit not involved without weak flow // DETERMINING strong boiler exit T[4], source boiler exit T[16] // 1.assume C source < C strong, so T[16] drops to T[3] for max ht // (100% effectiveness I define as max ht within appro ach limits, // with the approach limits usually taken as 0 for maximum theor ht) calcT[16]=calcT[3]; PTx(calcP[16],calcT[16],calcx[16]); h[16]=hm; // 2.check if pinch point is violated, and adjust T[16] accordingly Px(.5*(calcP[3]+calcP[4]),calcx[3]); // p. point at average str P in boiler // pinch point occurs in boiler since there is no recovery unit temp1=TT; // min source temperature at the pinch point temp2=hm h[3]; // rise in str solution spec enthalpy to boiling point PTx(.5*(calcP[15]+cal cP[16]),temp1,calcx[15]); temph=hm temp2*calcm[3]/calcm[15]; // enthalpy at 16 with p. point limit hPx(temph,calcP[16],calcx[16]); if(TT>calcT[16]) { // if the pinch point is violated calcT[16]=TT; h[16]=hm; // assign a higher T to state 16 } // 3 .determine T[4] from control volume energy balance PTx(calcP[15],calcT[15],calcx[15]); h[15]=hm; h[4]=calcm[15]/calcm[3]*(h[15] h[16])+h[3]; hPx(h[4],calcP[4],calcx[4]); calcT[4]=TT; // 4.determine max boiling with the C source < C strong assumption tem p4=m[3]*(h[4] h[3]); // 5.assume C source > C strong, such that T[4] rises to T[15] for max ht calcT[4]=calcT[15]; PTx(calcP[4],calcT[4],calcx[4]); h[4]=hm; // 6.check if pinch point is violated, and adjust T[4] accordingly Px(.5*(calcP[3]+calcP[4]),calc x[3]); // pinch point at average boiler P // pinch point occurs in boiler since there is no recovery unit temp1=TT; // min source temperature at the pinch point temp2=hm; // strong solution enthalpy at pinch point PTx(.5*(calcP[15]+calcP[16]),temp 1,calcx[15]); // source h at p. point temp3=h[15] hm; // rise in source spec enthalpy from p. point to 15 temph=temp2+temp3*calcm[15]/calcm[3]; // h at 4 within p. point limits hPx(temph,calcP[4],calcx[4]); if(TT C strong assumption temp5=m[3]*(h[4] h[3]); // 9.from the two max rec found, take the smallest one as the actual max if(temp5>temp4) Qboilermax=temp4; else Qboilermax=temp5;

PAGE 216

197 // 10.i mpose an effectiveness to the maximum, and recompute T[16], T[4] h[4]=h[3]+Qboilermax*bufboilerHE/calcm[3]; //evaluate new h[4] hPx(h[4],calcP[4],calcx[4]); calcT[4]=TT; //to get T[4] h[16]=h[15] Qboilermax*bufboilerHE/calcm[15]; //evaluate new h[16] hPx(h[16],calcP[16],calcx[16]); calcT[16]=TT; //to get T[16] // usually it only takes 5 iterations to converge. It will not converge if the // guessed T4 yields the calc T4 which again yields the guessed T4. If this // oscillation is across a small range, average the 2 after 10 iterations if(count2==15) { calcT[4]=calcT[4]+.05; goto iterate; } else if(count2==30) { calcT[4]=.5*(calcT[4]+guess); goto done; } count2++; if(fabs(guess calcT[4])>.01) { goto iterate; } } // THIRD look at case of par tial vaporization else { cycletype=1; // DETERMINING strong recovery exit T[3], weak recovery exit T[9] // 1.assume C weak < C strong, such that T[9] drops to T[2] for max rec // (100% effectiveness I define as max recovery within approach limits, // wit h the approach limits usually taken as 0 for maximum heat theor rec) calcT[9]=calcT[2]; PTx(calcP[9],calcT[9],calcx[9]); h[9]=hm; // 2.check if pinch point is violated, and adjust T[9] accordingly Px(.5*(calcP[2]+calcP[3]),calcx[3]); // pinch point at ave rage recovery P if(TTcalcT[9]) { // if the pinch point is violated calcT[9]=TT; h[9]=hm; // assign a higher T to state 9 } } // 3.determine T[3 ] from control volume energy balance PTx(calcP[8],calcT[8],calcx[8]); h[8]=hm; h[3]=calcm[8]/calcm[3]*(h[8] h[9])+h[2]; hPx(h[3],calcP[3],calcx[3]); calcT[3]=TT; // 4.determine max recovery with the C weak < C strong assumption temp4=m[2]*(h[3] h[2]); / / 5.assume C weak > C strong, such that T[3] rises to T[8] for max rec calcT[3]=calcT[8]; PTx(calcP[3],calcT[3],calcx[3]); h[3]=hm; // 6.check if pinch point is violated, and adjust T[3] accordingly Px(.5*(calcP[2]+calcP[3]),calcx[3]); // pinch point at average recovery P if(TT
PAGE 217

198 calcT[3]=TT; // if pinch point violated, lower T[3] } // 7.determine T[9] from control volume energy balance PTx(calcP[3],calcT[3],calcx[3]); h[3]=hm; h[9]=h[8] calcm[3]/calcm[9]*(h[3] h[2]); hPx(h[9],calcP[9],calcx[9]); calcT[9]=TT; // 8.determine max recovery with the C weak > C strong assumption temp5=m[2]*(h[3] h[2]); // 9.from the two max rec found, take the smallest one as the actual max if(temp5>temp4) Qrecmax=temp4; else Qrecmax=temp5; // 10.impose an effectiveness to th e maximum, and recompute T[9], T[3] h[3]=h[2]+Qrecmax*bufrecHE/calcm[2]; // evaluate new h[3] hPx(h[3],calcP[3],calcx[3]); calcT[3]=TT; // to get T[3] h[9]=h[8] Qrecmax*bufrecHE/calcm[8]; // evaluate new h[9] hPx(h[9],calcP[9],calcx[9]); calcT[9]=TT; / / to get T[9] // DETERMINING strong boiler exit T[4], source boiler exit T[16] // 1.assume C source < C strong, so T[16] drops to T[3] for max ht // (100% effectiveness I define as max ht within approach limits, // with the approach limits usually taken as 0 for maximum theor ht) calcT[16]=calcT[3]; PTx(calcP[16],calcT[16],calcx[16]); h[16]=hm; // 2.check if pinch point is violated, and adjust T[16] accordingly Px(.5*(calcP[3]+calcP[4]),calcx[3]); // p. point at average str P in boiler if(TT>calcT[3]) { // check if pinch point even occurs in boiler temp1=TT; // min source temperature at the pinch point temp2=hm h[3]; // rise in str solution spec enthalpy to boiling point PTx(.5*(calcP[15]+calcP[16]),temp1,calcx[15]); temph=hm temp2*calcm[3] /calcm[15]; // enthalpy at 16 with p. point limit hPx(temph,calcP[16],calcx[16]); if(TT>calcT[16]) { // if the pinch point is violated calcT[16]=TT; h[16]=hm; // assign a higher T to state 16 } } // 3.determine T[4] from control volume energy ba lance PTx(calcP[15],calcT[15],calcx[15]); h[15]=hm; h[4]=calcm[15]/calcm[3]*(h[15] h[16])+h[3]; hPx(h[4],calcP[4],calcx[4]); calcT[4]=TT; // 4.determine max recovery with the C source < C strong assumption temp4=m[3]*(h[4] h[3]); // 5.assume C source > C strong, such that T[4] rises to T[15] for max ht calcT[4]=calcT[15]; PTx(calcP[4],calcT[4],calcx[4]); h[4]=hm; // 6.check if pinch point is violated, and adjust T[4] accordingly Px(.5*(calcP[3]+calcP[4]),calcx[3]); // pinch point at average boiler P if(TT>calcT[3]) { // check if pinch point even occurs in boiler temp1=TT; // min source temperature at the pinch point temp2=hm; // strong solution enthalpy at pinch point PTx(.5*(calcP[15]+calcP[16]),temp1,calcx[15]); // source h at p. poi nt temp3=h[15] hm; // rise in source spec enthalpy from p. point to 15 temph=temp2+temp3*calcm[15]/calcm[3]; // h at 4 within p. point limits hPx(temph,calcP[4],calcx[4]); if(TT
PAGE 218

199 // 8.deter mine max recovery with the C source > C strong assumption temp5=m[3]*(h[4] h[3]); // 9.from the two max rec found, take the smallest one as the actual max if(temp5>temp4) Qboilermax=temp4; else Qboilermax=temp5; // 10.impose an effectiveness to the maximum, and recompute T[16], T[4] h[4]=h[3]+Qboilermax*bufboilerHE/calcm[3]; // evaluate new h[4] hPx(h[4],calcP[4],calcx[4]); calcT[4]=TT; // to get T[4] h[16]=h[15] Qboilermax*bufboilerHE/calcm[15]; // evaluate new h[16] hPx(h[16],calcP[16],calcx[1 6]); calcT[16]=TT; // to get T[16] // usually it only takes 5 iterations to converge. It will not converge if the // guessed T4 yields the calc T4 which again yields the guessed T4. If this // oscillation is across a small range, average the 2 after 10 ite rations if(count2==15) { calcT[4]=calcT[4]+.05; goto iterate; } else if(count2==30) { calcT[4]=.5*(calcT[4]+guess); goto done; } count2++; if(fabs(guess calcT[4])>.01) { goto iterate; } } done: if(cycletype==1||cycletype==2) { // finding tur bine exit T[6], given turbine exit P[6], mass fraction x[5], turbine inlet P[5], // turbine inlet T[5] and turbine isentropic efficiency PTx(calcP[5],calcT[5],calcx[5]); h[5]=hm;s[5]=sm; Psx(calcP[6],s[5],calcx[6]); h[6]=buftie*(hm h[5])+h[5]; hPx(h[6],c alcP[6],calcx[6]); calcT[6]=TT; // finding absorber vapor inlet T[7], given vapor inlet pressure P[7]=P[6], T[7] slightly less than ambient if(calcT[6]>=bufat lossT[2]) { // ie. if there is no refrig then h[7]=h[6]; // refrig unit acts as a throttle hPx(h[7],calcP[7],calcx[7]); // with the assumed pressure loss calcT[7]=TT; } else calcT[7]=bufat lossT[2]; // with refrig T7 will rise to this value // finding vapor flow meter temperature T[18] PTx(calcP[5],calcT[5],calcx[5]); h[5]=hm; hPx(h[5 ],calcP[6],calcx[6]); calcT[18]=TT; } if(cycletype==1||cycletype==3) { // finding absorber weak inlet T[10], assuming isenthalpic throttle h[10]=h[9]; hPx(h[10],calcP[10],calcx[10]); calcT[10]=TT; } } Calculation of Energy Transfers, Exergy Losses and Efficiencies The following uses standard units (enthalpy in kJ/kg, entropy in kJ/kgK, specific volume in m 3 /kg, pressure in bar, temperature in K, and mass fraction), and is applied

PAGE 219

200 after all the state points are determined with the previous function for simulations. This function can also be used for experimental data, provided all the state points are known. // make sure all units are standard void calculate() { // determine the entropy, enthalpy and availability at the state points n=1; while(n<=10) { PTx(calcP[n],calcT[n],calcx[n]); h[n]=hm;s[n]=sm;v[n]=vm; state[n]=istate; xNH3l[n]=xxNH3l;xNH3v[n]=xxNH3v; xH2Ol[n]=xxH2Ol;xH2Ov[n]=xxH2Ov; PTx(calcP[0],calcT[0],calcx[n]); av[n]=h[n] hm calcT[0]*(s[n] sm); if (xNH3v[n]<0) xNH3v[n]=0; if (xNH3l [n]<0) xNH3l[n]=0; if (xH2Ov[n]<0) xH2Ov[n]=0; if (xH2Ol[n]<0) xH2Ol[n]=0; Q[n]=(xNH3v[n]+xH2Ov[n])*100; n++; } PTx(calcP[15],calcT[15],0); h[15]=hm;s[15]=sm;v[15]=vm; state[15]=istate; PTx(calcP[0],calcT[0],calcx[15]); av[15]=h[15] hm calcT[0]*(s [15] sm); PTx(calcP[16],calcT[16],0); h[16]=hm;s[16]=sm;v[16]=vm; state[16]=istate; PTx(calcP[0],calcT[0],calcx[16]); av[16]=h[16] hm calcT[0]*(s[16] sm); // determine the energy transfers // boiler heat input if(boiler==0) // preference to calculat e from system side Qboiler=(h[4] h[3])*calcm[3]/1000; // boiler input kW else Qboiler=(h[15] h[16])*calcm[15]/1000; // preference to calculate from hot water side qboiler=Qboiler/calcm[1]*1000; // boiler heat input per unit bs kJ/kg if(absorbe r==0) { // preference to calculate from system side if(cycletype==2) { // if all vapor Qabsorber=(h[1]*calcm[1] h[7]*calcm[7])/1000; // kW Iabsorber=(1 calcT[0]/(calcT[1] lossT[1]))*Qabsorber+(calcm[7]*av[7] calcm[1]*av[1])/1000;//kW } else if(cycletype==3) { // if no vapor Qabsorber=(h[1]*calcm[1] h[10]*calcm[10])/1000; // kW Iabsorber=(1 calcT[0]/(calcT[1] lossT[1]))*Qabsorber+(calcm[10]*av[10] calcm[1]*av[1])/1000;//kW } else { Qabsorber=(h[1]*calcm[1] (h[10]*calcm [10]+h[7]*calcm[7]))/1000; // kW Iabsorber=(1 calcT[0]/(calcT[1] lossT[1]))*Qabsorber+ (calcm[7]*av[7]+calcm[10]*av[10] calcm[1]*av[1])/1000; // kW } } else Qabsorber=(h[11] h[12])*calcm[11]/1000; // preference to calculate from coolant side qa bsorber=Qabsorber/calcm[1]*1000; // kJ/kg bs Wpump=(h[1] h[2])*calcm[1]/1000; // kW wpump=Wpump/calcm[1]*1000; // kJ/kg bs

PAGE 220

201 // no option to calculate qrec from strong side strong side has phase change if(cycletype==2) { // if all vapor Qr ec=0;Irecovery=0;Ithrottle=0;qrec=0; } else { Qrec=calcm[8]*(h[8] h[9])/1000; // kW calculated from weak qrec=Qrec/calcm[1]*1000; // kJ/kg bs Irecovery=(calcm[2]*(av[2] av[3])+calcm[8]*(av[8] av[9]))/1000; Ithrottle=calcm[9]*(av[9] av[10])/1000 ; // kW } if(cycletype==3) { // if no vapor Qrefrig=0;Wturbine=0;Iturbine=0;qrefrig=0;wturbine=0; Qabsorber=Qboiler Wpump; qabsorber=qboiler wpump; } else { if(calcT[6]>calcT[1] lossT[1] lossT[2]) { Qrefrig=0;Irefrig=0; } else { Qref rig=(h[7] h[6])*calcm[5]/1000; // kW Irefrig=(1 calcT[0]/(calcT[1] lossT[1]))*Qrefrig+ calcm[6]*(av[6] av[7])/1000;//kW } if(turbine==0) // preference to calculate from system side Wturbine=(h[5] h[6])*calcm[5]/1000; // turbine work kW els e Wturbine=(h[14] h[13])*calcm[13]/1000; // calculate from coolant side Iturbine= Wturbine+(calcm[5]*(av[5] av[6]))/1000; // kW } qrefrig=Qrefrig/calcm[1]*1000; // kJ/kg bs wturbine=Wturbine/calcm[1]*1000; // turbine work per unit bs kJ/kg if (cycletype==1) Isep=(calcm[4]*av[4] calcm[8]*av[8] calcm[5]*av[5])/1000; // kW else Isep=0; Ipump= Wpump+(calcm[1]*(av[1] av[2]))/1000;//kW if(cycletype==1||cycletype==2) { // skip for no vapor if(refrig==1) // factor of inverse of ideal COP factor=(calcT[1] lossT[1] calcT[6])/calcT[6]; else if(refrig==2) // factor of 1 factor=1; else if(refrig==3) // refrigeration included as exergy of refrigeration if(calcT[7] calcT[6]<.001) // ie. the factor is Ec/Qc factor=0; else { PTx(calcP[6],calcT[6],calcx[6]); h[6]=hm; s[6]=sm; PTx(calcP[7],calcT[7],calcx[7]); h[7]=hm; s[7]=sm; factor=(h[6] h[7] calcT[0]*(s[6] s[7]))/(h[7] h[6]); } else if(refrig==0) //refrigeration not included factor=0; } // if the v ariable refrig=4, then the factor remains as is, being the inverse of a typical COP firstlaw=(Wturbine+Wpump+factor*Qrefrig)/Qboiler; //source as reservoir...if exp data, reservoir hst is approximated from T[4] if(source==0) {

PAGE 221

202 Iboiler=(1 calcT[0]/calc T[15])*Qboiler+calcm[3]*(av[3] av[4])/1000; // kW secondlaw=(Wturbine+Wpump+factor*Qrefrig)/((1 calcT[0]/calcT[15])*Qboiler); } // solar source else if(source==1) { Iboiler=calcm[15]*(av[15] av[16])/1000+calcm[3]*(av[3] av[4])/1000; // kW secondl aw=(Wturbine+Wpump+factor*Qrefrig)/(calcm[15]/1000*(av[15] av[16])); } // geothermal source else { Iboiler=calcm[15]*(av[15] av[16])/1000+calcm[3]*(av[3] av[4])/1000; //kW secondlaw=(Wturbine+Wpump+factor*Qrefrig)/(calcm[15]/1000*av[15]); } // dete rmine the recovery and boiler heat exchanger effectivess for exp data // effectiveness is q/qmax if(cycletype==3) firstlaw=secondlaw=0; if(display==0) { // effectivenesses has been assigned for designed cycle recHE=recHEdes; boilerHE=boilerHEdes; g oto down; } // for exp data, Qrec is the weak solution change in recovery unit as the strong solution phase is uncertain, // so base all effectivenss calculation on assuming the weak solution is of lower C, and so drops // at most to T[2] for Qrecmax // note only experimental data with partial boiling are analyzed correctly throughout this program!! // 1.estimate lowest T[9] temp=calcT[9]; calcT[9]=calcT[2]; // changing calcT[9] will not change exp T[9] PTx(calcP[9],calcT[9],calcx[9]); h[9]=hm; // 2.c heck if pinch point is violated, and adjust T[9] accordingly Px(.5*(calcP[2]+calcP[3]),calcx[3]); // pinch point at average recovery P if(TTcalcT[9]) { // if the pinch point is violated calcT[9]=TT; h[9]=hm; // assign a higher T to state 9 } } // 3.calculate maximum heat transfer in recovery Qrecmax=calcm[8]*(h[8] h[9])/1000; // reset calcT[9] from lowest possible T[9] back to actual T[9] calc T[9]=temp; // the strong solution phase is uncertain, so base all effectivenss calculation // on assuming the source is of lower C, and so drops at most to T[3] for Qboilermax // 1.estimate lowest T[16] temp=calcT[16]; calcT[16]=calcT[3]; PTx(calcP[16],ca lcT[16],calcx[16]); h[16]=hm; // 2.check if pinch point is violated, and adjust T[16] accordingly Px(.5*(calcP[3]+calcP[4]),calcx[3]); // p. point at average str P in boiler if(TT>calcT[3]) { // check if pinch point even occurs in boiler temp1=TT; // min source temperature at the pinch point temp2=hm h[3]; // rise in str solution spec enthalpy to boiling point PTx(.5*(calcP[15]+calcP[16]),temp1,calcx[15]); temph=hm temp2*calcm[3]/calcm[15]; // enthalpy at 16 with p. point limit hPx(temph, calcP[16],calcx[16]);

PAGE 222

203 if(TT>calcT[16]) { // if the pinch point is violated calcT[16]=TT; h[16]=hm; // assign a higher T to state 16 } } // 3.calculate maximum heat transfer in boiler Qboilermax=calcm[15]*(h[15] h[16])/1000; // reset calcT[16] fro m lowest possible T[16] back to actual T[16] calcT[16]=temp; recHE=Qrec/Qrecmax; boilerHE=Qboiler/Qboilermax; down: ; } Supporting Files for Program Operation The files listed here are examples of tempvalues.txt and filelist.txt, which are called up during normal operation of the dataprocess.exe program. Storage of Temporary Values The temporary values are called from the file tempvalues.txt when the program is started. This file must be in the same directory as the C ++ source code. The program t hen writes current values to the tempvalues.txt file when exiting. The format of the text file is given below. Note the text should be written in the file as is shown, except where in parentheses. Temperatures C (19 values, one per line) Temperature appr oaches (2 values, one per line) Pressures psig (19 values, one per line) Pressure losses (6 values, one per line) Mass fractions (16 values, one per line) Mass flow rates reading (16 values, one per line) Turbine isentropic efficiency 0.900 Pump isentro pic efficiency 0.800 Recovery heat exchanger effectiveness 0.850 Boiler heat exchanger effectiveness 0.850

PAGE 223

204 Recovery at 100% effectiveness 4.557 Boiler input at 100% effectiveness 5.705 Display integer value 2 Factor for refrigeration inclusi on 0.000 GC calibration constant 1.0678 Absorber analysis preference 0 Boiler analysis preference 1 Refrigeration analysis preference 3 Source analysis preference 1 Turbine analysis preference 0 Cycletype designation 1 Filenames in use 110602c.txt 110 602c 110602cr.txt Filename for range calculations range4.txt Filename for results calculations results.txt Filename for property calculations propjunk.txt Low and high temperature calibrations (15 lines, one low and one high per line) Uncertainties in the thermocouples (14 values, one per line) Low pressure calibrations (5 values, one per line) Uncertainties in the pressures and fs (5 lines, one uncertainty and one full scale value per line) Uncertainty in GC constant 0.0118 Uncertainty in liquid and vapor GC 0.0068 0.0031 Uncertainties in the flow meters and fs (6 lines, one uncertainty and one full scale value per line) Number of thermocouple readings (14 values, one per line) Number of pressure readings (5 values, one per line) Number of GC readings (3 va lues, one per line) Number of flowmeter readings (6 values, one per line)

PAGE 224

205 List of Raw Data Files The raw data files are cataloged into the file filelist.txt, which allows for quick inspection of what has been tested, when it was tested, key operating co nditions, and some comments. The parameters that are recorded into this file are the heat source temperature, boiler temperature, boiler pressure, absorber temperature, absorber pressure, basic solution mass fraction (as an area percentage of ammonia from the GC), and the heat source flow ratio. The data can be flagged if deemed not complete or inaccurate for full analysis. Comments are listed as well. The format is shown below as a truncated listing. This file is added onto when it is modified. file so urce T boiler T boiler P absorb T absorb P bsmf hsfr flag comments name (C) (C) (psig) (C) (psig) (%Area) 030701b 72.298 59.038 52.263 12.143 2.251 42.20 2.863 1 refrig, unsteady END 100502a 91.052 82.877 62.825 37.051 21.360 39.83 4.182 1 Helium added to abs END 102802a 98.743 90.890 78.000 35.658 22.715 39.83 4.543 1 largely approx & unsteady m str, time delay btw m and v mf sampling END

PAGE 225

206 LIST OF REFERENCES Armstead, H. C. H., 1978, Geothermal Energy E. & F. N. Spon Ltd., London. Bhatt, M. S., Srinivasan, K., Krishna Murthy, M. V. and Seetharamu, S., 1994, Absorption Resorption Heating Cycles with the New Working Paris R21 NMP and R21 DMA, Energy Conversion Management Vol. 35, pp. 443 451. Bliem, C. J., 1980 Design and Off design Operation of a Dual Boiling Binary Geothermal Power Plant, AIChE Symposium Series: Heat Transfer pp. 163 172. Bogart, M., 1981, Ammonia Absorption Ref rigeration in Industrial Processes Gulf Publishing Company, Houston. Celata, G. P., Cumo, M. and Setaro, T., 1994, Critical Heat Flux in Upflow Convective Boiling of Refrigerant Binary Mixtures, International Journal of Heat and Mass Transfer Vol. 37, no. 7, pp. 1143 1153. Cengel, Y. A. and Boles, M. A., 1998, Thermodynamics: An Engineering Approach 3 rd ed., McGraw Hill Inc., New York. Chai, L. H., Peng, X. F., Wang, B. X. and Ochterbeck, J. M., 1998, Interfacial Behavior of Growing Bubbles in Conc entration Boundary Layers, International Journal of Heat and Mass Transfer Vol. 41, pp. 3529 3535. Chilingar, G. V., Edwards, L., Fertl, W. and Rieke, H. H. III., 1982, Introduction, Handbook of Geothermal Energy edited by Edwards, L. M., Chilingar, G. V., Rieke, H. H. III and Fertl, W. H., Gulf Publishing Company, Houston, pp. 1 43. Collier, J. G. and Thome, J. R., 1996, Convective Boiling and Condensation Clarendon Press, Oxford. Dickson, M. H. and Fanelli, M., 1995, Geothermal Background, Geot hermal Energy edited by Dickson, M. H. and Fanelli, M., John Wiley and Sons, New York, pp. 1 37. Drbal, L. F., Boston, P. G., Westra, K. L. and Erickson, R. B., 1996, Power Plant Engineering Chapman and Hall, New York. Duan, Z., Mller, N. and Weare, J H., 1996, Equation of State for the NH 3 H 2 O System, Journal of Solution Chemistry Vol. 25, no. 1, pp. 43 50.

PAGE 226

207 El Sayed, Y.M and Tribus, M., 1985, Thermodynamic Properties of Ammonia Water Mixtures: Theoretical Implementation for Use in Power Cycle An alysis, ASME Special Publication: Analysis of Energy Systems ASME AES Vol. 1, pp. 89 95. El Shaarawi, M. A. I. and Al Nimr, M. A., 1990, Equations for Use with Computers to Evaluate the Performance of NH3 H2O Intermittent Solar Refrigerators, Energy C onversion Management Vol. 30, no. 3, pp. 315 327. Energy Information Administration, 2001, Annual Energy Review U.S. Department of Energy. Enick, R. M., Donahey, G. P. and Holsinger, M., 1998, Modeling the High Pressure Ammonia Water System with WATAM and the Peng Robinson Equation of State for Kalina Cycle Studies, Ind. Eng. Chem. Res. Vol. 37, pp. 1644 1650. Frank, M., Kuipers, J. and Swaaij, W., 1996, Diffusion Coefficients and Viscosities of CO 2 + H 2 O, CO 2 + CH 3 OH, NH 3 + H 2 O, and NH 3 + CH 3 OH Li quid Mixtures, Journal of Chemical and Engineering Data Vol. 41, pp. 297 302. Goswami, D. Y., 1995, Solar Thermal Power: Status of Technologies and Opportunities for Research, Heat and Mass Transfer 95, Proceedings of the 2 nd ASME ISHMT Heat and Mass Transfer Conference Tata McGraw Hill Publishers, New Delhi, India, pp. 57 60. Goswami, D. Y., and Xu, F., 1999, Analysis of a New Thermodynamic Cycle for Combined Power and Cooling Using Low and Mid Temperature Solar Collectors, Journal of Solar Energy Engineering Vol. 121, pp. 91 97. Hayden, J. G. and OConnell, J. P., 1975, A Generalized Method for Predicting Second Virial Coefficients, Industrial Engineering Chemistry Process Design Division Vol. 14, no. 3, pp. 209 216. Helvarg, D., 2001, Oi l and Water, Popular Science August 2001, pp. 44 50. Hewitt, G. F., Delhaye, J. M. and Zuber, N., 1982, Multiphase Science and Technology Hemisphere Publishing Corporation, New York. Holman, J. P., 1966, Experimental Methods for Engineers McGraw Hill New York. Hudson, R. B., 1995, Electricity Generation, Geothermal Energy edited by Dickson, M. H. and Fanelli, M., John Wiley & Sons, New York, pp. 39 71. Ibrahim, O. M. and Klein, S. A., 1993, Thermodynamic Properties of Ammonia Water Mixtures, A SHRAE Transactions: Symposia, pp. 1495 1502.

PAGE 227

208 Ibrahim, O. M. and Klein, S. A., 1996, Absorption Power Cycles, Energy Vol. 21, no. 1, pp. 21 27. International Energy Agency, 1991, Guidelines for the Economic Analysis of Renewable Energy Technologies In ternational Energy Agency Workshop on the Economics of Renewable Energy Technologies, Quebec, Canada, International Energy Agency. Kalina, A. I., 1984, Combined Cycle System with Novel Bottoming Cycle, Journal of Engineering for Gas Turbines and Power Vol. 106, October, pp. 737 742. Kuwada, J. T., 1972, Geothermal Power Plant Design, AIChE Symposium Series: Water pp. 439 444. Lu, S., 2001, Thermodynamic Analysis and Optimization of a New Ammonia Based Combined Power/Cooling Cycle, Ph.D. dissertat ion, University of Florida. Magee, J. W. and Kagawa, N., 1998, Specific Heat Capacity at Constant Volume for {xNH 3 + (1 x)H 2 0} at Temperatures from 300 to 520 K and Pressures to 20 Mpa, Journal of Chemical and Engineering Data Vol. 43, pp. 1082 1090. Marston, C. H., 1990, Parametric Analysis of the Kalina Cycle, Journal of Engineering for Gas Turbines and Power Vol. 112, January, pp. 107 116. Modell, M. and Reid, R. C., 1983, Thermodynamics and its Applications Prentice Hall, Inc., Englewood Cliff s, New Jersey. Moran, M. J. and Shapiro, H. N., 1992, Fundamentals of Engineering Thermodynamics John Wiley & Sons, New York. Morrison, G., 1999, Stirling Renewal, Mechanical Engineering May 1999, pp. 62 65. Norton, E., 2001, Ammonia Liquid Recircul ation, ASHRAE Journal Vol. 43, no. 10, pp. 50 51. Nowarski, A. and Friend, D. G., 1998, Application of the Extended Corresponding States Method to the Calculation of the Ammonia Water Mixture Thermodynamic Surface, International Journal of Thermophysi cs Vol. 19, no. 4, pp. 1133 1142. Pillis, J. W., 1993, Expanding Ammonia Usage in Air Conditioning, Proceedings of the 1993 ASHRAE/NIST Refrigerants Conference Gaithersburg, Maryland, pp. 103 107. Polak, J. and Lu, B. C. Y., 1975, Vapor Liquid Equil ibria in System Ammonia Water at 14.69 and 65 Psia, Journal of Chemical and Engineering Data Vol. 20, no. 2, pp. 182 191.

PAGE 228

209 Prausnitz, J. M., Lichtenthaler, R. N. and de Azevedo, E. G., 1999, Molecular Thermodynamics of Fluid Phase Equilibria Prentice Ha ll PTR, Upper Saddle River, New Jersey. Rizvi, S. S. H. and Heldemann, R. A., 1987, Vapor Liquid Equilibria in the Ammonia Water System, Journal of Chemical and Engineering Data Vol. 32, pp. 183 191. Sassen, C. L., van Kwartel, A. C., van der Kooi, H. J. and de Swaan Arons, J., 1990, Vapor liquid Equilibria for the System Ammonia + Water up to the Critical Region, Journal of Chemical and Engineering Data Vol. 35, no. 2, pp. 140 144. Skogestad, S., 1983, Experience in Norsk Hydro with Cubic Equatio ns of State, Fluid Phase Equilibria Vol. 13, pp. 179 188. Smolen, T. M., Manley, D. B. and Poling, B. E., 1991, Vapor Liquid Equilibrium Data for the NH3 H2O System and Its Description with a Modified Cubic Equation of State, Journal of Chemical and E ngineering Data Vol. 36, no. 2, pp. 202 208. Stryjek, R. and Vera, J. H., 1986, PRSV: An Improved Peng Robinson Equation of State for Pure Compounds and Mixtures, The Canadian Journal of Chemical Engineering Vol. 64, April, pp. 323 333. Subbiah, S. a nd Natarajan, R., 1988, Thermodynamic Analysis of Binary Fluid Rankine Cycles for Geothermal Power Plants, Energy Conversion Management Vol. 28, no. 1, pp. 47 52. Szargut, J., Morris, D. R. and Steward, F. R., 1988, Exergy Analysis of Thermal, Chemical and Metallurgical Processes Hemisphere Publishing Comp., New York, pp. 7 39. Tamm, G., Goswami, D. Y., Lu, S. and Hasan, A., 2001, Theoretical and Experimental Investigation of an Ammonia Water Power and Refrigeration Thermodynamic Cycle, Proceedings of the ISES 2001 Solar World Congress Adelaide, Australia, in press. Thorin, E., Dejfors, C. and Svedberg, G., 1998, Thermodynamic Properties of Ammonia Water Mixtures for Power Cycles, International Journal of Thermophysics Vol. 19, no. 2, pp. 501 5 10. Tsiklis, D. S., Linshits, L. R. and Goryunova, N. P., 1965, Phase Equilibria in the System Ammonia Water, Russian Journal of Physical Chemistry Vol. 39, no. 12, pp. 1590 1592. Tsonopoulos, C., 1974, An Empirical Correlation of Second Virial Coeff icients, AIChE Journal Vol. 20, no. 2, pp. 263 272. Twidell, J. and Weir, T., 1986, Renewable Energy Resources E. & F. N. Spon, London.

PAGE 229

210 Valenti, M., 1999, Solar Flares: Technology Hones the Efficiency of Sun Powered Energy Systems, Mechanical Engine ering Power July 1999, pp. 15 17. Vidal, J., 1983, Equations of State Reworking the Old Forms, Fluid Phase Equilibria Vol. 13, pp. 15 33. Villamaan, M. A., Gonzalez, C. and van Ness, H. C., 1984, Excess Thermodynamic Properties for Water/Ethylene Glycol, Journal of Chemical and Engineering Data Vol. 29, pp. 427 429. Wardell, C., 2001a, Blackout, Discover May 2001, pp. 63 67. Wardell, C., 2001b, Nuclear Energy Comes Full Circle, Discover August 2001, pp. 38 43. Weber, L. A., 1999, Estim ating the Virial Coefficients of the Ammonia + Water Mixture, Fluid Phase Equilibria Vol. 162, pp. 31 49. Wilson, E. B., 1952, An Introduction to Scientific Research McGraw Hill, New York. Wolcott, B., 1999, Sun Worship, Mechanical Engineering June 1999, pp. 62 64. Xu, F. and Goswami, D. Y., 1999, Thermodynamic Properties of Ammonia Water Mixtures for Power Cycle Applications, Energy Vol. 24, pp. 525 536. Ziegler and Trepp, 1984, Equation of State for Ammonia Water Mixtures, International Jou rnal of Refrigeration Vol. 7, no. 2, pp. 101 106.

PAGE 230

211 BIOGRAPHICAL SKETCH Gunnar Olavi Tamm was born in New York in 1974. He received his M.S. degree from Rutgers University in 1998 and B.E. degree from The Cooper Union in 1996, both in mechanical engineering. Prior academic research includes buoyancy induc ed flows and solar generated hydrogen power systems. The research presented in this dissertation has been recognized with the 2002 American Solar Energy Society John and Barbara Yellott Graduate Student Award and the 2002 American Society of Mechanical E ngineers Solar Energy Division Graduate Student Award.


Permanent Link: http://ufdc.ufl.edu/UFE0000802/00001

Material Information

Title: Experimental Investigation of An Ammonia-Based Combined Power and Cooling Cycle
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0000802:00001

Permanent Link: http://ufdc.ufl.edu/UFE0000802/00001

Material Information

Title: Experimental Investigation of An Ammonia-Based Combined Power and Cooling Cycle
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0000802:00001


This item has the following downloads:


Full Text










EXPERIMENTAL INVESTIGATION OF AN AMMONIA-BASED
COMBINED POWER AND COOLING CYCLE














By

GUNNAR OLAVI TAMM


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

UNIVERSITY OF FLORIDA


2003















ACKNOWLEDGMENTS

There are a number of individuals who have contributed to the completion of this

dissertation and the enrichment of my educational experience at the University of Florida.

The gracious assistance, guidance and friendship of these individuals are recognized.

As my Ph.D. committee chair, Dr. D. Yogi Goswami provided guidance but

allowed me to follow my own research design. This afforded me the opportunity and

responsibility to think independently and motivate myself on a real project. Dr. Goswami

also encouraged interaction with industry, writing journal papers, soliciting grants and

attending conferences, which have exposed me to the professional stage and built

confidence in my transition towards an engineering career.

My Ph.D. committee members, Dr. Jacob N. Chung, Dr. James F. Klausner, Dr.

Ulrich H. Kurtzweg and Dr. S. A. Sherif, provided direction early on and helped me set

realistic goals. The suggestions made regarding clarity, consistency, and analysis of

results in my Ph.D. proposal were useful in preparing this dissertation and other technical

papers on the study.

The senior engineering technician at the Solar Energy and Energy Conversion

Laboratory, Chuck Garretson, exhibits a demeanor that motivates self-reliance. Skilled

and willing to teach, Chuck is an asset to those willing to learn. His practical assistance,

advice and instruction were critical for my project.

Barbara Graham provided much editorial and secretarial assistance to the solar lab

group in writing papers and reports on our projects. In addition, her activism in solar and









environmental issues helped to stimulate the group and increase our involvement in

related events.

The weekly meeting of the solar lab group was a good arena to practice

presentation and discussion skills. Attentive graduate students in the group, such as

Sanjay Vijayaraghavan, are commended for giving honest criticism, constructive

arguments and intelligent suggestions during and after these meetings. Sanjay and Chris

Martin, also a graduate student, are recognized for their assistance with my experimental

work. As my predecessor, Shaoguang Lu assisted with the first design of the

experimental setup and the initial procurement of system components, granting me

challenging opportunities in his wake.

I reiterate my appreciation to those that contributed to this dissertation and to

making the solar lab a constructive, pleasant, and humorous working environment.

Thanks go to my family for their encouragement and support throughout my lengthy

educational experience. I also acknowledge financial support from the U.S. Department

of Energy, Florida Solar Energy Center and NASA.
















TABLE OF CONTENTS


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

LIST OF TABLES ................................ .. ........... ............................. ix

LIST O F FIG U RE S .............. ......................... ........................... ....................... .. ....... .x

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

A B S T R A C T ................................................................................................................... x v iii

CHAPTER

1 M O TIV A TIO N ................................................................................................ ................

Energy Breakdown and Renewable Resources............................................................ 1
L ow Tem perature R sources .................................................................. .............. 4
S o lar R esou rces ................................................................................................. 4
G eoth erm al R esou rces .... .. ...................................... ........................ .............. 6
W aste H eat R sourcess. ................................................................ .............. 10
B iom ass R sources ........................................... .. ....................... ... ........... .. 10
Available Methods For Thermal Energy Conversion............................................. 10
D irect P ow er P reduction ........................................ ......................... ................. 11
Indirect P ow er P roduction....................................... ........................ .............. 12

2 BACKGROUND AND SUMMARY OF PREVIOUS WORK ...............................17

O v erv iew o f th e C y cle .................................................................................................. 17
Therm odynam ics of the Cycle ................................... ..................... .............. 18
C om prison to O their C ycles ..................................... ...................... .............. 20
T theoretical B background .................................................................. ....................... 2 1
Properties of Ammonia-Water Mixtures.......................................................... 21
Efficiency Based on Source and Sink Temperatures ....................................... 21
Efficiency Based on Energy Transfers.................................................. 23
P previous T heoretical Studies......................................... ......................... .............. 25
Param etric A analysis .... .. ................................. .......................... .............. 26
O p tim iz atio n ........................................................................................................... 2 8
Irreversibility Analysis ....................................................... .............. 33
Sum m ary of Theoretical Studies ........................ ........................................... 35









3 EXPERIM EN TAL SY STEM ..................................... ........................ ................ 36

Ammonia-Water Side ............................. .......... ....................... 37
H ot W after Side ...................................................................................................... 44
C oolant Side ............................................................................................ .............. 44
Instrum entation ............................................................................................................. 47
T h e rm o c o u p le s ....................................................................................................... 4 8
P pressure T ransdu cers............................. ............................................. .............. 49
Gas Chromatograph and Syringe Sampling.......................................... .............. 49
Flow M eters ............................................................................ ... ......................... 51
Data Acquisition Hardware and Analysis Software .................................... .............. 51
S a f e ty ............................................................................................................................ 5 2

4 EXPERIMENTAL METHODOLOGY....................................................... ...............54

Parameters in Simulations and Experiments ................................................ .............. 54
Limits and Selection of Operating Conditions ................. .................................. 56
H eat Source Tem perature ................................................................... .............. 59
B o ile r P re ssu re ....................................................................................................... 6 1
Ambient Temperature .............................................. 62
Basic Solution Ammonia M ass Fraction ............................................... .............. 63
H eat Source Flow R atio ................................................................................... 63
System B behavior ............................................................................................ 63
U uncertainty of M easurem ents ................................................................... .............. 65

5 EXPERIM EN TAL RESULTS ....................................................................................69

V apor G generation .......................................................................................................... 69
O observations .......................................................................................... 70
R covered H eat ....................................................................................... 71
B oiler H eat Input ...................................................................................... 73
Vapor Fraction Leaving the Separator .................................................. .............. 75
Vapor and Weak Solution Ammonia Mass Fractions ......................................... 77
Absorption ................................................... ... .............. 80
O b serve atio n s ........................................................................................................... 8 0
Coolant Flow Rate and Temperature ........................................................ 81
Absorber H eat Rejection .................................................................................. 82
Potential Work Output, Cooling Capacity and Cycle Efficiencies............................... 83
Param etric D dependence ..................................................................... ............... .. 84
Pum p W ork Input .................................................................................... 87
T urbine W ork O utput ..................................................... .............. .............. 88
C o o lin g C ap city ..................................................................................... ... 9 0
F first L aw E ffi ciency ................................................................................... 9 1
Second L aw E efficiency .............................................. ........................................ 92

6 RECOMMENDATIONS FOR FUTURE WORK .................................................. 93



v









Target A reas for Im provem ent.................................................................. .............. 93
Turbine/Generator Set and Refrigeration Unit................................... .............. 93
Absorber Design ........................................ ......... ....................... 94
B oiler H eat E xchanger .................................................................... .............. 94
A m m onia-W ater Pum p ......................................... .......................... .............. 95
Coolant Tem perature Fluctuations ....................... ......................................... 95
H eat Source C capacity .... ................................................................ .............. 96
Control Methods ........................................ ......... ....................... 96
Operating Limits ........................................ ......... ....................... 97
In su la tio n ...................... .................................................................................. . 9 8
Component M odeling in Simulations ................................................ 98
Suggested Strategy for System M modifications .......................................... .............. 98
Short Term Plan ................................... ............................ .. 98
Long Term Plan.............................. .......... ....................... 100

7 C O N C L U SIO N S .......................................................................................................... 10 1

APPENDIX

A A M M O N IA T O X IC IT Y .............................................................................................102

B BINARY FLUID PROPERTY EVALUATION .............................................103

Characteristics of Ammonia and W ater.............. ............................................ 103
Id e al M o d els.............................................................................................................. .. 10 4
Sem i-Em pirical C orrelations.................................... ......................... .............. 105
C ubic E quations of State .................................... ........................ .............. 107
V irial E equations of State ................................... ......................... .............. 108
Gibbs Excess Energy..................................................... ............ .. 108
Law of Corresponding States ...... ........... ............ .................... 109
P erturbation T heory .... ............................................................... .............. 110
G roup C contribution M ethod ............................................................................. 111
Property Correlations Used in the Current Study .............. ................. ................. 111
A m m onia and W ater M ixtures......................................................... ................. 111
Ethylene-Glycol and Water Mixtures ................................ 119

C EXPERIMENTAL COMPONENT LIST............... .........................122

D ata A acquisition and Electrical ................................. ....................... .............. 122
F lu id s......................................................................................................... .......... 12 3
H eat E x ch an g e ............................................................................................................ 12 3
In stru m en tatio n ........................................................................................................... 12 4
P u m p in g ...................................................................................................................... 12 6
S safety ........................................................................................................ ........... 12 7
V alv e s ........................................................................................................ ........... 12 8
V e ssels ...................................................................................................... .......... 12 9









D EXPERIMENTAL PROCEDURES ................ ........ ......................130

C alibration of Instrum entation ................................... ....................... .............. 130
Therm couple C alibration .................................. ....................... .............. 130
Pressure Transducer Calibration ...... ........ .. ........ .................... 131
Gas Chrom atograph Calibration ....... ........ ........ .................... 133
Flow M eter C alibration ..................................... ......................... .............. 134
Syringe Sam pling Techniques ................................................................ .............. 136
Observations on the Behavior of Ammonia-Water Mixtures............................... 138
N orm al O operating Procedures................................... ........................ .............. 139
Startup .................................................................................................. ...... 139
O operation and T testing ................................................................ .............. 142
S h u td o w n .............................................................................................................. 14 4
System M maintenance and Initial Setup ...... ........ ...... .................... 145
D ata A acquisition Setup .................................... .......................... .............. 145
System Draining and Cleaning...... ......... ......... .................... 147
System Charging ...................................... ......... ....................... 149
Idling in Cold W weather ....................................................... ............ .. 151
Idling in Hot W weather ....................................................... .............. 151
E m ergency Procedures.... .................................................................. .............. 152

E UNCERTAINTY ANALYSIS ......................................................... 153

Theoretical Background of Analysis Method......... ..................................... 153
Uncertainty of M measured Values...... ......... ........ .................... 153
Uncertainty of Calculated Values ....... ...... ...... .................... 156
Uncertainty of State Point Measurements................ ........................ 157
Tem perature M easurem ents ................................. ...................... .............. 157
Pressure M easurem ents .................. .......................................................... 159
Ammonia M ass Fraction M easurements...... .......... ..................................... 162
Flow Rate Measurements ........................................................ 165
Uncertainty of Energy Transfers and Efficiencies...... .................................... 167
B oiler H eat Input..... .. .................................. ........................... ............ .. 169
A bsorber H eat R ejection .................................... ........................ .............. 170
C ooling C capacity ....... .. ....................................... ......................... . ........ .. 17 1
Internal H eat R ecovery...................................... ......................... .............. 172
P u m p W ork In put ...... .. ........................................ ........................ .............. 17 3
Turbine W ork O utput ................................................................ .............. 174
V a p o r F ra ctio n .................................................... .. .......................................... 17 4
Boiler Heat Exchanger Effectiveness................ ........................ 175
Recovery Heat Exchanger Effectiveness ....... ......... .................................... 175
F irst L aw E efficiency ......................................................................... .............. 176
Second Law Efficiency ................... ......................................................... 177
Conclusions of the Uncertainty Analysis................ ........................ 178

F COMPUTER PROGRAM FOR DESIGN AND DATA ANALYSIS........................ 179









Program Features ........................... .. .......... ...................................... 179
Property E valuation ... ................................................................. .............. 179
D esign and Sim ulation .............................................................. .............. 180
Experim ental Data Processing ....... .......... ............ .................... 180
Experimental Data Organization and Comparison..................... ................. 181
U uncertainty A analysis ................................................................. .............. 182
U nits Preferences ........................................................... .... .. .. ........ .. .......... .. 182
Calculation Preferences and Calibration Constants ................... ................. 182
Program Structure ...................................... ............................ 182
Property E valuation Functions ...................................................................... ............. 183
Critical Temperature and Pressure of the Mixture......................................... 184
Bubble and Dew Point Temperature, Pressure and Mass Fraction................... 184
Enthalpy of Pure Components and the Mixture...... .................. ................. 186
Entropy of Pure Components and the Mixture...... ................... ................. 187
Specific Volume of Pure Components and the Mixture ................................ 188
Determination of All Properties at a State Point...... ................. ................. 189
Therm odynam ic Analysis Functions ................... .......................... .............. 193
Determination of All State Points in the Cycle................................. ............. 193
Calculation of Energy Transfers, Exergy Losses and Efficiencies................... 199
Supporting Files for Program Operation........................................ 203
Storage of Tem porary Values ....... .......... ............ .................... 203
L ist of R aw D ata Files ................................................................ .............. 205

LIST OF REFEREN CE S ..................... ................................................................. 206

BIOGRAPH ICAL SKETCH .................. .............................................................. 211

























viii















LIST OF TABLES


Table page

4.1 Fluid and component parameters in the simulation and experiment .......................56

4.2 Lim its of prim ary operating conditions ................................................. ................ 59

4.3 Experimental control measures for maintaining operating parameters ...................65

A .1 A m m onia exposure lim its .................. ........................................................... 102

B. 1 Coefficients for pure water and pure ammonia.............................................. 114

B.2 Coefficients for the ammonia-water Gibbs energy functions ................................116

B.3 Coefficients for determining bubble and dew point temperatures, critical
temperatures, and critical pressures of ammonia-water mixtures ......................117

B.4 Empirical constants in ethylene glycol and water mixture property evaluation ......121

E.1 Chauvenet's criterion for elimination of data points ..................... ................... 155

E.2 Uncertainties in single measurements within confidence intervals........................156

E.3 Device history of pressure measurement in the system................. ...................162

E.4 Liquid flow m eter uncertainties...... ............. ............ ..................... 167

E.5 The uncertainties of calculated values and their dependence on measurements for a
typical test of the experimental system, within 90% confidence .......................168















LIST OF FIGURES


Figure page

1.1 U .S. energy consum ption in 1999 ............................................................ ...............1...

1.2 Installed U.S. capacity of renewable electricity production in 1999 .........................2...

1.3 U .S. energy research funding in 2001 ...................................................... ...............3...

1.4 The consumption of geothermal and solar energy in the U.S..................................8...

1.5 Temperature-entropy diagram for the Stirling cycle ........................ ..................... 13

1.6 Schem atic diagram of the K alina cycle.................................................. ................ 16

2.1 Schematic of the power and cooling cycle concept ......................... ..................... 19

2.2 Temperature-entropy diagram of an ideal cycle with a sensible heat source and heat
sin k ...................................................................................................... ....... .. 2 2

2.3 Effect of turbine inlet pressure on the thermal efficiency of the cycle....................27

2.4 Effect of turbine inlet pressure on the cooling capacity of the cycle.......................27

2.5 Efficiencies of the optimized cycle at various heat source temperatures, optimized for
second law efficiency .................................................................. .............. 29

2.6 Pressure ratio of the optimized cycle at various heat source temperatures, optimized
for second law efficiency ........................................ ....................... ................ 30

2.7 Ratio of refrigeration to work of the optimized cycle at various heat source
temperatures, optimized for second law efficiency ..........................................31

2.8 Exergy destruction of the optimized cycle at various heat source temperatures,
optim ized for second law efficiency................................................. ................ 32

3.1 Schematic of the experimental system concept .............. ..................................... 36

3.2 Schematic of the experimental system components for the system side .................38

3.3 Photograph of the system side and hot water side of the experiment......................39









3.4 Photograph of the system side of the experiment..................................................39

3.5 Photograph of the system side pump and expansion tank .....................................40

3.6 Photograph of the system side heat exchangers.....................................................40

3.7 Schem atic of the absorber design........................................................... ................ 42

3.8 Photograph of the flow indicator to the storage tank.............................................43

3.9 Photograph of the hot water side of the experiment .............................................43

3.10 Schematic of the experimental system components for the hot water side ..............45

3.11 Schematic of the experimental system components for the coolant side...............46

3.12 Photograph of the coolant side of the experiment................................................47

3.13 Location of data measurements for system analysis...........................................48

3.14 Photographs of pressure measurement instruments on the system. ........................49

3.15 Photograph of a syringe sampling port and sight glass on the absorber................50

3.16 Photograph of the gas chromatograph setup ........................................................50

3.17 Photograph of the temperature displays and coolant flow meter...........................51

3.18 Photograph of the storage tank and emergency tank ...........................................53

4.1 Schematic of the cycle concept used in experiment simulations, showing the main
fluid parameters ............................ ............ ............................. 55

4.2 Expected turbine work output per unit basic solution flow, based on simulation with
n o lo sse s ............................................................................................................. .. 5 7

4.3 Expected refrigeration output per unit basic solution flow, based on simulation with
n o lo sse s ............................................................................................................. .. 5 8

4.4 Vapor fraction of ammonia-water fluid leaving the boiler, based on simulation with
n o lo sse s ............................................................................................................. .. 6 0

4.5 Boiler heat input per unit basic solution flow, based on simulation with no losses ...61

4.6 Expected absorber heat rejection per unit basic solution flow, based on simulation
w ith n o lo sse s ......................................................................................................... 6 2

4.7 Expected first law efficiency, based on simulation with no losses and with typical
lo sse s .................................................................................................... ........ .. 6 7









4.8 Expected second law efficiency, based on simulation with no losses and with typical
lo sse s .................................................................................................... ........ .. 6 8

5.1 Qualitative representation of the change in properties during the boiling of the strong
so lu tio n ................................................................................................................ .. 7 0

5.2 Expected internal heat recovery per unit basic solution flow, based on simulation
with typical losses ... ................................................................ 72

5.3 Boiler heat input per unit basic solution flow, for various vapor fractions leaving the
b o ile r ................................................................................................................ . 7 4

5.4 Boiler heat exchanger effectiveness for the initial and improved systems .............74

5.5 Vapor fractions in the initial system tests for various boiler temperatures and
pressures, with a basic solution ammonia mass fraction of 45.6% ........................76

5.6 Vapor fractions in the improved system tests for various boiler temperatures and
pressures, with a basic solution ammonia mass fraction of 38.3% ........................76

5.7 Ammonia mass fraction in the weak solution for various boiler exit temperatures and
pressures, with a basic solution ammonia mass fraction of 45.6% ........................78

5.8 Ammonia mass fraction in the vapor for various boiler exit temperatures and
pressures, with a basic solution ammonia mass fraction of 45.6% ........................78

5.9 Ammonia mass fraction in the weak solution and vapor for a basic solution ammonia
mass fraction of 45.6%, for various boiler pressures.........................................79

5.10 Coolant flow rates for various temperature differences between the coolant inlet and
th e ab so rb er p o ol.................................................................................................... 82

5.11 Absorber heat rejection per unit basic solution flow, for various vapor fractions
leav in g th e b oiler.................................................................................................... 83

5.12 Simulated first law efficiency for various turbine and pump isentropic efficiencies,
and w ith other typical losses............................................................. ................ 85

5.13 Simulated second law efficiency for various turbine and pump isentropic
efficiencies, and w ith other typical losses......................................... ................ 85

5.14 Simulated cooling capacity per unit basic solution flow, for various turbine
isentropic efficiencies and heat source temperatures........................................86

5.15 Pump work input per unit basic solution flow, for various pressure rises from the
absorber to the boiler ..................................................................... 87

5.16 Measured pressure loss across the vapor bubble inlet to the absorber...................89









5.17 Measured pressure loss across the weak solution spray inlet to the absorber...........89

5.18 Expected work output per unit basic solution flow from a 90% efficient turbine, for
observed vapor fractions and expansion ratios................................. ................ 90

5.19 Expected first law efficiency with a 90% efficient turbine, for observed vapor
fractions and expansion ratios........................................................... ................ 91

5.20 Expected second law efficiency with a 90% efficient turbine, for observed vapor
fractions and expansion ratios........................................................... ................ 92

B. 1 Bubble and dew point diagram for a non-azeotropic mixture................................104















NOMENCLATURE


A area [m3]

C gas chromatograph calibration constant

Cp isobaric heat capacity [kJ/K]

COP coefficient of performance

D diffusion coefficient [m2/s]

E exergy [kW]

AE change in exergy [kW]

F general function or general property

G Gibbs energy [kW]

H enthalpy [kW]

HE heat exchanger effectiveness

K number of samples in the set

N number of samples in the population

P pressure [bar]

Q heat transfer rate [kW] or volumetric flow rate [m3/s]

R molar specific gas constant [kJ/K] or ratio of maximum deviation to

standard deviation

S entropy [kW]

T temperature [C]

V volume [m3] or voltage [V]









W work transfer rate [kW]

Cp isobaric specific heat capacity [kJ/kgK]

g molar specific Gibbs energy [kJ/kmol] or acceleration due to gravity

h molar specific enthalpy [kJ/kmol]

hst heat source temperature [K]

m mass flow rate [kg/s]

q specific heat transfer rate [kJ/kg]

s molar specific entropy [kJ/kmol]

v molar specific volume [m3/kmol]

w specific work transfer rate [kJ/kg]

x mass fraction of ammonia in the liquid [kg NH3 (liq) / kg total (liq)] or

general measurement


x average of a general measurement


x mole fraction of ammonia in the liquid [mol NH3 (liq) / mol total (liq)]

y mass fraction of ammonia in the vapor [kg NH3 (vap) / kg total (vap)]


y mole fraction of ammonia in the vapor [mol NH3 (vap) / mol total (vap)]

Greek

91 first law (thermal) efficiency

92 second law (exergy) efficiency

Cp pump isentropic efficiency

9t turbine isentropic efficiency

i dynamic viscosity [Pa-s]

fi density [kg/m3]










Co




subscripts

0

B

H

H20

L

NH3

a

abs

avg

b

c

cycle

d

ex

f

g

h

hot

hs

i


standard deviation

uncertainty



reference value or dead state

reference value

high

water

liquid or low

ammonia

ammonia or absorber

absorber

average

bubble or boiler

critical or cooling or calibration

based on the whole cycle

dew

exit

final or float

vapor

heating

hot water

heat source

pertaining to component i or initial










in

m

max

min

p

r

rec

refrig

s

str

sys

t

tur

vap

w

weak

superscripts

E

L

g

in

mix


inlet

mixture or measured or pertaining to the mean

upper limit

lower limit

pump

reduced

recovery

based on refrigeration

standard conditions or superheater

strong solution

system

turbine

turbine

vapor

water

weak solution




excess

liquid

vapor

inlet

mixing


xvii















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

EXPERIMENTAL INVESTIGATION OF AN AMMONIA-BASED
COMBINED POWER AND COOLING CYCLE

By

Gunnar Olavi Tamm

May 2003


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

A novel ammonia-water thermodynamic cycle, capable of producing both power

and refrigeration, was proposed by D. Yogi Goswami. The binary mixture exhibits

variable boiling temperatures during the boiling process, which leads to a good thermal

match between the heating fluid and working fluid for efficient heat source utilization.

The cycle can be driven by low temperature sources such as solar, geothermal, and waste

heat from a conventional power cycle, reducing the reliance on high temperature sources

such as fossil fuels.

A theoretical simulation of the cycle at heat source temperatures obtainable from

low and mid temperature solar collectors showed that the ideal cycle could produce

power and refrigeration at a maximum exergy efficiency, defined as the ratio of the net

work and refrigeration output to the change in availability of the heat source, of over

60%. The exergy efficiency is a useful measure of the cycle's performance as it compares

the effectiveness of different cycles in harnessing the same source.


xviii









An experimental system was constructed to demonstrate the feasibility of the

cycle and to compare the experimental results with the theoretical simulations. In this

first phase of experimentation, the turbine expansion was simulated with a throttling

valve and a heat exchanger. Results showed that the vapor generation and absorption

condensation processes work experimentally. The potential for combined turbine work

and refrigeration output was evidenced in operating the system. Analysis of losses led to

modifications in the system design, which were implemented to yield improvements in

heat exchange, vapor generation, pump performance and overall stability.

The research that has been conducted verifies the potential of the power and

cooling cycle as an alternative to using conventional fossil fuel technologies. The

research that continues is to further demonstrate the concept and direct it towards

industry. On the large scale, the cycle can be used for industrial power production or as a

central power plant for a community, with refrigeration produced as required by the

application. On the small scale, an affordable residential or commercial unit could allow

independent electricity generation for the home or business while also cooling it.















CHAPTER 1
MOTIVATION

To identify the intended domain of the proposed power and cooling cycle, an

overview of low temperature energy resources is presented in this chapter, giving insight

into the requirements of systems that convert them to useable energy. A discussion of

available methods for low temperature resource conversion follows as background and

motivation for the development of the proposed power and cooling cycle as an alternative

to the established technologies.

Energy Breakdown and Renewable Resources

The global energy consumption in 1999 was 381.9x1015 Btu, with the U.S.

contributing a quarter of the total energy usage. Fossil fuels account for a majority of the

energy supply, and nuclear and renewable resources make up the remainder.


biomass 3.3%/ geothermal 0.4% solar <0.1%
hydro 3.6%w <

nuclear 8.0%1----

petroleum
39.1%


coal 22.4%




natural gas 23.0%


Figure 1.1. U.S. energy consumption in 1999 (Energy Information Administration,
2001).










Renewable energy is defined as a continuous and self-propagating flow of energy

in the natural environment. There is no initiation energy required to release the stored

potential, as in the case of finite or non-renewable energy resources (Twidell and Weir,

1986). As renewable energy is cyclical in the environment, the exploitation of it is

naturally with less environmental impact than in using finite energy resources.



hydro
80000 -












8L 4DOO


20 4000o




biornam55


9eolhernmal rp
g__1nI I I-1
Solar PV id




Figure 1.2. Installed U.S. capacity of renewable electricity production in 1999 (Energy
Information Administration, 2001).

The installed capacity of renewable electricity production is shown in Fig. 1.2.

Hydro is by far the most developed renewable resource. The major reason that the global

dependence on finite energy resources has not shifted faster to renewable energy is cost.

In 2001, only 10% of the U.S. electricity needs were met with renewable resources, while









54% came from coal, 20% from nuclear and 16% from natural gas (Wardell, 2001a). In

looking at 2001 production costs for power companies, electricity from coal only costs

$0.021 per kilowatt hour and nuclear even less at $0.017/kWh. Meanwhile, natural gas

and wind generation costs twice as much at $0.04/kWh, and other renewable resources

cost much more (Wardell, 2001b). As the renewable market grows and technology

improves, these figures will improve but it will take some time.

The U.S. government realizes the environmental and political importance of

diversifying its energy market from fossil fuels into alternatives, as does the rest of the

world. Federal research spending in the U.S. in 2001 was $955 million, with Fig. 1.3

showing the largest allotment for coal. However, considering the large current U.S.

dependence on coal, it is positive to note that only 27% of research dollars were for coal

research and 30% for renewable resources.


wind 3.7% geothermal 2.6%
natural gas 4.2%

h y d r o g e n 4 7 % c a 2 7 2

nuclear6.3%






petroleum 7.3%

biomass 8.4% fusion 25.7%
solar 9.4%


Figure 1.3. U.S. energy research funding in 2001 (Helvarg, 2001).









Low Temperature Resources

The proposed power and cooling cycle is designed for use with low temperature

resources, which can include solar and geothermal. The combustion of biomass, although

a renewable resource, produces a resource temperature that exceeds the scope of the

proposed cycle's design. Indirectly, finite resources can also be used. The proposed cycle

can serve as a bottoming cycle for conventional power plants, which reject spent fluids

that are cooled beyond primary use but can serve as the source for a low temperature end

cycle. In this way more work can be extracted to improve the overall resource utilization

of conventional systems.

The following is an overview of resources that can be used to operate the

proposed cycle. Namely solar, geothermal, and waste heat resources are discussed, to

establish their potentials for utilization and to give background on their development.

Solar Resources

The approximate solar radiation arriving on the earth's surface is 1000 W/m2,

reduced by atmospheric interaction from 1377 W/n? on average arriving extra-

terrestrially (Twidell and Weir, 1986). The actual radiant flux density varies by 1.5%

according to solar fluctuations and by 4% according to the distance of the earth from

the sun. Ultraviolet light accounts for 9% of the electromagnetic spectrum of solar

radiation, visible light for 45% and infrared for 46%. The total daily solar flux to the

earth is about 3-30 MJ/m2, depending on the location, time of the year and weather

(Twidell and Weir, 1986). Solar radiation arrives on the earth's surface as beam and

diffuse radiation. On a clear day, the beam radiation can account for 90% of the total

while on an overcast day diffuse radiation can account for 100% of the total (Twidell and

Weir, 1986). Only beam radiation can be focused.









Owing to the electromagnetic nature of solar radiation, it can be converted

directly to electricity through photovoltaic means. Photovoltaic cells were first introduced

in 1954, and developed largely for the space industry. Photovoltaic devices operate with

the principal of solar radiation exciting electrons to separate from certain materials and

move in the form of a current to produce electricity. Photocells typically produce

potentials of 0.5 V, which can be compounded and current densities of 200 A/m2 with

clear sky radiation of 1000 W/m2 (Twidell and Weir, 1986). Typical efficiencies of

commercial units are 10-15%, producing 1-2 kWh/m2 per day. The National Renewable

Energy Laboratory set a laboratory thin film solar cell conversion record of 18.8% in

1998 (Wolcott, 1999). Other PV technologies can extend efficiencies by a few more

percent, although at a very high cost. The cost of solar cells has dropped sharply from

$100 per Watt in 1974 to about $5 per Watt in 2001. However, the cost should fall to less

than $1 per Watt to become commercially competitive in the electricity market.

Solar thermal technology is much more competitive currently than photovoltaics

for residential use. Domestic hot water accounts for 18% of a home's energy usage,

which has prompted the development of cheap solar thermal collection at a cost of $100

to $300 per square meter (International Energy Agency, 1991; Valenti, 1999). With flat

plate collectors, temperatures around 100 C can be achieved in the collection medium. In

regards to capital cost and energy savings, residential and commercial solar thermal

heating systems have reasonable payback periods and have gained popularity globally.

The typical annual efficiency of solar hot water systems is 35-40% (International Energy

Agency, 1991).









Solar thermal use for power generation has been limited to mid and high

temperature collection, as efficient power cycles for low temperature resources have not

been thoroughly explored. The power production cost has been found to go down

dramatically with system size, so most systems are on the large commercial scales.

Systems using parabolic trough concentrators get transfer fluid temperatures around 700

C. Parabolic dish concentrators can achieve temperatures up to 3000 K (Twidell and

Weir, 1986). For large solar systems, two methods are used for mid temperature cycles.

With distributed collection, many collectors are networked which heat a fluid to high

temperatures. This fluid can be steam, which is expanded through a turbine, or an

intermediary fluid such as ammonia, which is dissociated to store chemical energy. The

recombination gives off heat that can be continuously extracted, even at night. The heat

from this can power a steam turbine. Power towers are the other method, with an array of

reflectors focused on a single collection point (Twidell and Weir, 1986). Both of these

methods for solar thermal electricity are infeasible on residential scales, as the cost would

be tremendous to the homeowner. The typical efficiency for mid and high temperature

solar thermal power generation is about 20% (International Energy Agency, 1991).

The cost of solar collection is minimized in using flat plate collectors, although

the temperatures produced in the transfer fluid are not sufficiently high to efficiently fuel

conventional power cycles. The proposed cycle addresses this issue by being able to

utilize low temperature resources, providing a cost savings potential by being able to use

inexpensive flat plate collectors.

Geothermal Resources

Geothermal energy is loosely defined as the natural heat of the earth. Practically,

geothermal energy is economically viable if it is concentrated into finite regions, and









close enough to the surface so it can be readily extracted. The heat flow outwards from

the earth is relatively constant geographically, with pockets of variation where there are

cracks in the crust in tectonically active regions. This is where geothermal resources

abound.

The three primary geothermal sources can be characterized as hot-igneous,

conduction-dominated and hydrothermal (Chilingar et al., 1982; International Energy

Agency, 1991). Hot-igneous sources include molten rock and dry rock. Molten rock is

liquid magma chambers that are too high in temperature (650-1200 C) and too deep (>3

km) to harvest any energy from with current technologies. Dry rock is solidified magma,

closer to the surface at less than 3 km, and cooler at less than 650 C. A dry rock

geothermal system has been proven at Los Alamos by creating a heating water loop

through it, but at too much cost to become practical (Chilingar et al., 1982). Conduction

type sources transfer heat to the surface through conduction from great depths. In order to

reach temperatures over 100 C to become useable in power cycles, the well depth would

need to exceed 5-10 km which becomes too expensive to be practical. The hydrothermal

sources are the most convenient and common geothermal sources to harvest. Hot water

and steam saturate porous and permeable rocks, with natural convection currents moving

the hot fluids towards the surface while cooler fluids sink.

The total heat flow from the earth is estimated at 42 x 1012 W, or roughly 82

mW/m2 on the surface. The internal generation is largely by radioactive decay of isotopes

of uranium, thorium and potassium. The thermal content of the earth is vast, estimated at

12.6 x 1024 MJ above a reference temperature of 15 C. However, the fraction of this

geothermal energy that can be utilized is rather small. Geothermal energy for electrical










generation has been exploited in Italy for nearly 100 years, beginning in 1904. New

Zealand has been exploiting geothermal power for 50 years, and the western U.S. for 40

years (Armstead, 1978; Chilingar et al., 1982). Fig. 1.4 shows the early development of

geothermal power was well before solar. Most U.S. geothermal sites are on federal lands,

which were opened up for private leasing by energy companies in 1970 with the

Geothermal Steam Act (Kuwada, 1972). In 1993 the global geothermal power generation

was only 5915 MWe, of which 43.9% was generated in the U.S., 15% in the Philippines,

12.7% in Mexico, 10.8% in Italy, and the balance in other countries. However the

industry is gaining ground internationally, as the total generation was up 72% from 10

years prior (Dickson and Fanelli, 1995).


400



3OO
M 300
0


0
'. 200
E
-n


100
4)


- geothermal
.. solar


0+-
1960


1970 1980 1990


2000


Figure 1.4. The consumption of geothermal and solar energy in the U.S. (Energy
Information Administration, 2001).









A deterrent for the geothermal industry growth is the economical risk in

identifying the source. Location of geothermal sources can prove costly. Indications of

potential sources may include geysers, seismic or volcanic activity, and hot springs.

Given the initial surface testing, shallow boreholes are made and finally deep wells. The

success rate rests at about 10%, with a $1 million cost for each site that is tested, in a

1982 estimate (Chilingar et al., 1982). The high cost has hindered extensive exploration.

Once a geothermal site is found, it can also be difficult to determine the total

available energy of the site, which is necessary to gauge the worth of the investment.

Strictly speaking, geothermal energy is not locally renewable. Exploitation of geofluids

will undoubtedly prompt a geological response, which could result in a finite lifetime for

a geothermal site before it cools off (Dickson and Fanelli, 1995). So although the global

geothermal potential is virtually limitless, local concentrations with high geothermal

gradients are bounded and not renewable (Armstead, 1978). Estimates must be made for

the geothermal gradient, heat flow to the surface, and heat capacity of the fluid holding

rocks, which are often only accurate to within one order of magnitude. The potential site

must be selected prudently.

As an example of the resource potential stored at a geothermal site, the

Appalachian region of Virginia, North Carolina and Tennessee could provide the

geothermal energy equivalent of 30-35 trillion barrels of crude oil from its dry rock heat

store (Chilingar et al., 1982). The availability of geothermal resources is not disputed.

However, economical exploitation of such an immense resource becomes the issue in

resource location and utilization. The immediate environmental impact of geothermal

systems is also not negligible. As the geofluid is not pure water, there may be issues with









surface water pollution, air pollution from gases dissolved in the geofluid that are freed as

it cools, and underground pollution. Other concerns likewise are not taken lightly,

although it is anticipated that geothermal energy will have much less long term

environmental impact than conventional energy sources (Dickson and Fanelli, 1995).

Waste Heat Resources

Waste heat resources can be loosely defined as the unused energy of conventional

power cycles. This can be from conventional fossil fuel plants, from biomass combustion

plants, or nuclear reactors. The waste heat is rejected in fluids that are too low in

temperature for useful conversion in the primary cycle but can be hot enough to power a

bottoming cycle. This allows extraction of more work from the original energy source,

and raises the overall efficiency of the system.

Biomass Resources

Biomass is defined as energy derived from the sun and converted through plant

photosynthesis into organic material. The stored energy can be in the form of plants or

animal waste. The largest potential for biomass harvesting is in wood products. The

estimated annual global resource of biomass is 170 billion tons, with less than 1%

currently harvested (International Energy Agency, 1991). The most common conversion

scheme for biomass is direct combustion, although the fuel can be converted to more

practical liquid, gas or solid forms. Biomass combustion occurs at 500-600 C, which is

beyond the design of the proposed cycle's use. However, biomass is mentioned here as it

is used as an energy source in comparative cycles.

Available Methods For Thermal Energy Conversion

The thermal conversion process is initiated with the collection of the resource

heat. The heat in the collection fluid is either converted to power directly, or the fluid









serves as a storage medium to later power another working fluid. The advantage of the

storage medium is that it allows for off-hour conversion of solar radiation. The

disadvantage is that multiple intermediary conversion processes all exhibit additional

inefficiencies.

Some methods of conversion to electricity are specific to the resource, and some

are applicable to multiple resources. Available methods for thermal conversion of low

and mid temperature resources are presented for comparison to the proposed cycle.

Direct Power Production

Conversion methods that exhibit direct power production include processes in

which the fluid heated by the source is the only intermediary, such that power is produced

from it directly as the working fluid.

Solar steam expansion. High temperature steam is possible using solar

concentrators. Water is pumped to high pressures, then boiled in the concentrator. The

high temperature and pressure steam then expands through a turbine producing power.

The steam is condensed and can be recycled back through the system to preserve exergy.

Geothermal steam expansion. Geothermal plants operate with a variety of

cycles. In the simplest cycle, steam is separated out from the geofluid and allowed to

expand through a turbine, exiting to the atmosphere. These systems tend to be the

cheapest in capital cost ($1000/kWnet) for geothermal use, although they require roughly

twice the steam input per unit power production than systems that condense the expanded

steam. The goal of the condensation is to expand the steam through the turbine to much

lower pressures than atmospheric, allowing for greater turbine work output. The

condensed steam must then be pumped to atmospheric pressures for discharge.

Drawbacks of the condensing design include additional components needed to pump non-









condensable gases from the condenser, and cooling towers to extract heat from the

condenser. This all adds to the capital cost by roughly 50%, although overall the cost is

less per amount of steam extracted (Hudson, 1995).

Solar chimney. Based on the principles of buoyancy, solar heated air rises

through a converging tower, speeding up as it rises. The fast moving air rises through a

turbine, which is coupled to a generator to produce electricity. The efficiencies of solar

chimneys are on the order of 1-2% percent, and they must be extremely tall to produce

significant output. The design is simple but becomes impractical at great heights.

Indirect Power Production

Conversion methods that exhibit indirect power production include processes in

which the fluid heated by the source is an intermediary transfer fluid, which subsequently

transfers energy to the working fluid that produces power. The power production can

occur simultaneously with the heat extraction form the source, or the charged transfer

fluid can be stored for later generation.

Stirling engine. The allure of the Stirling engine has been the high potential

efficiency, operating ideally with the same efficiency of a Carnot cycle between the same

source and sink reservoir temperatures, as shown in Fig 1.5. Although the cycle was first

developed in 1816, it subsided to the internal combustion engine until recent growing

interest (Morrison, 1999). The basic cycle includes an isothermal compression, heat

addition at constant volume, isothermal expansion, and finally heat rejection at constant

volume.

The Stirling engine is an external combustion engine and tends to produce less

pollution than internal combustion engines (Moran and Shapiro, 1992). It can be powered

by concentrated solar power, hot-igneous geothermal resources, or with the heat from










burning biomass (Morrison, 1999). Biomass is combusted in the 500-600 C range,

similar to mid temperature solar and geothermal collection, at which the Stirling cycle

exhibits high thermal efficiencies over 60%. Low temperature solar and geothermal can

also be used to fuel Stirling engines, but at lower thermal efficiency limits of 20-35%.






H 2 3





o. o




L- 1 4



entropy


Figure 1.5. Temperature-entropy diagram for the Stirling cycle.

Rankine cycle. Rankine cycles used as secondary or binary cycles in low

temperature applications have been limited to geothermal resources. Binary Rankine

systems are the most common of geothermal power cycles. Solar power generation at low

temperatures is too costly using Rankine cycles, and so solar thermal applications have

been limited to mid and high temperature regimes.

For low temperature resource fluids under 150 C in geothermal systems, it

becomes more economical to use a secondary Rankine cycle for power production than a

direct power production method. Typically the secondary fluid for geothermal

applications has been isobutane (Bliem, 1980), although other fluids can be used









(Armstead, 1978). The working fluid has a lower boiling point than water, such that the

evaporation process in the Rankine cycle typically occurs at low temperatures. The high

pressure working fluid is vaporized and expands through a turbine, finally being

condensed by heat exchange with cooling water.

For geothermal applications the heat source is a geothermal brine, which is

reinjected into the ground after heat is extracted from it by the working fluid. Rankine

cycles are appealing in situations that require the geofluid to remain liquid, as it can be

pressurized during the extraction and reinjection processes to prevent flashing. Flashing

is harmful to the well as it allows for fouling by dissolved substances, leaving a residue

on components (Hudson, 1995; Kuwada, 1972).

Improvements to the geothermal Rankine cycle can be made in the form of

supplemental heating. It has been shown that raising a geothermal low temperature

resource temperature with an additional heating from solar, biomass, fossil fuel, or other

means results in a large payback from the primary Rankine cycle. The extra energy

output from the combined system is about 4 times the supplemental energy added to it

(Subbiah and Natarajan, 1988).

Low to mid temperature geothermal systems operating with binary Rankine

cycles can utilize source temperatures of 85-175 C. Typical plant costs depend highly on

the geothermal fluid temperature. For a 140 C resource temperature, the plant capital

cost will be about $1900/kWnet, excluding the cost of the wells (Dickson and Fanelli,

1995).

Kalina cycle. Improvements in the thermal efficiency over the conventional

Rankine cycle can be achieved by increasing the number of boiling steps, operating at









supercritical conditions, or using a mixed working fluid. The mixed working fluid

provides a varied boiling temperature, but with a conventional Rankine cycle this will

cause a varied condensation temperature for only moderate overall improvement (Kalina,

1984).

The Kalina cycle, shown in Fig. 1.6, is designed for use as a bottoming cycle or

with low temperature heat sources. A multicomponent working fluid is used, with a

concentration dependent boiling temperature. This allows for a good thermal match in the

boiler with the sensible heat source and yields less exergy loss. Better heat utilization

allows for lower component costs of the evaporator and condenser, as less surface area is

required for the same power output. Besides the variable boiling temperature being able

to extract more heat from the heat source, there is also more internal heat recovery. A

high pressure stream can be boiled by a low pressure condensing stream, since the low

pressure stream can have phase changes at higher temperatures than the high pressure

stream according to the ammonia mass fraction of the mixture. This cannot happen in a

single component Rankine cycle (Enick et al., 1998).

Parametric analysis and optimization were performed for the Kalina cycle

operating with a binary ammonia and water mixture, with properties determined

according to the method outlined by El-Sayed and Tribus. Analysis showed turbine inlet

concentration and separator temperature as the dominant parameters. The vapor exiting

the turbine is then used for distillation of the basic solution. This is more efficient than

using it for evaporation (Marston, 1990).

The addition of the Kalina cycle as a bottoming cycle raises the overall thermal

efficiency of the conventional gas turbine system to 50-52%. The thermal efficiency of









the Kalina cycle alone is sometimes misleading. It is more appropriate to present a

second law or exergy efficiency for low temperature cycles. The Kalina cycle has been

shown to be 16 to 19% more efficient than a comparative Rankine cycle operating with

the same heat source and sink. Payback is estimated at 1.5 years versus 6.5 for Rankine

cycle systems used as bottoming cycles, based on $0.06 per kWH electricity cost (Kalina,

1984).

turbines


Figure 1.6. Schematic diagram of the Kalina cycle (Kalina, 1984).














CHAPTER 2
BACKGROUND AND SUMMARY OF PREVIOUS WORK

This chapter introduces the combined power and cooling cycle as an alternative to

the established methods of low temperature energy conversion. Following an overview of

the cycle concept, previous theoretical work is outlined and key results are summarized

as a motivation for the current experimental work.

Overview of the Cycle

A new energy conversion cycle has been proposed by D. Y. Goswami (1995),

which can be used as a bottoming cycle using waste heat from a conventional power

cycle or an independent cycle using low temperature sources. The motivation is to

improve the effectiveness of harnessing renewable energy resources such as solar and

geothermal energy. With the subsequent reduction in cost from such improvement in

energy conversion, renewable resources can become more competitive to conventional

energy technologies, and the global reliance on fossil fuels may be reduced.

The proposed cycle derives both usable power and cooling from the heat of the

energy source. On the large scale, the combined cycle could be used for industrialized

power production or as a centralized power plant for a community, with refrigeration

produced as required by the application. On the small scale, an affordable residential or

commercial unit could allow independent electricity generation for the home or business

while also cooling it, all with the energy of the sun.









Thermodynamics of the Cycle

Multi-component working fluids in power cycles exhibit variable boiling

temperatures during the boiling process which make them suitable for a sensible heat

source (Ibrahim and Klein, 1996; Kalina, 1984). The temperature difference between the

heat source and the working fluid remains small to allow for a good thermal match

between the source and working fluid, such that less irreversibility results during the heat

addition process.

The proposed cycle combines the Rankine and absorption refrigeration cycles,

using a binary ammonia-water mixture as the working fluid. An ammonia-water mixture

is used as it exhibits desirable thermodynamic properties, such as a large heat capacity.

Ammonia is relatively inexpensive, can accommodate system design modifications well

and separates easily from internal lubricating oils (Norton, 2001; Pillis, 1993). Lower

viscosities of ammonia require smaller piping sizes and are less taxing on pumps.

Ammonia is also environmentally friendly, with no ill effects on global warming or the

ozone layer (Pillis, 1993). Ammonia-water mixtures as the working fluid in power cycles

have shown higher efficiencies than with the conventional Rankine cycle using water or

another single component fluid alone (Bogart, 1981; Thorin et al., 1998). Drawbacks of

using ammonia include it being toxic (see Table A. 1 in Appendix A) and corrosive, with

a mild flammability range of 16 to 27% in air (Pillis, 1993).

A schematic of the cycle is shown in Fig. 2.1. The relatively strong basic solution

of ammonia-water leaves the absorber as saturated liquid at the cycle low pressure. It is

pumped to the system high pressure and is preheated before entering the boiler by

recovering heat from the weak solution returning to the absorber. As the boiler operates

between the bubble and dew point temperatures of the mixture at the system high









pressure, partial boiling produces a high concentration saturated vapor and relatively low

concentration saturated liquid. The liquid weak solution gives up heat in the recovery unit

and throttles into the absorber. The rectifier condenses out water to further purify the

vapor, by rejecting heat to a secondary strong solution stream, before entering the boiler.

The vapor is superheated and expanded through the turbine to produce work. Due to the

low boiling point of ammonia the vapor expands to low temperatures yielding the

potential for refrigeration. The vapor is finally absorbed back into the liquid, giving off

heat that is rejected as the cycle heat output.







211











Ahbmxw 9 ig



Figure 2.1. Schematic of the power and cooling cycle concept.

The main parameters that can be varied to influence the cycle are the heat source

temperature, system high pressure, basic solution ammonia mass fraction, ratio of

working and heating fluid flow rates, and absorber pressure and temperature. Saturation

in the absorber reduces the number of independent main parameters to five that govern









the cycle. Rectifier and superheater temperatures can also be modified, and the conditions

of heat transfer from the source to the ammonia-water mixture as well.

The cycle can be driven by different heat sources including solar, geothermal, and

low temperature waste heat. The use of mid- and low-temperature solar collectors to

drive the combined cycle was investigated by Goswami and Xu (1999), while using

geothermal energy as a heat source was analyzed by Lu (2001) and Tamm et al. (2001).

Typical working conditions of a 400 K boiler superheated to 410 K, and an

ambient at 280 K yield a first law efficiency of 23.5% if work and cooling are added as

the net output. In comparison, the equivalent Carnot efficiency is 31.7%. Conventional

power cycles operating between the same temperatures have lower first law efficiencies

than the proposed cycle, as no cooling output is included. At higher temperatures,

however, their thermal efficiencies are better. The thermal efficiency is deceiving though,

and the strength of this cycle lies rather in the heat source utilization. It exhibits much

higher second law efficiencies than conventional power cycles at the same temperatures.

Comparison to Other Cycles

For utilization of low temperature resources, the proposed cycle offers several

advantages in comparison to other thermal energy conversion methods. In most instances,

low temperature heat from geothermal and solar resources is not an option to provide

direct steam expansion, as the resource temperature is too low. The solar chimney is

highly inefficient unless built to large scales, at which point it becomes impractical. The

Stirling and Rankine cycles become dominant only for mid and high temperature

resources. The Kalina cycle is really the only one that merits a closer look.

The primary advantage of the proposed cycle over the Kalina cycle is the

possibility of refrigeration output. In the proposed cycle, the heat rejection occurs at









much lower temperatures than in the Kalina cycle because the vapor is absorbed into the

bulk fluid while giving off latent heat. Thus the vapor is allowed to expand to much

lower temperatures than in the Kalina cycle, allowing for refrigeration to be a byproduct

of the power production process. The Kalina cycle also operates best for higher heat

source temperatures, such that it is not as suitable for low temperature solar, geothermal

or waste heat resources.

Theoretical Background

Thermodynamic definitions and principles to evaluate the cycle are discussed in

this section, for comparing the proposed cycle to other established cycles that also use

low temperature, sensible heat sources. The theoretical limit in the performance of the

cycle is given in terms of source and sink temperatures, while the actual performance is

gauged from calculated energy transfers.

Properties of Ammonia-Water Mixtures

There are several studies on the evaluation of ammonia-water mixture properties

in the literature. A convenient semi-empirical scheme is used here that combines the

Gibbs free energy method for mixtures and bubble and dew point temperature

correlations for phase equilibrium. The calculated results have been compared to

experimental mixture properties in the literature with good agreement (Xu and Goswami,

1999). The property evaluation method is further described in Appendix B.

Efficiency Based on Source and Sink Temperatures

The Carnot efficiency of a cycle operating between two reservoirs at constant

temperature is given as Eq. 2.1, where TH is the source temperature and TL is the sink

temperature.










S= 1--T (2.1)
TH

The heat source is typically sensible, however, and exhibits a temperature change

during the energy transfer. Assuming that the heat source enthalpy is a linear function of

the temperature, the cycle thermal efficiency can be derived as Eq. 2.2 for a sensible heat

source, using an entropic average temperature.


lT In(TH 'THf ) (2.2)
q =1 iL J(2.2)








H,f 3



H,i -








entropy


Figure 2.2. Temperature-entropy diagram of an ideal cycle with a sensible heat source
and heat sink.

If the sink temperature varies as well, as shown in Fig. 2.2, then with the same

assumptions the cycle thermal efficiency can be expressed as Eq. 2.3 for a sensible heat

source and sink. Variation of the heat sink temperature in addition to the heat source

temperature allows for greater exergy potential.









'IT,- Tf.,) ln(TL,,TLf)
Ll=n1- v H'/' Ll /*1 (2.3)


The thermal efficiency in Eq. 2.3 multiplied by the heat source input gives the

maximum possible power output for a cycle operating between a sensible source and

sink. The bracketed terms represent the minimum part of the source input that must be

rejected as heat.

Efficiency Based on Energy Transfers

Knowing properties at the state points in the cycle of Fig. 2.1 allows for mass and

energy conservation equations to be written over components and for the cycle as a

whole. With minimal assumptions, the work and heat transfers can be determined for the

cycle in steady state.

The cycle first law or thermal efficiency is defined as the net useful energy

product divided by the total energy input, given by Eq. 2.4 or equivalently by Eq. 2.5.

ne- (2.4)
Qh

wt w + q, (2.5)
= -1 : (2.5)
qb +q,

The net work includes both the turbine output and pump input. Qc is the

refrigeration capacity and Qh is the total heat added to the cycle from the heat source in

both the boiler and superheater. These are calculated based on simple mass and energy

balances over the cycle components.

This definition for first law efficiency can be deceiving, as the availability in

refrigeration is less than that in work. The addition of work and cooling output in a single

cycle is not documented in the literature, as such a combined power and refrigeration









cycle is a new idea. Finding and justifying an appropriate definition for first and second

law efficiencies including both work and cooling is a new concept, and is being explored

as motivated by this cycle.

A first comparison can be made by scaling the refrigeration term with the

coefficient of performance (COP) of a refrigeration cycle operating between the

appropriate temperatures. The Carnot COP is used as a limit, although typically the COP

is much lower at about 3. Eq. 2.6 gives a modified first law efficiency if power output is

the primary goal and Eq. 2.7 can be considered as a new cooling efficiency or coefficient

of performance, if refrigeration is the desired output.

Wne +Qc/COPerig (2.6)
rq, = (2.6)
Qh


COPIe = w Or'g +Q (2.7)
Qh

Exergy, or availability, is defined as the maximum reversible work a substance

can do during the process of reaching equilibrium with its environment. The second law

efficiency, or exergy efficiency, for a solar heat source is defined as the exergy output

divided by the exergy input to the cycle (Cengel and Boles, 1998). The exergy input is

taken as the exergy change of the heat source. The exergy output is the exergy of the net

work and the exergy of the refrigeration. The second law efficiency for a solar heat

source is given by Eq. 2.8. It measures how much useful output can be derived from a

change in heat source exergy.


AEW +









The exergy of refrigeration, Ec, is the refrigeration capacity divided by the

coefficient of performance of a Carnot refrigeration cycle operating between the ambient

and cycle low temperatures, as given by Eq. 2.9 (Szargut et al., 1988).


E =Q(T-T) (2.9)


Note that in Eq. 2.8, the definition was based on the exergy change of the heat

source. Solar systems recycle any unused heat source fluid back to the solar collector, so

no exergy is wasted. On the other hand, the unused heat source fluid in geothermal

systems is dumped, therefore a more appropriate second law definition for these systems

is given by Eq. 2.10. This measures how efficiently the maximum amount of available

energy is converted to useful output.

We +Ec
72-- + (2.10)
Ehs

In the definitions of Eqs. 2.9 and 2.10, there is some difficulty in justifying the

addition of work with an equivalent cooling term. There is no standard definition

available for the second law efficiency of a combined power and cooling cycle. Other

definitions can be argued and are being investigated for this cycle, but the above are used

for a first analysis.

Previous Theoretical Studies

The theoretical work that has been performed on this cycle is documented by

Goswami and Xu (1999), Lu (2001) and Tamm et al. (2001), and is summarized here as

background and motivation for the current work. A parametric analysis of the cycle and

optimization of its performance are reviewed, with the inclusion of losses in an









irreversibility study. This summary verifies the feasibility of the cycle concept and

establishes the necessity for an experimental investigation.

Parametric Analysis

Operating conditions were individually varied in a straightforward parametric

analysis to study the effects on the energy transfers and efficiencies of the combined

cycle (Goswami and Xu, 1999). The parametric analysis gave insight into the behavior of

the cycle, and showed that optimization of the cycle would be possible for first or second

law efficiency, as well as work or cooling output. Figure 2.3 is a sample of the parametric

study, showing a peak in the thermal efficiency as defined by Eq. 2.4.

Figure 2.3 shows that for higher ammonia mass fractions in the basic solution,

more vaporization occurs for the given boiler temperature and pressure. This higher

vapor fraction allows for greater flow through the turbine, and more work production for

a higher thermal efficiency. Figure 2.4 concludes that more vapor is available for

refrigeration output also, per kg of basic solution that is boiled. Figures 2.3 and 2.4 were

evaluated for a boiler at 400 K (260 F), superheater at 410 K (278 F), absorber at 280 K

(44 F) and rectifier at 360 K (188 F). The low pressure in the system at the absorber

was set at 2 bar (14.3 psig).

For increasing turbine inlet or system high pressure, the figures show that the

work and cooling outputs peak. Determining the location of these peaks is necessary a

priori to optimize a working system's performance. The effect of the higher pressure in

limiting vapor production begins to dominate as the boiler exit fluid is shifted towards

saturated liquid. The peak shifts for higher basic solution ammonia mass fractions as a

two-phase equilibrium can be sustained at higher pressures for higher mass fractions.
































14i I
16 18 20 22 24 26 28 30 32 34
turbine inlet pressure (bar)


Figure 2.3. Effect of turbine inlet pressure on the thermal efficiency (%) of the cycle
(Goswami and Xu, 1999).


16 18 20 22 24 26 28 30 32 34
turbine inlet pressure (bar)

Figure 2.4. Effect of turbine inlet pressure on the cooling capacity (kJ/kg) of the cycle
(Goswami and Xu, 1999).

Note that a series of similar plots can be determined by changing any of the


operating parameters. Each plot could conceivably provide a visual location of the


optimum over the range of interest for the single parameter. For practical operation, the


--x=0.47
* x=0.5
--x=0.53


-*-x=0 47
--x=0 5
-*-x=0 53









cycle has several parameters that are varied together, presenting a multidimensional

surface on which an optimum can be found. A mathematical approach at locating this

optimum is necessary.

Optimization

Initial parametric studies of the cycle showed the potential for the cycle to be

optimized for first or second law efficiency, as well as work or cooling output. For a solar

heat source, optimization of the second law efficiency is most appropriate, since the spent

heat source fluid is recycled through the solar collectors and the unused exergy is not

wasted. For a geothermal heat source, the second law efficiency allows an interpretation

of the quality of energy usage from the resource fluid. The second law efficiency is a

useful measure of the cycle's performance as it may be used to compare the effectiveness

of different cycles in harnessing the same source.

The optimization work presented here was performed for solar heat sources, and

the second law efficiency is that given by Eq. 2.8. Work has been conducted for

geothermal systems as well (Lu, 2001), with the second law efficiency as in Eq. 2.10.

Optimization methodology. A Generalized Reduced Gradient (GRG) scheme is

used for the optimization (Tamm et al., 2001). The GRG method searches a feasible

region bounded by equality and inequality constraints. Moving finitely towards a better

value with a newly determined search direction at every step, the optimum is ultimately

reached within a limit of convergence.

The optimization scheme searches over eight free variables for the optimal second

law efficiency as defined by Eq. 2.4. The parameters are the absorber or ambient

temperature, boiler, superheater and rectifier temperature, the boiler pressure (high

pressure), absorber pressure (low pressure), and heat source inlet and exit temperatures.











From these eight free variables all other state points in the cycle can be determined with

minimal and reasonable assumptions, neglecting potential and kinetic energies.

Optimization results. The optimization results verified that the cycle could be

optimized using the Generalized Reduced Gradient method. The desired heat source

temperature will vary according to the intended use of the cycle. The effects of heat

source temperature on the optimized cycle performance are shown in Figs. 2.5 to 2.8,

which are optimized for the second law efficiency at the chosen heat source temperature.

Except for the heat source and sink temperatures, which are set, the operating parameters

are determined from the optimal solutions.




80 -

7 ---- cnlQh
70 --Q&Qh

60

50

I 40

30 -




10 -

0 I
300 350 400 450
heat source temperature (K)


Figure 2.5. Efficiencies of the optimized cycle at various heat source temperatures,
optimized for second law efficiency (Tamm et al., 2001).

The refrigeration as a fraction of the heat addition, Q/Qh, changes little as the

heat source temperature increases as shown in Fig. 2.5. As the heat source temperature

approaches the ambient temperature, refrigeration approaches zero. The highest











refrigeration fraction is near a source temperature of 390 K (242 F). The refrigeration


fraction decreases to zero near 480 K (404 F), as the higher temperature vapor can no


longer be expanded to sub-ambient temperatures.




16 -

14

12

S 10

? 8

6" 6

4

2

0 I I I
300 350 400 450
heat source temperature (K)


Figure 2.6. Pressure ratio of the optimized cycle at various heat source temperatures,
optimized for second law efficiency (Tamm et al., 2001).

The net power as a fraction of heat addition, Wnet/Qh, increases as the heat source

temperature increases. As turbine work output is related mainly to the pressure ratio

across the turbine, the net power curve can be explained in relation to the pressure ratio in

Fig. 2.6, which shows a continuous increase with heat source temperature.

The first law efficiency curve, which is a sum of the refrigeration and power

curves, shows similar behavior to the power curve up to the maximum value of 23.6% at


400 K (260 F). After the maximum point the efficiency starts decreasing slowly in a


similar manner to the refrigeration curve.










The second law efficiency shows a maximum value of 65.2% at 380 K (224 F).

The sharp increase in Fig. 2.5 of the second law efficiency between 320 and 380 K (116

to 224 F) is due to the increase of both power and refrigeration outputs. The second law

efficiency reaches a maximum where the refrigeration output begins to decrease above

400 K (260 F).

Figure 2.7 shows the refrigeration to net power ratio versus heat source

temperature. In the temperature range between 320 and 360 K (116 to 188 F) this ratio

changes rapidly, while above 360 K (188 F) the ratio decreases slowly, reaching 0.12 at

460 K (368 F). Thus, increasing the heat source temperature favors production of power

rather than refrigeration.







08


06


04


02



300 350 400 450
heat source temperature (K)


Figure 2.7. Ratio of refrigeration to work of the optimized cycle at various heat source
temperatures, optimized for second law efficiency (Tamm et al., 2001).

Figure 2.8 shows normalized exergy destruction in the cycle as a function of the

heat source temperature. The total exergy destruction in the cycle increases with an










increase in the heat source temperature, such that more of the heat source availability is

wasted. It can be seen in Fig. 2.8 that the exergy destruction in both the absorber and heat

exchanger changes little as the source temperature increases. The superheater has almost

no exergy destruction because of its small heat load. The boiler exergy destruction is

much lower than that of the absorber.



02



0 15 -
t


01



0 05




300 350 400 450
heat source temperature (K)


Figure 2.8. Exergy destruction of the optimized cycle at various heat source
temperatures, optimized for second law efficiency (Tamm et al., 2001).

From the exergy analysis, if the heat source is between 320 and 460 K (116 to 368

F), then the best operating heat source temperature is around 380 K (224 F), since it

gives the maximum second law efficiency.

Exergy destruction in the rectifier increases throughout in Fig. 2.8, as the heating

load also increases in the rectifier. The sharp increase of rectifier exergy destruction is

due to more rectification needed for higher heat source temperatures in order to obtain a

high purity vapor stream and thus refrigeration. At even higher temperatures, it becomes









too costly to produce refrigeration from a second law efficiency standpoint, such that for

optimized cases there is less refrigeration according to Fig. 2.5. Figure 2.8 shows that

finally there is less rectification and thus fewer losses in the rectifier at high heat source

temperatures.

It has been shown that the cycle can be optimized for a range of heat source

temperatures. Similarly, the cycle can be optimized for each heat sink or ambient

temperature, and other parameters. This way the cycle can be customized to the intended

application for optimal performance.

Irreversibility Analysis

In realistic systems, there are irreversibilities associated with every component as

with this ammonia-water cycle. These irreversibilities will have negative effects on the

performance of the cycle. The effects of each loss were studied individually and jointly

on the cycle performance (Tamm et al., 2001). Typical working conditions used in this

analysis were 400 K (260 F) and 30 bar (421 psig) at the boiler exit, 360 K (188 F)

rectification, 410 K (278 F) at the turbine inlet, 280 K (44 F) and 2 bar (14.3 psig) in

the absorber, and a basic solution ammonia mass fraction of 0.53.

A typical turbine efficiency of 90% was assumed as suggested in the literature

(Drbal et al., 1996). The thermal efficiency drops from 23.3% to 19.7%, a decrease of

15.4%. Due to the irreversibility in the turbine, although the pressure ratio is the same,

the exhaust temperature of the turbine is higher. Less energy is converted into mechanical

work in the turbine, and the turbine work output drops from 76.1 kW to 68.5 kW, a

decrease of 10.0%. At the same time, a higher turbine exhaust temperature provides less

cooling capacity. The cooling capacity drops 29.2% from 26.0 kW to 18.4 kW.









A pump efficiency of 80% was assumed as suggested in the literature (Drbal et

al., 1996). The pump work requirement increases from 3.4 kW to 4.2 kW. This small

increase causes the thermal efficiency to drop slightly.

A pressure loss of 5% of the inlet pressure was assumed across the boiler as

suggested in the literature (Bhatt et al., 1994). The results show this pressure loss has

almost no negative effect on the cycle performance. Only slightly more pump work is

required to boost the boiler inlet pressure to compensate for the pressure loss in the

boiler.

A pressure loss of 5% was assumed for the superheater (Bhatt et al., 1994). The

results show only a minor negative effect on the cycle performance. Thermal efficiency

decreases by 3% of that in the ideal cycle, from 23.3% to 22.7%. Due to the pressure loss

in the superheater, the turbine inlet pressure drops. Therefore, less expansion is possible

producing 1.2% less work and higher exhaust temperatures. The cooling capacity

decreases by 6.5%.

A pressure loss of 5% was assumed for both streams in the recovery heat

exchanger (Bhatt et al., 1994). The effects on the cycle performance are minimal, with a

negligible decrease in thermal efficiency owing to an increase in the pump work

requirement.

A pressure loss of 5% was assumed in the refrigeration heat exchanger, for

comparison to other component pressure losses. The thermal efficiency drops by 2.7%

from 23.3% to 22.8%, as the higher turbine exhaust pressure limits the expansion

possible. The work output decreases by 1.6%. The reduced turbine pressure ratio also

raises the exhaust temperature, reducing the cooling capacity by 4.6%. In a typical cooler,









however, the heat exchanger experiences a 3% pressure loss (Bhatt et al., 1994), which is

the value used in the combined irreversibility study.

Finally, the overall effect of the irreversibility associated with the cycle was

analyzed for combined losses. The thermal efficiency decreases from 23.3% under ideal

conditions to 18.5%. The turbine work output drops by 11.8%, from 76.1 kW to 67.1 kW.

The cooling capacity decreases by 37.7%, from 26.0 kW to 16.2 kW. It can be seen that

the greatest loss is attributed to the imperfect expansion in the turbine.

Summary of Theoretical Studies

The parametric analysis showed the potential for the cycle to be optimized.

Optimization of the operating parameters is possible for each heat source and heat sink

temperature, using a Generalized Reduced Gradient (GRG) method. The cycle may be

optimized for the first law efficiency, second law efficiency, power output or cooling

output, depending on the intended application and the heat source. For a solar heat

source, optimization for the second law efficiency is most appropriate, since the spent

heat source fluid is recycled back to the solar collectors. It is found from simulation that

optimization for second law efficiency produces no refrigeration at high heat source

temperatures, while for low heat source temperatures it does. Inclusion of realistic losses

in the analysis reduces the cycle thermal efficiency by 20.6%, with 11.8% less work

output and 37.7% less cooling capacity. The largest source of irreversibility in the cycle

is the imperfect turbine expansion.












CHAPTER 3
EXPERIMENTAL SYSTEM
An experimental system has been built to demonstrate the feasibility of the

ammonia-based combined power and cooling cycle. Operation of a working system gives

practical experience and a means to improve the basic cycle design, in order to advance

the concept towards industry.










HE
I













Figure 3.1. Schematic of the experimental system concept. In the current study, the
shown turbine and refrigeration processes are simulated.
The conceptual design of the experimental system is shown in Figure 3.1. The

experimental study is being approached in two phases. In Phase 1 the vapor generation

and absorption processes are studied, while a heat exchanger and throttling valve model









the expansion process until a turbine is installed in Phase 2 to form a complete power

cycle. In comparison to the initial cycle concept of Fig. 2.1, the rectifier and consequently

the superheater have been left out for simplicity as purification of the high concentration

ammonia-water vapor is not critical without a turbine in place. The refrigeration unit is

also simulated by the turbine heat exchanger. Phase 1 is presented in this dissertation, and

is used in preparation for later Phase 2 studies.

Over the course of the experimental work, an initial system was built. An

experimental study revealed where modifications could be made, yielding an improved

system that was also studied. Changes made in the system are noted below, and the

results are compared in Chapter 5.

Ammonia-Water Side

This main part of the system is shown in detail by the schematic in Fig. 3.2, with

photographs in Figs. 3.3 and 3.4. A list of components is given in Appendix C, and

experimental operating procedures are given in Appendix D.

In the initial system, the strong solution of ammonia and water was pumped from

the absorber with a rotary vane positive displacement pump capable of producing the

high boiler pressures of interest. A diaphragm pump, shown in Fig. 3.5, is used in the

improved system as the rotary vane pump was found to cavitate and failed after

prolonged operation. Also, the diaphragm pump flow rate is not limited by the

downstream pressure, in the range of operating pressures of interest. A strainer is at the

pump inlet to prevent loose rust pieces from damaging the pump. The collected particles

can be purged from the system through a valve. An expansion tank is positioned after the

pump to dampen the pulses in the flow rate. Leaving the pump, the strong solution passes

through a nickel-brazed, stainless steel, vertically stacked plate heat exchanger where it









recovers heat from the weak solution returning to the absorber. The heat exchange area in

the recovery unit was increased by 286% for the improved system. The system heat

exchangers are shown in Fig. 3.6.


from turbine to turbine


Figure 3.2. Schematic of the experimental system components for the system side.











re d


Figure 3.3. Photograph of the system side and hot water side of the experiment.


Figure 3.4. Photograph of the system side of the experiment.




















Figure 3.5. Photograph of the system side pump and expansion tank.


-r L r
I^ l l1


Figure 3.6. Photograph of the system side heat exchangers.









The boiler is also a vertically stacked plate heat exchanger, using resistance

heated hot water to partially boil the strong solution which flows into a simple carbon

steel tank where gravity separates out liquid from vapor. The heat exchange area in the

boiler was also increased for the improved system, by 47%.

The relatively low concentration weak solution exits the bottom while high

concentration vapor leaves through the top of the separator. The weak solution flows

through the solution heat exchanger to yield energy to the strong solution. After a

throttling valve to reduce pressure, the weak solution enters at the top of the absorber

through a spray nozzle across the cooling elements inside, as shown in Fig. 3.7. The

vapor is throttled and passes through a stacked plate heat exchanger against coolant flow.

This expansion lowers the pressure and temperature of the vapor before it enters the

carbon steel absorber through 18 holes of 0.041" diameter distributed along a horizontal

S-shaped tube section, intended to bubble the vapor into the pool for better absorption.

A storage tank is in place to contain the working fluid while the system is

undergoing modifications or repairs. The storage tank is also used when the system is

initially charged, and can be used during distillation of the working fluid to alter its

concentration. A flow indicator, shown in Fig. 3.8, connects the storage tank with the

system in order to monitor whether liquid or vapor is flowing.

Stainless steel tubing connects the components, with Swagelok pressure fittings

and valves where needed. Care was taken to minimize flow losses in the system. All the

tubing is insulated, as are the absorber and separator lanks. Additional insulation was

added to flow meters and fittings in the improved design. Preventing heat losses in vapor

lines discourages condensation on the tubing walls.













weak solution
coolant inlet
























vapor


strong solution


vent! -- to emergency tank


Sprayyr







finned heat
exchangers











liquid level
indicator
bubble diffuser
___ _' 4 + 4 + t -


:~~* (&~


Figure 3.7. Schematic of the absorber design.

Material compatibility is a serious concern with ammonia solutions. All

components are selected to have no reaction with the working fluid. The tubing and

fittings are made of stainless steel and the tanks are of mild steel. The absorber

evaporator units are made of aluminum. Other heat exchangers are with nickel-brazed

stainless steel. O-rings in the valves are made of Kalrez, and are coated with Krytox

lubricant. Gaskets, packing and septa elsewhere are suitably of neoprene, Teflon,

EPDM or other compatible material.


olant exit


I































Figure 3.8. Photograph of the flow indicator to the storage tank.


Figure 3.9. Photograph of the hot water side of the experiment.









Hot Water Side

The hot water side is shown in Fig. 3.9, with a schematic in Fig. 3.10. A 5 kW

single-pass boiler supplies hot water to a hot water storage tank, to build up ample high

temperature storage. This way the system can provide more than 5 kW heat input, for a

finite period of operation. The water is pressurized for the storage temperature to exceed

100 C (212 F), and can go as high as 150 C (302 F). A circulating pump sends the hot

water to the plate heat exchanger used as a boiler of the ammonia-water mixture, then

back to the single pass boiler.

Lines are set up which can divert the hot water from the plate heat exchanger to

an advanced boiling apparatus of the ammonia-water mixture, still in the design stage,

that will allow a more detailed study of the boiling process in the cycle in future work.

Hot water from this boiling apparatus will also be returned to the hot water storage tank.

The hot water temperature is controlled with a mixing valve, which mixes hot

water drawn from the storage tank with some of the relatively cooler water returning

from the system boiler. The valve regulates the flow from each of these lines to maintain

a steady outlet temperature. The single pass boiler, storage tank, and associated tubing are

well insulated.

Coolant Side

The coolant side is shown by the schematic in Figure 3.11, with a photo shown in

Fig. 3.12. The thermostatically controlled chiller maintains a constant coolant

temperature, within + 3 C (+ 5 F), in the chilled water storage tank. It is capable of

producing coolant temperatures of -30 C (-22 F). The coolant used is a 50/50 mixture

of ethylene-glycol and water, which has a freezing point that is low enough for the









application and liquid viscosity at such low temperatures that can be handled by

conventional circulation pumps. The chiller serves to provide a range of coolant

temperatures for simulating ambient sink conditions for the absorption process.


Figure 3.10. Schematic of the experimental system components for the hot water side.









pump


Figure 3.11. Schematic of the experimental system components for the coolant side.

From the chilled water storage tank, the coolant is pumped to the absorber and

passes through a finned coil heat exchanger in the absorber. The finned coil heat

exchanger consists of several aluminum car evaporators connected in parallel to

minimize coolant flow losses. The fins on the evaporators enhance the heat transfer

between the cooling liquid and ammonia solution. They also act as baffles to increase the

residence time of the weak solution before it gets to the bottom of the tank so that the









ammonia vapor and weak solution have enough time to be mixed. A second coolant line

provides the cooling for the heat exchanger in the ammonia vapor line to simulate the

turbine. The storage tank and piping on the coolant side are well insulated.



























Figure 3.12. Photograph of the coolant side of the experiment.

Instrumentation

Pressures (P), temperatures (T), ammonia mass fractions (x) and flow rates (m)

are recorded at locations shown in Figure 3.13. Properties at other state points can be

extrapolated with minimal assumptions for energy transfers to be calculated. Changes in

potential and kinetic energies are neglected. Some of the measurements are redundant

and are used to compare calculated results from two sets of independent measurements.

Properties of the hot water and coolant can be used to determine energy transfers and to

verify those found from the ammonia-water side.











hot water






separator -



recovery
HE




simulated
pumrp (turbine

coolant




absorber --------
I refrigeration





coolant

Figure 3.13. Location of data measurements for system analysis.

Thermocouples

In total, there are 19 temperatures monitored by T-type thermocouples of which

14 are connected to the DaqBook data acquisition system to record real-time and time-

averaged values with the DaqViewTM software. The 14 are located as shown in Fig 3.13.

Note that in the initial system, the liquid and vapor exit temperatures from the separator

were not measured, such that only 12 thermocouple values were then recorded. Multiple









channel temperature displays are near the experiment to allow quick readouts. A two-

point calibration was used for each of the thermocouples, as outlined in Appendix D. The

error in the measurement is discussed in Appendix E.

















Figure 3.14. Photographs of pressure measurement instruments on the system.

Pressure Transducers

Pressure transducers at 5 locations record data directly through the data

acquisition system and compliment the analog gauge displays on the system. Transducer

locations are noted in Fig 3.13. The transducers were calibrated using a one-point method

as outlined in Appendix D, with the uncertainty discussed in Appendix E. Gauge readings

were used in place of certain transducers while they were being repaired, as listed

according to the measurement device history in Table E.4 of Appendix E.

Gas Chromatograph and Syringe Sampling

Samples are taken by syringe through septum ports in the strong, weak and vapor

lines. A sampling port on the absorber is shown in Fig. 3.15. The samples are transported

to a gas chromatograph (GC) in the lab by a locked syringe, and then analyzed using a

thermal conductivity detector (TCD) in the GC. The GC setup is shown in Fig. 3.16.







50


Procedures for syringe sampling, using the GC, and calibration are given in Appendix D.


Uncertainty discussion is given in Appendix E.




































Figure 3.15. Photograph of a syringe sampling port and sight glass on the absorber.


i ~
a 41Gm
SI
S
U


ji


1. OF .J


Figure 3.16. Photograph of the gas chromatograph setup.









Flow Meters

Variable-area float-type flow meters are in place to measure strong and weak

solution, vapor, hot water and coolant flow rates. A secondary spring/piston-type flow

meter is connected to the vapor coolant for trials with higher flow rates. The liquid flow

meter readings were verified but not calibrated. Corrections for the measured flow rates

are given in Appendix D. The uncertainty is discussed in Appendix E.



















Figure 3.17. Photograph of the temperature displays and coolant flow meter.

Data Acquisition Hardware and Analysis Software

DaqBook was used as the data acquisition electronic interface for the pressure

and temperature measuring devices. Expansion cards are connected directly to the

transducer and thermocouple wires from the system. More about the DaqBook

components of the data acquisition system is presented in Appendix D.

The readings provided by the DaqBook system are shown in real-time and sent

to file by the DaqViewTM software. To process the raw binary or text files provided by the

acquisition system, a data analysis program was written to extract and compile the









multiple readings into more useful time averaged data. The key elements of the computer

code are given in Appendix F. The program applies assumptions as needed to yield the

pressure and temperature at each state point. With manual entry of concentration

measurements from the GC and flow rates from the flow meters, the program accesses a

routine similar to the one used in the theoretical studies to determine properties at each

state point, and furthermore provides key results such as cycle efficiencies, work output

and cooling capacity. The experimental results are then compared to a simulated cycle

with the program.

The program also allows for the design of experiments, extrapolating desired

system operating conditions consisting of thermocouple, transducer and flow meter

readings from a theoretical starting point consisting of heat source temperature, ambient

sink temperature, boiler high pressure, absorber low pressure and a ratio of the heat

source and ammonia-water flow rates. Component parameters such as heat exchanger

effectiveness, pump and turbine efficiencies, and other losses can be included and

modified according to the real results of the experiment, such that the design program

evolves with the experimentation. The goal of the design program is to predict actual

system parameters efficiently to minimize the experimentation time whenever new design

conditions are to be tested.

Safety

The safety measures needed for system operation are extensive as ammonia is

toxic and corrosive. The components have been mounted on the wall outside the lab

building, allowing adequate ventilation in case of leaks. Pressure relief valves are

included to prevent critical pressure buildup in the system, which might otherwise rupture

a component. Upon cracking, the relief valves discharge through tubing into a large









emergency storage tank, which can be safely emptied at a later time. The emergency tank

is shown in Fig. 3.18. However if a rupture does occur, overhead water sprinklers can

immediately dilute the surrounding air to safe concentrations of ammonia.





























Figure 3.18. Photograph of the storage tank (left) and emergency tank.

An emergency power cutoff switch can cut power to both the system side and hot

water pumps. A gas mask fitted with ammonia filters is available for emergency use. An

eyewash station and shower are also located adjacent to the experiment. For normal

system operation, however, only chemical spill resistant goggles are necessary. Lab coats

and face shields are recommended for certain maintenance operations. Emergency

procedures are outline in Appendix D.














CHAPTER 4
EXPERIMENTAL METHODOLOGY

Strategies for the design of experiments, system operation and results analysis

methods are discussed in this chapter to best match the limits and behavior of the system

and to yield the results of interest.

Parameters in Simulations and Experiments

In simulating the modified cycle that the experimental system is based on, the

design parameters can be classified as either fluid or component parameters. The fluid

parameters can sufficiently describe the cycle for the special case of ideal components.

For realistic simulation, the complexity of the design increases as the components are

given non-ideal characteristics.

Fig. 4.1 shows the fluid parameters used in the simulation of the cycle. The mass

flow rates of the working fluid and hot water are extensive properties and can be

combined into an intensive heat source flow ratio parameter, thus leaving 5 fluid

parameters to simulate the ideal cycle. For the inclusion of realistic losses, 12 component

parameters are used in the simulation as shown in Table 4.1. The parameters include

pump and turbine efficiency, pressure losses, effectiveness for heat transfers between two

fluid streams for which the heat exchange process is specific, and approach temperature

limits for heat transfers with the ambient for which the heat exchange process is not yet

specific. Typical values for the losses are also given in Table 4.1.

For the experimental system trials, data is taken only at critical locations and

certain component parameters are estimated to complete the set of properties at each









location. Minimal error is expected from these assumptions. The component parameters

used in conjunction with the experimental data are shown in Table 4.1. Several pressure

losses are estimated as pressures are only measured at limited locations in the system.

The pump efficiency is estimated since the immediate pump exit pressure is not recorded

in each trial. The other component parameters of Table 4.1 that are assigned in the

experimental trials are necessary to model the turbine and ambient conditions.

heat source


I.
/
V


ambient

Figure 4.1. Schematic of the cycle concept used in experiment simulations, showing the
main fluid parameters.









Table 4.1. Fluid and component parameters in the simulation and experiment.

Simulation Experiment
Fluid Parameters

heat source temperature Set Measured
boiler pressure Set Measured
ambient temperature Set Set (absorber T is
measured)
basic solution ammonia mass fraction Set Measured
heat source flow ratio Set Measured

Component Parameters ideal typical

approach T ambient and absorber 0 5 V Set Set
approach T ambient and refrig. HE 0 5 Vi Set Set

effectiveness recovery heat exchanger 100 85 % Set Measured
effectiveness boiler heat exchanger 100 85 % Set Measured

isentropic efficiency turbine 100 90 % Set Set
isentropic efficiency pump 100 80 % Set Set

pressure loss vapor inlet to absorber 0 5 % Set Measured
pressure loss weak inlet to absorber 0 5 % Set Measured
pressure loss recovery HE strong side 0 5 % Set Set
pressure loss recovery HE weak side 0 5 % Set Set
pressure loss boiler HE system side 0 5 % Set Set
pressure loss refrig HE system side 0 5 % Set Set


Limits and Selection of Operating Conditions

The operating conditions of the investigations were discretely selected to

demonstrate the vapor generation and absorption processes, and to determine the power

production and refrigeration potential of the system in comparison to simulations.

Contour plots such as Figures 4.2 and 4.3 show this potential for a typical cycle design at

two different ambient temperatures, and are used to determine the necessary operating

conditions. These simulations, shown for the ideal cycle with no losses here, allow the

anticipation of certain results and system behavior prior to experimentation.











ambient temperature = 5 C
basic solution mass fraction = 40 % heat source flow ratio = 3.5







cycle with no losses






yc no / ii//// I
SI I/ I/ / Y I] |
'**- ^"^ ^^ *T -7 7

.^^ y././ ^^A,


70 80 90
heat source temperature (C)


ambient temperature = 25 C
basic solution mass fraction = 40 % heat source flow ratio = 3.5







cycle with no losses







/ll/ /
'- -_ / / /i \_


70 80 90
heat source temperature (C)


Figure 4.2. Expected turbine work output (kJ/kg) per unit basic solution flow, based on
simulation with no losses.


14

13

12

. 11


U)
o 9

o 8
'0
-Q 7

6

5


14

13

12
. 11


U)
U)
o 9

o 8
-Q 7

6

5


100


110


100


110






58


ambient temperature = 5 C
basic solution mass fraction = 40 % heat source flow ratio = 3.5


60 70 80 90 100
heat source temperature (C)

ambient temperature = 25 C
basic solution mass fraction = 40 % heat source flow ratio
I I I I ///


110


= 3.5
AL/


/7/


:_________ I ____________ //^/___________
- yl wt.o los e/ //s
cycle with no losse.//




I ___________^ I I I I I I I I I I I I I I I I


70 80 90
heat source temperature (C)


100


110


Figure 4.3. Expected refrigeration output (kJ/kg) per unit basic solution flow, based on
simulation with no losses.









The plots above are 3-dimensional, but are viewed in two dimensions. In other

words, each contour represents the locus of all work outputs of the same magnitude in

Fig. 4.2, and cooling capacities of the same magnitude in Fig. 4.3.

Prior to experimentation, the limits of the system were evaluated, to which the

desired operating conditions must conform. The feasible range of operating conditions is

dictated by experimental and practical limits, and is more stringent than in simulations.

The limits of operating conditions are given in Table 4.2. Limits of system components

are listed in Appendix C, which cannot be exceeded by the operating conditions.


Table 4.2. Limits of primary operating conditions.

Lower limit Upper limit
heat source temperature 50 C (122 F) 110 C (230 F)
boiler pressure 3.8 bar (40 psig) 14.8 bar (200 psig)
ambient temperature 5 C (41 F) 35 C (95 F)
basic solution NH3 mass fraction 30% 50%
heat source flow ratio 0.5 8

Heat Source Temperature

The heat source temperature was varied over a sufficiently large range to

demonstrate trends in system performance and to compare them to simulations. Since the

application can be used for geothermal sources, a practical upper temperature limit is

typically 140 C (284 F) for the boiler. However the experimental limit is taken as 110

C (230 F), as for low temperature solar sources, so not to tax the hot water storage and

boiler beyond capacity. For low heat source temperatures, vaporization is possible at low

pressures. However, low pressures limit the expansion capability of the system so the

heat source temperature lower limit was set at 50 C (122 F).









The heat source temperature must be high enough to partially boil the strong

solution at the given pressure and ammonia mass fraction, in order to generate vapor for

power and refrigeration production. However if too high, there will be no liquid leaving

the separator and the cycle will resemble a binary fluid Rankine cycle. Limits of vapor

generation for a typical solution concentration and absorber temperature are shown in

Fig. 4.4, which gives the vapor fraction of the mixture leaving the boiler for various

temperatures and pressures. The vapor fraction, or quality, is defined here as the ratio of

the mass flow rate of vapor leaving the separator to the mass flow rate of the basic

solution entering the boiler.


ambient temperature = 25 C
basic solution mass fraction = 40 % heat source flow ratio = 3.5

-E------7/


12
.a 11
5 10
U)
S 9
C.
_ 8
0
- 7
6
5
4


// //cycle with no losses

:--- ccl wthnolose ---- //~^-/7- / -'//.--s


50 60 70 80 90 100 110
heat source temperature (C)

Figure 4.4. Vapor fraction (%) of ammonia-water fluid leaving the boiler, based on
simulation with no losses.

The corresponding specific heat input is shown in Fig. 4.5. Knowing the

experimental limitations for heat input, Fig. 4.5 shows whether the design is feasible a

priori. With this system, 5 kW can be continuously supplied by the hot water heater. Any










higher heat input is finite and depends on the heat stored in the hot water storage tank.

Similarly, Fig. 4.6 is used to remain within absorber heat rejection limits set by the

coolant serving as the heat sink.


ambient temperature = 25 C
basic solution mass fraction = 40 % heat source flow ratio = 3.5

I I I A /


14

13

12

. 11
01
*5 10
U)
o 9
a.
o 8

.0 7

6

5

4


- cycle with no losses


50 60 70 80 90 100 110
heat source temperature (C)


Figure 4.5. Boiler heat input (kJ/kg) per unit basic solution flow, based on simulation
with no losses.

Boiler Pressure

The system performance is sensitive to the boiler pressure, and so it was varied

over a large range to show diverse results. The boiler pressure must be such as to allow

partial boiling at the given boiler temperature and basic solution ammonia mass fraction.

The upper limit is set at the separator pressure rating of 14.8 bar (200 psig). The lower

limit is arbitrary, provided it is higher than the saturation pressure of the absorber. At low

boiler pressures, not much expansion would be possible through the turbine and







62


refrigeration would not be seen. The lower limit is thus taken as 3.8 bar (40 psig), but was

higher in most studies.


ambient temperature = 25 C
basic solution mass fraction = 40 % heat source flow ratio = 3.5

14 / __


13

12
n 11

5 10
U)
U)
S 9
0.
_ 8
0
Q 7

6

5

4


: / / /




*- X / ./ ,/

_- / ./ / ?


/ _____ / I^ I S I / .


50 60 70 80 90 100 110
heat source temperature (C)


Figure 4.6. Expected absorber heat rejection (kJ/kg) per unit basic solution flow, based
on simulation with no losses.

Ambient Temperature

The ambient temperature is a key design parameter, although it was simulated

through the measured absorber temperature and the estimated approach limit. The

absorber temperature affects the absorber pressure as the absorber operates near

saturation, and thus the expansion ratio that is available for the turbine is affected. The

boiling process is not expected to depend highly on absorber temperature and so only a

few ambient temperatures were studied. The limits of absorber temperatures reflect the

ambient temperature limits, which were taken for a range of climates and seasons.









Basic Solution Ammonia Mass Fraction

The basic solution ammonia mass fraction may fully range between 0 and 100%

ammonia by mass, provided the mixture can be partially evaporated at the given boiler

pressure and temperature. The strong solution should leave the absorber as liquid, and the

saturation pressure must remain below the upper limit of the absorber for a given ambient

temperature. For simulating the ambient temperatures of interest, a 30-50% ammonia

mass fraction was selected as it gives a range of absorber saturation pressures most

compatible with the current experimental system. It is desired to remain lower than the

3.1 bar (30 psig) absorber pressure rating and above 1 bar (0 psig) to maintain positive

pressure as a rule of thumb. Usually the lower limit is higher, according to the saturation

pressure of the absorber.

Heat Source Flow Ratio

The limits on the heat source flow ratio arise mostly from limits in pumping

capacity and flow meter useful range. The hot water pump produces less than 5.7 1pm

(1.5 gpm), and the system pump gets around 1.9 1pm (0.5 gpm) depending on the

absorber pressure. A typical heat source flow ratio is around 3.5, and the system was

operated near this value usually. The heat source flow ratio was not varied deliberately.

System Behavior

The system reaches a pseudo-steady state typically in 20-30 minutes between

trials, and longer when first started or if the experiment borders on component limits such

as pump flow rate. As pressures, temperatures and flow rates are closely tied, a slow

fluctuation over several minutes of one property causes a predictable, slow fluctuation in

other properties. This fluctuation is generally self-correcting after the system has been

running for a while, but can be sped up by operator intervention. This pseudo-steady









fluctuation is taken as acceptable, but is an inconvenience and should be reduced with the

implementation of control methods.

Steady state during operation is determined by monitoring the above system

parameters. The indirect source of deviation from steadiness arises from a poor absorber

design, in which the cooling occurs above the liquid pool. The absorber pressure is

sensitive to the coolant temperature and flow rate, dropping the absorber to very near its

saturation pressure. The absorber should operate slightly above its saturation pressure, so

that the pump does not cavitate and is able to draw a sufficient flow rate.

Helium has been added to the absorber during operation to raise the absolute

pressure about 10% above saturation, thus compressing the liquid leaving the absorber.

The pump flow rate improves significantly if pumping liquid that is compressed, and

does not work at all if the liquid is near saturation. At times when the pump cavitates to

the point of producing no flow rate, either the pressure in the absorber can slowly rise by

regulating the coolant flow or the absorber can be shocked with excess vapor from the

separator. A revision in the absorber design to have the cooling elements in the liquid

pool would slightly subcool the fluid steadily, with the same effect to improve the pump

flow rate. Elevating the absorber would also generate head pressure to compress the

liquid at the pump inlet.

Another strategy to maintain the absorber at slightly higher than saturation

pressure is to permit slow pooling of the liquid solution in the separator. The reduced

weak solution return to the absorber slows absorption and raises the absorber pressure.

Liquid level indicators on the absorber verify the imbalance in flow rates. After an hour

or two, the weak solution can be returned before the absorber liquid level drops below the









vapor inlet. If the level is below the vapor inlet, the absorber pressure becomes very

sensitive to weak solution flow rate and may rise quickly to above vessel limits.

Therefore, the pump flow rate is the direct source of deviation from steadiness. A

change in the pump flow rate will change the temperature of the strong solution exiting

the boiler, and how much vapor has been formed. The effects compound and the system

is shifted from the desired operating conditions.

Usually control efforts are localized to the absorber to maintain steady operation.

A general control is to vary the coolant flow rate according to the coolant inlet

temperature. For changing operating parameters and for other areas of control, Table 4.3

lists the corresponding control measures. These should be used in conjunction with the

operating procedures of Appendix D.


Table 4.3. Experimental control measures for maintaining operating parameters.

System parameter... Controlled primarily by...
boiler pressure vapor flow through throttling valve and basic
solution flow rate
heat source temperature storage tank temperature and mixing valve setting
absorber temperature coolant temperature and flow rate to absorber
absorber pressure basic solution ammonia mass fraction, coolant
temperature and flow rate, vapor and weak solution
flow rates
basic solution mass fraction system charging before operation
basic solution flow rate absorber pressure, diverting flow from basic
solution line back to absorber
vapor inlet temperature to absorber coolant temperature and flow rate to vapor H.E.
fluid pooling in separator weak solution flow rate leaving separator

Uncertainty of Measurements

Data acquisition in the ammonia-water side during operation is limited to the

pressures, temperatures, ammonia mass fractions and flow rates taken at locations in the

system shown in Fig. 3.12. Properties at all other state points can be extrapolated with









minimal and reasonable assumptions, and estimation of component parameters in Table

4.1. The losses are expected to reduce the performance of the simulated cycle. Fig. 4.7

shows the first law efficiency for a cycle simulated as ideal and with the typical losses of

Table 4.1. Fig. 4.8 shows a similar plot for the second law efficiency. Note that the

assumption of an ideal heat exchanger causes instabilities in the computations near the

lower limit of vapor generation, and rough contours arise in the simulation of the cycle

with no losses.

Measurement errors are compounded in calculated energy transfers, efficiencies

and other key results according to the uncertainty analysis discussion in Appendix E. In

addition to the error attributed to measurements, the phase of the ammonia-water mixture

is not well known at each location, which contributes to error in the property evaluation.

As property evaluation of liquid is of greater accuracy than that of two-phase ammonia-

water mixtures, data is acquired redundantly in locations where the phase is more certain.

The required measurements to obtain key results from the ammonia-water side of the

experiment are reduced, as measurement from the hot water side is also used. The coolant

side data is not used, as the response time of the system to coolant flow rate is very slow

and thus the coolant flow rate may be in error. The coolant flow meters also exhibit a

worse rated accuracy than the hot water and ammonia-water flow meters.

The uncertainty of the key results should not be greater than 5%. In order to

minimize the experimental error, the instrumentation should be calibrated carefully,

repeated measurements should be taken, and the measurements should be taken only

when the system is deemed sufficiently steady.











ambient temperature = 25 C
basic solution mass fraction = 40 % heat source flow ratio = 3.5







cycle with no losses


70 80 90
heat source temperature (C)


100


ambient temperature = 25 C
basic solution mass fraction = 40 % heat source flow ratio = 3.5







cycle with typical losses I
------ ------ ------ -- .y .' *' / -- /------ l

I______z Zv_/;


70 80 90
heat source temperature (C)


100


110


110


Figure 4.7. Expected first law efficiency (%), based on simulation with no losses and
with typical losses.







68



ambient temperature = 25 C
basic solution mass fraction = 40 % heat source flow ratio = 3.5







cycle with no losses


amben-tmpratre--5-




Al --- y^-


ambient temperature = 25 C
basic solution mass fraction = 40 % heat source flow ratio = 3.5







cycle with typical losses









i____^' i iiiii


70 80 90
heat source temperature (C)


100


110


110


Figure 4.8. Expected second law efficiency (%), based on simulation with no losses and
with typical losses.


70 80 90
heat source temperature (C)


100














CHAPTER 5
EXPERIMENTAL RESULTS

The results show that essential processes of the proposed cycle can be reproduced

experimentally. Generation of vapor at high pressures and its expansion back to low

pressures verify the potential to produce work from heat. Modeling the turbine and

refrigeration unit with an equivalent expansion and temperature change, experiments

support the potential for refrigeration at certain operating conditions. Absorption of the

vapor back into the basic solution confirms that the cycle can regenerate itself.

As the turbine is simulated by an equivalent expansion, the calculated results that

follow are grouped into those that are independent of the expansion process and those

that depend on the models used for the turbine and refrigeration unit. To best convey the

key findings of each section, a qualitative interpretation of the results and explanation of

relevant phenomena precedes the presentation of the results. Generally the expected

trends of simulating the cycle are evident in the experimental results, although

quantitative comparisons to simulated results show inconsistencies.

Vapor Generation

Experiments suggest that there are non-equilibrium phase change processes

during the generation of vapor from the basic solution. Many of the results can be

attributed to a difference between the wall temperature and the fluid bulk temperature

during conditions at or near saturation. The phase change at the wall will not be in

proportion to the phase change in the bulk fluid, which is assumed to exist in equilibrium








at the given temperature, pressure and ammonia mass fraction. The conditions near the

wall are very sensitive to the wall temperature.

Observations

The vaporization of the strong solution may initiate in the recovery unit and

continues in the system boiler. In the boiler and recovery heat exchanger, the wall

temperature is higher than the bulk temperature of the strong solution. This causes vapor

production of lower ammonia mass fractions as higher temperature boiling can vaporize

proportionally more water, which has a higher boiling point than ammonia.





"IIlIII tI I I I I I tIlllllllllllll










m ----- &im


Figure 5.1. Qualitative representation of the change in properties during the boiling of
the strong solution. Property measurement locations are shown.

However, away from the heat source, the vapor bubbles condense into the

relatively cooler bulk fluid and on the cooler tube walls upon leaving the heater section.

This absorption lowers the vapor fraction of the fluid, and raises the bulk temperature.

Again due to the boiling point of water, more water will condense than ammonia such









that the remaining vapor will be of a higher ammonia mass fraction and the liquid part

will be of a lower ammonia mass fraction. Equilibrium will eventually be reached far

downstream from the boiler exit. The bulk temperature when measured too close to the

boiler exit, as shown in Fig. 5.1, will not necessarily be the equilibrium temperature.

Upon entering the separator, the vapor is extracted from the liquid, with the two

streams exiting at saturation in principle. Even with adequate insulation there are external

losses, and the vapor continues to condense on the separator vessel walls. This raises the

vapor temperature and the ammonia mass fraction in the vapor region. The temperature

of the vapor leaving the separator was found to be 0.74 + 0.16 C higher on average than

the temperature of the weak solution exiting the tank. The vapor fraction, defined as the

ratio of the mass flow rate of vapor to the mass flow rate of the two-phase inlet stream, is

also reduced. This reduces the potential for work and cooling output, as the vapor flow

rate is reduced.

Although water condenses more readily than ammonia from the vapor stream, the

condensate is still very high in ammonia concentration. This raises the ammonia mass

fraction in the weak solution as well, which pools in the separator. The weak solution

temperature at the separator exit was essentially the same as the bulk fluid temperature at

the boiler exit. The incomplete liquid and vapor separation also contributes to the greater

ammonia content in the weak solution than expected. The results confirm that the weak

solution is supersaturated with ammonia as it leaves the separator.

Recovered Heat

The internal recovery is determined from the heat transfer across the recovery

heat exchanger as measured from the weak solution side of the unit, since the weak










solution is not expected to undergo a phase change. The certainty in the determination of

the phase leads to greater certainty in the properties at that state.

The maximum possible heat transfer is the energy change of the lower heat

capacity fluid in having a maximum temperature change. The lower heat capacity fluid is

the weak solution, as it has a lower mass flow rate and should not incur a phase change.

The maximum heat it can reject is from cooling to the strong solution recovery inlet

temperature. For a real heat exchanger of finite surface area, the maximum heat exchange

will not occur and the heat exchanger effectiveness is used to gauge the performance of

the unit.


ambient temperature = 25 C
basic solution mass fraction = 40 % heat source flow ratio = 3.5

14

13
12-- -

S11 cycle with typical losses
10

w 9 --- - -- -- -- ---_ _--_^. -----








4 P= -,t =s .
50 60 70 80 90 100 110
heat source temperature (C)


Figure 5.2. Expected internal heat recovery (kJ/kg) per unit basic solution flow, based on
simulation with typical losses.

Simulated results are shown in Fig. 5.2, where the typical losses are those listed in

Chapter 4. In particular, 85% effectiveness was used for the recovery heat exchanger. In

further studies, correlation of the effectiveness with other parameters could yield a better









model. Results suggest that there is more recovery when the weak solution is hotter, as

for higher boiler temperatures. There is also more recovery for higher pressures, since

there is more weak solution return from the separator as less strong solution is vaporized

in the boiler. The upper left region is where the conditions do not allow boiling, and only

liquid exits from the separator. The observed behavior follows the expected trends.

It is seen from the experimental results that for trials in which boiling initiates in

the recovery unit, the effectiveness of the heat exchange improves. This occurs if the

system operates at pressures well below the dew point pressure. This low pressure allows

the wall temperature of the recovery unit to be high enough to initiate boiling. The initial

system averaged 98.1 7.0% effectiveness for when the boiling initiated in the recovery

unit, and otherwise 75.4 8.2% for when boiling did not initiate until the boiler. There

was no boiling in the recovery unit for the improved system trials, as higher boiler

pressures were evaluated. But with the addition of 286% more heat exchange area, the

recovery effectiveness for trials without boiling initiation increased to 90.9 0.9%.

Boiler Heat Input

The boiler heat input is expected to increase with an increase in vapor production,

which occurs at higher boiler temperatures and lower pressures for a given strong

solution ammonia mass fraction. The combined experimental results for the initial and

improved systems are given in Fig. 5.3, which verify this trend. Using the vapor fraction

is a way to combine the effects of various boiler temperatures, pressures and solution

ammonia mass fractions into one parameter. The mass flow rates of the vapor, strong

solution and heat source are also normalized in using the specific boiler heat input and

the vapor fraction.















400



350



S300
0)


250

C
200
(V
4-

S150
0

100


50 1-


I I I I I I I I I I I I I I I 1 1
5 10 15 20
vapor fraction leaving boiler (%)


Figure 5.3. Boiler heat input per unit basic solution flow, for various vapor fractions

leaving the boiler.


100 -


NJ I
'-, S
I '~


a


70

-

60 -^



50
improved system
0 improved system
40 - initial system
o initial system

30 1 I 1 1 I 1 I 1 1
0 1 2 3 4
heat source flow ratio


, I


Figure 5.4. Boiler heat exchanger effectiveness for the initial and improved systems.









The boiler heat exchanger effectiveness was found decrease with the heat source

flow ratio, such that at lower heat source flow rates, the hot water temperature at the

boiler exit approaches the strong solution inlet temperature. Thus, the heat source can

provide more of the maximum possible boiler input at lower flow rates. Fig 5.4 supports

that the addition of 47% more boiler surface area in the improved system also increased

the boiler effectiveness. Although the heat source flow ratio has a significant effect on the

boiler effectiveness, it is not the only parameter that does. The fluctuations in the figure

are due to other parameters such as the difference in fluid temperatures, boiler pressure,

and solution ammonia mass fraction not being held constant.

Vapor Fraction Leaving the Separator

A vapor fraction between 0 and 1 is obtained from operating the boiler between

the bubble and dew points. More vaporization will occur for higher temperatures and

lower pressures. In the initial system, only 59 3% of possible vapor production was

realized on average as shown in Fig. 5.5. The experimental trends are reasonable,

although the data do not correlate well with the equilibrium based simulations.

With added insulation around the separator and vapor lines, and greater heat

exchange area in both the recovery unit and boiler, the improved system generated 103

7% of possible vapor production on average for the given boiler conditions. A value over

100% of the maximum is the indirect result of a time lag between the strong solution and

vapor flow rates. The vapor flow rate is steady as it is pressure driven from a large tank.

However, the vapor fraction, as a ratio of the vapor and strong solution flow rates, can

seemingly increase past the maximum if the pump flow rate temporarily decreases.












60 C simulated
o 60 C data
- 70 C simulated
70 C data
------- 80 C simulated
0 80 C data


boiler pressure (bar)


Figure 5.5. Vapor fractions in the initial system tests for various boiler temperatures and
pressures, with a basic solution ammonia mass fraction of 45.6%.


- 73.7 C simulated
a 73.7 C data
- 82.9 C simulated
A 82.9 C data
-------- 97.3 C simulated
0 97.3 C data
101.1 C simulated
0 101.1 C data


25.00N




20.00N


I I I ,,,,mm m I iiii I ,,,,I IIII ,I ,,, I ,,,,I ,1 ,,,I, ,,I ,,.,I, ,,,I ,,. I, i,
4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10 10.5
boiler pressure (bar)


Figure 5.6. Vapor fractions in the improved system tests for various boiler temperatures
and pressures, with a basic solution ammonia mass fraction of 38.3%.


40


35


30






0.
1C
| 15


20 1-


\
i\


?\


et~









Likewise, if the pump flow rate increases, then the boiler exit temperature will

drop and equilibrium calculations will reduce the expected vapor output such that the

actual vapor output seems to excel past the calculated limit. For either explanation, the

results are encouraging, as vapor production has increased due to the system

improvements. The vapor fractions of the improved system are shown in Fig. 5.6, where

the maximum values are shown as the simulated results. Note that the highest vapor

fraction seen in the experimentation was 19.8%.

Vapor and Weak Solution Ammonia Mass Fractions

The weak solution and vapor ammonia mass fractions will not be equal to that in

the strong solution, as shown in the bubble and dew point diagram in Fig. B. 1. The vapor

will be formed at a high ammonia concentration owing to its lower boiling point than that

of water, and the corresponding weak solution will have a lower ammonia mass fraction

to maintain an overall mass balance in equilibrium.

The concentrations in the system are not those predicted under equilibrium

conditions. The weak solution leaves the separator at a 2.0 + 0.6% (absolute; ie. if a 30%

ammonia mass fraction is expected, the fluid leaves at 32%) greater ammonia mass

fraction than that predicted from equilibrium at the measured boiler exit temperature,

pressure and strong solution ammonia mass fraction. This is shown in Fig. 5.7, for trials

at various boiler pressures and temperatures. For higher boiler pressures, the trends show

that the weak solution is richer, as more ammonia can be sustained in the liquid for the

given boiler temperature. For higher boiler temperatures the weak solution is leaner, as

more ammonia can vaporize. Water will vaporize more with lower pressure and higher

temperature also, but not in proportion to the ammonia vaporization.








78




0.44 60 C simulated
0.44 6 C data
70 C simulated
C0.42 70 C data
O 80 C data
0 0.4

,0.38

0.36 ,. --

0.34

E 0.32
E -
S0.3-


0.28 --

0.26 -

U0.24 --4

0.22 ,*^

0.21 I I -I I II-- I
4 5 6
boiler pressure (bar)


Figure 5.7. Ammonia mass fraction in the weak solution for various boiler exit
temperatures and pressures, with a basic solution ammonia mass fraction of 45.6%.







0.98 -



.0.96 -

0' -6 C
E0.94 ""'""
E-


092- 0C simulated

7' 0 C data
() 80 C simulated
U) 0.9 0 C data
E -'


0.88 -"
I I 1 1 1 11 1 1 1 I
4 5 6
boiler pressure (bar)


Figure 5.8. Ammonia mass fraction in the vapor for various boiler exit temperatures and
pressures, with a basic solution ammonia mass fraction of 45.6%.







79



The vapor exits the separator also at a higher concentration, with a 1.9 + 0.3%


greater ammonia mass fraction. Some results are shown in Fig. 5.8 and compared to


simulations based on equilibrium. Trend lines for the experimental data are not included,


as more data should be taken to produce a reliable fit. The weak solution and vapor


ammonia mass fractions are shown on the same scale in Fig. 5.9 for one set of trials. The


error in the data that is noted is based on measurement uncertainties. Other discrepancy


with the simulated results can be attributed to inadequate modeling of losses and


unsteady operation.


1 -- __ _.---- ------

0.9

0.8

=- 0.7
0
E
E 0.6

0
a 0.5

0.4 -

a 0.3 -
E
0.2
vaporsimulated
A vapor data
0.1 weak solution simulated
V weak solution data

4 5 6
boiler pressure (bar)


Figure 5.9. Ammonia mass fraction in the weak solution and vapor for a basic solution
ammonia mass fraction of 45.6%, for various boiler pressures. The boiler exit
temperature is 60 'C.

The basic solution concentration is measured only while the system is idle and


assumed to be in equilibrium. The strong solution does not necessarily remain at the basic


solution concentration during operation, and later testing included the measurement of the









operating strong solution ammonia mass fraction. The strong solution tends to become

leaner during operation by 1.5 + 1.0%, on average, relative to the basic liquid ammonia

mass fraction while the system is idle. The increase in ammonia content in the separator

should balance the decrease in the absorber.

Absorption

Experiments suggest that the mixing and absorption of the vapor stream into the

weak solution in the absorber does not occur under equilibrium conditions. Much of the

discrepancy can be attributed to the design of the absorber, in which the heat rejection

occurs above the liquid pool in the vapor region. The two regions operate at different

temperatures, which does not allow for equilibrium to regulate the pressure in the vessel.

This has been a major source of instability while operating the system.

Observations

The vapor is bubbled into the liquid pool of the absorber through small holes to

optimize the surface area available for the absorption to take place. Tested with ammonia

vapor bubbled into an open water pool before installation into the system, the bubbles did

not appear to break the surface. Most of the absorption is expected to take place in the

liquid pool. The vapor could also be absorbed above the pool by the cooled weak solution

spray, although this will not be as effective. Tests conducted with a pool level below the

vapor inlet saw the absorber pressure rise quickly.

The absorption heat is expelled as the cycle heat rejection to the ambient, which is

simulated by the coolant flow. Although the heat is released mostly in the liquid pool, the

cooling occurs above the pool. In order to maintain the desired pool temperature, the

weak solution spray needs to reach the pool below the desired pool temperature. This is

where the problem lies. Essentially the coolant operates below the bubble temperature of









the bulk solution, which will lower the pressure in the absorber as the fluid condenses on

the cooling elements. In order to maintain the desired saturation pressure, excess vapor

must be present in the vapor region.

Control of the absorber pressure is critical for the operation of the solution pump.

If the fluid at the pump inlet is not a compressed liquid, the pump intake stroke will flash

boil the fluid, and the pump flow rate will decrease significantly. Cavitation can be heard

and may damage certain pumps. Fluctuations in pump flow rate will affect other

parameters downstream and the system will not be steady.

There are three ways to regulate the absorber pressure that have been effective. If

less weak solution is returned to the absorber and allowed to instead pool in the separator,

the vapor will build up in the absorber above the pool and the liquid will leave the

absorber compressed. This works, but the imbalance of mass flow rates has to be

accounted for, and the weak solution needs to be brought back to the absorber eventually.

Secondly, an inert, insoluble gas can be added to the vapor region in the absorber to

increase the pressure by about 10-15%. Helium was used and improved the pump flow

rate, but leaked through seals and had to be replenished after a few days. The increase in

absorber pressure also limits the expansion possible across the turbine. Finally, careful

matching of the coolant inlet temperature and flow rate reduces sway in pressure. Later

trials were conducted with the latter two options applied.

Coolant Flow Rate and Temperature

The best coolant temperature was found by trial and error to be about 20 C (36

F) below the desired absorber pool temperature. The chiller maintains the coolant in a 6

C (11 F) range, which is still enough of a fluctuation to change the absorber pressure by