ARTI-21CR/612-10050-01

FUNDAMENTAL UNDERSTANDING OF HEAT AND MASS TRANSFER IN THE AMMONIA/WATER ABSORBER

Interim Report Date Published May 2006

Author:

Dr. Srinivas Garimella

Georgia Tech Research Institute George W. Woodruff School of Mechanical Engineering Atlanta, GA 30332-0841

Prepared for

AIR-CONDITIONING AND REFRIGERATION TECHNOLOGY INSTITUTE 4100 N. Fairfax Drive, Suite 200, Arlington, Virginia 22203 Distribution A Approved for public release; further dissemination unlimited.

Funding for this project was provided by: - Air Conditioning and Refrigeration Technology Institute, Inc., (ARTI 21CR program) Funding for the ARTI 21CR program provided by:
U.S. Department of Energy through DOE Cooperative Agreement No. DE-FC05-99OR22674 Air-Conditioning & Refrigeration Institute (ARI) Copper Development Association (CDA) New York State Energy Research and Development Authority (NYSERDA) California Energy Commission (CEC) Refrigeration Service Engineers Society (RSES) Heating, Refrigeration and Air Conditioning Institute of Canada (HRAI) DISCLAIMER This report was prepared as an account of work sponsored by the Air-Conditioning and Refrigeration Technology Institute (ARTI) under its HVAC&R Research for the 21st Century (21CR) program. Neither ARTI, the financial supporters of the 21CR program, or any agency thereof, nor any of their employees, contractors, subcontractors or employees thereof - makes any warranty, expressed or implied; assumes any legal liability or responsibility for the accuracy, completeness, any third partys use of, or the results of such use of any information, apparatus, product, or process disclosed in this report; or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute nor imply its endorsement, recommendation, or favoring by ARTI, its sponsors, or any agency thereof or their contractors or subcontractors. The views and opinions of authors expressed herein do not necessarily state or reflect those of ARTI, the 21CR program sponsors, or any agency thereof.

ARTI-21CR/612-10050-01

FUNDAMENTAL UNDERSTANDING OF HEAT AND MASS TRANSFER IN THE AMMONIA/WATER ABSORBER

Interim Report Date Published May 2006 Author: Dr. Srinivas Garimella

Prepared for

AIR-CONDITIONING AND REFRIGERATION TECHNOLOGY INSTITUTE Under ARTI 21CR Program Contract Number 612-10050

EXECUTIVE SUMMARY An experimental investigation of heat and mass transfer in falling-film ammonia-water absorbers was conducted. A tube bank consisting of four columns of six 9.5 mm (3/8) nominal OD, 29.2 cm (11.5) long tubes was installed in an absorber shell that allowed heat and mass transfer measurements and optical access. A test facility consisting of all the components of a functional absorption chiller was fabricated specifically for this investigation. Thus, a steamheated desorber was used to generate ammonia-water vapor over a wide range of conditions. The ammonia-water vapor was rectified and condensed, followed by recuperative heat exchange before being expanded to the evaporator pressure. The cooling load was supplied to the evaporator by a combination of resistance heating and closed loop fluid heating. The evaporated refrigerant was preheated recuperatively by the fluid exiting the condenser and then flowed to the absorber to be absorbed in the test section. A solution heat exchanger between the absorber and desorber (and in some tests, a solution pre-conditioner upstream of the absorber) completed the ammonia-water loop. Several variations of the basic system set up were fabricated to enable testing over the wide range of conditions (nominally, desorber solution outlet concentrations of 5 - 40% for three nominal absorber pressures of 150, 345 and 500 kPa, over solution flow rates of 0.019 0.034 kg/s.) Heat transfer rates were measured independently for both sides of each component and energy balances were rigorously established based on mass, species and enthalpy balances for each component before test results were deemed acceptable. Care was also taken throughout the study to not only establish the desired conditions, but also to maintain the solution-side thermal resistance as the governing resistance so that absorption heat and mass transfer phenomena could be measured accurately. Component level measurements at the absorber were used to determine heat transfer rates, overall thermal conductances, and solution-side heat transfer coefficients for each test condition. The trends in heat transfer coefficients were plotted and discussed, highlighting the effect of solution flow rate, with preliminary interpretations of the effects of solution concentration and absorber pressure on the solution heat transfer coefficients. For the range of experiments conducted, the solution heat transfer coefficient varied from 558 to 1140 W/m2-K, depending on the test condition. The solution heat transfer coefficient increased with increasing solution flow rate. The preliminary heat transfer coefficient estimates reported here were based on approximate analyses that treat the absorber as a single, lumped component, and do not account in detail for the internal variations in heat and mass transfer coefficients. Some of the local measurements also taken during the experiments will be used in the ongoing second phase of this study, in which the development of more detailed models of the absorption process will be attempted.

TABLE OF CONTENTS INTRODUCTION .......................................................................................................................... 1 Background ................................................................................................................................. 1 Issues in Ammonia-Water Absorption........................................................................................ 1 RELEVANT LITERATURE.......................................................................................................... 3 Horizontal and Vertical Tube Falling-Film Absorption ............................................................. 3 Bubble-Type Absorbers .............................................................................................................. 3 Miniaturization of Absorbers...................................................................................................... 4 Research Needs........................................................................................................................... 5 PROJECT SCOPE AND OBJECTIVES ........................................................................................ 6 Scope........................................................................................................................................... 6 Specific Objectives ..................................................................................................................... 7 EXPERIMENTAL APPROACH.................................................................................................... 9 Test Facility Overview................................................................................................................ 9 Details of Individual Components .............................................................................................. 9 Absorber and Absorber Coolant Loop.................................................................................... 9 Desorber and Steam-Heated Auxiliary Loop........................................................................ 12 Condenser and Condenser Coolant Loop ............................................................................. 13 Evaporator and Evaporator Heating Loop ............................................................................ 14 Separator ............................................................................................................................... 15 Rectifier................................................................................................................................. 15 Solution Heat Exchanger ...................................................................................................... 16 Pre-Cooler ............................................................................................................................. 16 Expansion Valve ................................................................................................................... 17 Pressure Reduction Valve ..................................................................................................... 17 Solution Pump....................................................................................................................... 18 Instrumentation ......................................................................................................................... 18 Temperature and Pressure Measurement .............................................................................. 18 Flow Meters .......................................................................................................................... 18 Data Acquisition System....................................................................................................... 19 Tailoring the Test Facility for Wide Operating Range ............................................................. 21 Test Facility, Version I (Initial Approach to Stable Operation) ........................................... 21 Test Facility (Version II) ...................................................................................................... 24 Test Facility (Version III) ..................................................................................................... 26 Experimental Procedures .......................................................................................................... 28 Safety Precautions................................................................................................................. 28 Leak Testing and Charging................................................................................................... 29 Testing................................................................................................................................... 30 DATA ANALYSIS....................................................................................................................... 33 Calculation Methodology.......................................................................................................... 33 Absorber................................................................................................................................ 35 Desorber................................................................................................................................ 37 Evaporator............................................................................................................................. 38 Condenser ............................................................................................................................. 39 Rectifier................................................................................................................................. 40

ii

RESULTS AND DISCUSSION ................................................................................................... 41 Test Matrix................................................................................................................................ 41 Energy Balance Quantification and Preliminary Uncertainties ................................................ 44 Absorber Heat Duty, Overall U, and Solution-Side Heat Transfer COefficient ...................... 47 Effect of Dilute Solution Concentration ................................................................................... 48 Effect of Absorber Pressure ...................................................................................................... 50 CONCLUSIONS........................................................................................................................... 52 WORK PLANNED FOR PHASE II............................................................................................. 53 REFERENCES ............................................................................................................................. 54 APPENDIX A............................................................................................................................... 57 OVERALL, COOLANT, AND SOLUTION HEAT TRANSFER COEFFICIENTS ................. 57

iii

LIST OF TABLES Table 1. Originally Proposed Effort................................................................................................6 Table 2. Instrumentation Specifications Summary........................................................................20 Table 3. Test Conditions for the Present Study ..........................................................................42 Table 4. Primary Challenges and Techniques Used .....................................................................43 Table 5. Energy Balances for Absorber, Condenser and Evaporator ...........................................44 Table 6. Uncertainties in Key Parameters.....................................................................................45

iv

LIST OF FIGURES Fig 1. Falling-Film Absorption Inside and Outside Vertical Tubes ..............................................2 Fig 2. Representative Forced Convective (Bubble) Absorber ........................................................2 Fig 3. Falling-Film Absorption on Horizontal Tube Banks............................................................2 Fig 4. Proposed and Approved Microchannel Absorber Test Geometry........................................6 Fig 5. Single-Pressure Absorber Heat and Mass Transfer Test Facility.........................................7 Fig 6. Drawing of the Absorber Outer Shell.................................................................................10 Fig 7. Drawing of the Absorber Tube Array..................................................................................10 Fig 8. Revised Absorber Photograph (Installed Assembly, Tube Bundle)...................................10 Fig 9. Tube Array Dimensions.......................................................................................................11 Fig 10. Photograph of the Desorber and Evaporator ....................................................................12 Fig 11. Steam Heat Exchanger......................................................................................................13 Fig 12. Photograph of the Condenser ............................................................................................13 Fig 13. Plate Heat Exchangers .......................................................................................................14 Fig 14. Photograph of the Separator ..............................................................................................15 Fig 15. Photograph of the Rectifier...............................................................................................16 Fig 16. Solution Heat Exchanger ...................................................................................................16 Fig 17. Photograph of the Pre-Cooler ............................................................................................17 Fig 18. Test Facility Schematic, Version I ....................................................................................21 Fig 19. Photograph of the First Version of the Test Facility .........................................................22 Fig 20. Test Facility Schematic, Version II ..................................................................................24 Fig 21. Test Facility Photograph, Version II .................................................................................25 Fig 22. Dilute Solution SubCooler................................................................................................28 Fig 23. Test Facility Schematic, Version III (Current)..................................................................29 Fig 24. Test Facility Photograph, Version III (Current) ...............................................................30 Fig 25. Schematic of the Desorber.................................................................................................33 Fig 26. Schematic of the Rectifier .................................................................................................34 Fig 27. Separator Liquid-Phase Balance.......................................................................................35 Fig 28. Schematic of the Absorber ...............................................................................................35 Fig 29. Desorber and Separator Balance.......................................................................................38 Fig 30. Evaporator Energy Balance ..............................................................................................38 Fig 31. Condenser Energy Balance...............................................................................................39 Fig 32. Concentration Ranges at Nominal 150 kPa (21.8 psi).......................................................45 Fig 33. Concentration Ranges at Nominal 345 kPa (50.0 psi)......................................................46 Fig 34. Concentration Ranges for Nominal 500 kPa (72.5 psi)....................................................46 Fig 35. Absorber Duty vs. Solution Flow Rate..............................................................................47 Fig 36. Overall U vs. Solution Flow Rate......................................................................................47 Fig 37. Solution-Side Heat Transfer Coefficient vs. Flow Rate ....................................................48 Fig 38. Solution Heat Transfer Coefficient vs. Flow Rate at 150 kPa (21.8 psi) .........................48 Fig 39. Solution Heat Transfer Coefficient vs. Flow Rate at 345 kPa (50.0 psi) .........................49 Fig 40. Solution Heat Transfer Coefficient vs. Flow Rate at 500 kPa (72.5 psi) .........................49 Fig 41. Solution Heat Transfer Coefficient vs. Flow Rate (xdes,out = 5% Nominal)......................50 Fig 42. Solution Heat Transfer Coefficient vs. Flow Rate (xdes,out = 15% nominal) ....................50 Fig 43. Solution Heat Transfer Coefficient vs. Flow Rate (xdes,out = 25% nominal) .....................51 Fig 44. Solution Heat Transfer Coefficient vs. Flow Rate (xdes,out = 40% nominal) ....................51

Fig A.1. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 5%, 150 kPa (21.8 psi) ..........................................................................................................58 Fig A.2. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 5%, 345 kPa (50.0 psi) ............................................................................................................59 Fig A.3. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 5%, 500 kPa (72.5 psi) ............................................................................................................60 Fig A.4. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 15%, 150 kPa (21.8 psi) ............................................................................................................61 Fig A.5. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 15%, 345 kPa (50.0 psi) ............................................................................................................62 Fig A.6. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 15%, 500 kPa (72.5 psi) ............................................................................................................63 Fig A.7. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 25%, 150 kPa (21.8 psi) ............................................................................................................64 Fig A.8. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 25%, 345 kPa (50.0 psi) ............................................................................................................65 Fig A.9. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 25%, 500 kPa (72.5 psi) ............................................................................................................66 Fig A.10. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 40%, 150 kPa (21.8 psi) ............................................................................................................67 Fig A.11. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 40%, 345 kPa (50.0 psi) ............................................................................................................68 Fig A.12. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 40%, 500 kPa (72.5 psi) ............................................................................................................69

vi

INTRODUCTION Background National and international interest in the global climate change problem has focused renewed attention (Ryan 2002) on the use of absorption heat pumps, which are environmentally sound and energy-efficient alternatives to CFC-based, ozone-depleting space-conditioning systems. The Second Assessment Report of the Intergovernmental Panel on Climate Change (1995) prepared by 2,500 of the worlds leading experts on climate change concludes that the balance of evidence suggests that human activities are having a discernible influence on global climate. The IPCC further concluded that without specific policies that reduce the growth of emissions, the Earths average temperature will increase by about 1 to 3.5oC by 2100 - the fastest rate since the last ice age 10,000 years ago. The heat-driven absorption systems, particularly those employing the more advanced cycles such as the Generator-Absorber-Heat Exchange (GAX) cycle (Engler et al. 1996; Garimella et al. 1996) could offer higher source efficiencies than corresponding vapor compression systems in the heating mode, and when powered by waste heat, offer the potential to be part of an overall Integrated Energy System that provides building cooling, heating, and power. Advanced cycles investigated for absorption space-conditioning systems include double-effect ((Gommed and Grossman 1990; Garimella et al. 1992; McGahey et al. 1994), triple-effect cycles (DeVault and Marsala 1990; Grossman et al. 1994; Ivester and Shelton 1994; Garimella et al. 1997), and even quadruple- and multiple-effect absorption chillers (Ziegler and Alefeld 1994; Grossman et al. 1995). However, the increased cycle complexity also results in the need for numerous heat exchangers, which have thus far proven to be complicated and expensive to fabricate, representing one of the biggest hurdles to commercialization. Issues in Ammonia-Water Absorption The absorber, in which refrigerant vapor is absorbed into the dilute solution with the release of a substantial amount of heat of absorption, governs the viability of the entire cycle and has been referred to as the bottleneck in the heat pump (Beutler et al. 1996). However, a lack of understanding of this inherently complicated process has led to the use of poor designs employing inappropriate and often grossly expensive and oversized heat and mass exchangers. The ammonia-water fluid pair (unlike the LiBr/H2O fluid pair) has a volatile absorbent, thus presenting both heat and mass transfer resistances across the respective temperature and concentration gradients in both the liquid and vapor phases. The highly non-ideal ammoniawater fluid pair releases a considerable amount of heat of absorption at the vapor-liquid interface that must be transferred across a liquid film into the coolant. Some of this heat released at the interface is also transferred to the vapor, depending on the local temperature differences. Furthermore, at different regions in the overall absorption process, particularly over the extremely wide concentration and temperature range of cycles such as the GAX cycle, the local concentrations in the bulk and interface liquid and vapor phases could be such that species are desorbed into the vapor phase, rather than being absorbed. The coupled heat and mass transfer processes in both phases have presented challenges for analysis, modeling, experimental validation, and design. In addition, the fluid flow in the respective phases is rarely well defined (as a smooth, laminar film, for example). In-tube, forced-convective absorbers operating in the annular flow regime, and vertical falling-film absorbers with a countercurrent vapor core (Fig 1), are often subject to, and governed by, turbulent mixing within the phases, and interfacial waves.

Similarly, when operating in the intermittent or bubble flow regime (Fig 2), velocity, temperature and concentration gradients within the liquid slug, the vapor bubble, the annular film surrounding the bubble, and the acceleration/deceleration of the two phases as the liquid slows down or speeds up while traveling between the slug region and the annular film region must be taken into account. Falling-film absorbers (especially horizontal tube banks, Fig 3) must also account for the wavy flows and their effects on the transfer processes. In many of these configurations, incomplete wetting of the transfer surfaces leads to considerable deterioration of the performance from the fully wetted rates. Comprehensive critical reviews of absorption heat and mass transfer Fig 1. Falling-Film Absorption Inside and models (Killion and Garimella 2001) and Outside Vertical Tubes experiments (Killion and Garimella 2003) are available in the literature. These reviews have pointed out that research on ammonia-water absorption undertaken by various investigators has often resulted in seemingly conflicting conclusions, although further analysis could point out that the conclusions are not so much conflicting as being limited in the extent of their applicability. Based on the above discussion, any of a number of transport operations may control the entire absorption process, also with different resistances being dominant in different regions and for different conditions. A brief review of the literature on absorption heat and mass transfer is provided in the Fig 2. Representative following Forced Convective section. Based (Bubble) Absorber on this review, the deficiencies in the understanding of absorption phenomena are pointed out, and the objectives for the current project are formulated. This is followed by a discussion of the project scope and tasks, the experimental approach and data analysis, and a preview of the analytical work to be done in Phase II of this project. 2

Fig 3. Falling-Film Absorption on Horizontal Tube Banks

RELEVANT LITERATURE Horizontal and Vertical Tube Falling-Film Absorption Among the ammonia-water investigations pertinent to the subject project is the experimental study of a coiled-tube absorber (Jeong et al. 1998) in which dilute solution flowed over the outside of the tubes with the ammonia vapor flowing upward in the shell and coolant flowing within the tube. They conducted experiments with and without absorption and found that film heat transfer coefficients were lower for experiments with absorption than those without absorption, perhaps due to insufficient wetting caused by vapor shear. For laminar, steady flow at low Reynolds numbers with uniform wetting, Perez-Blanco (1988) presented a simple 1-D model for the absorption process in a horizontal-tube, falling-film absorber, while accounting for water transport both into and out of the solution film. To avoid unrealistic excessively high mass transfer rates at the interface, he used a two-film model to decouple the interface concentrations of ammonia in the liquid and vapor. The film heat transfer coefficient was found using simple penetration theory (Higbie 1935; Ruckenstein and Berkente 1968). The solution flow rate and liquid-phase mass transfer coefficient influenced absorber performance significantly, with sensitivity to other parameters being small, although the relative importance of the coolant heat transfer coefficient increased as the solution approached saturation conditions. He also found that excessive water evaporation and migration from hot to cold areas in the vapor phase resulted in performance degradation. Potnis et al. (1997) developed a computer program that simulated the GAX process with liquid-film absorption over a coiled fluted tube with countercurrent vapor flow, and convective desorption inside the fluted tube. The Colburn and Drew (1937) approach was used to model absorption as the condensation of binary mixtures. They found that the mass transfer resistance was primarily in the vapor phase, although the liquid-phase mass transfer resistance was not negligible. Panchal et al. (1997) developed a model of condensation of ammonia-water mixtures flowing on the outside of vertical tubes using the Colburn and Drew (1937) approach, and compared the results of their model with tests on similar geometries conducted at Oak Ridge National Laboratory. Their work focused not on absorption into an incoming liquid solution stream, but rather on condensation of a vapor stream of varying ammonia-water concentrations. Using different assumptions for liquid-phase mass transfer including complete mixing and no mixing, they predicted that the complete mixing in the liquid phase assumption approximated the data better than the other limiting cases. They also concluded that vapor-phase diffusion presented a significant resistance to the binary-fluid condensation process. Takuma et al. (1993) analyzed condensation of ammonia-water mixtures on horizontal tube bundles using a heat and mass transfer analogy and confirmed the predictions using measurements on a full-scale shell-and-tube condenser as well as a test apparatus using the Coherent Anti-Stokes Raman Spectroscopy technique for ammonia concentration measurement. They concluded that the accumulation of ammonia at the interface presents an important resistance to condensation. Attempts at obtaining compact ammonia-water absorber geometries include counter-current fluted-tube absorbers (Kang and Christensen 1994, 1995). A generalized design tool capable of modeling several components within an absorption system was presented by Kang et al. (1997). Bubble-Type Absorbers A vertical-tube bubble absorber with co-current solution and vapor flow in an inner tube, and countercurrent coolant flow in the annulus, was modeled by Herbine and Perez-Blanco 3

(1995). Several simplifying assumptions were made including: the absorption process was at steady state, the process occurred at a constant pressure, all bubbles had the same diameter and velocity at a given location along the absorber length, bubble breakup and coalescence were negligible, resistance to mass transfer in the bubble was negligible, and no direct heat transfer took place between the vapor and the coolant. They found that as the ammonia concentration within the bubble decreased to that at equilibrium, water transfer was in the opposite direction of ammonia transfer; however, when the ammonia concentration dropped below that at equilibrium, water and ammonia transfer were in the same direction. Dence et al. (1996) proposed a GAX absorber-desorber component, which used a complicated annulus-shaped co-flow bubble absorber with a helical bubble injector and simulated desorption on both absorber walls using helical cooling water channels. Very simple two-phase heat transfer correlations (without mass transfer analyses) were used to design the component, and the paper does not report experimental results. Kang et al. (1998) evaluated the heat and mass transfer resistances in both the liquid and vapor regions in a countercurrent ammonia-water bubble absorber composed of a plate heat exchanger with offset strip fin inserts. They concluded that the heat transfer resistance was dominant in the vapor phase, while the mass transfer resistance dominated in the liquid phase. In addition, the interfacial mass transfer area and vapor hold up were found to be significant in determining the size of the absorber. Merrill et al. (1994; 1995) used numerous passive enhancement techniques on verticaltube bubble absorbers for the GAX cycle such as repeated roughness elements, internal spacers, and increased thermal conductivity metal to improve heat transfer, and mass transfer improvement was achieved through the use of static mixers, variable cross-section flow areas, and numerous vapor injector designs. In a further study on bubble absorption, Merrill and PerezBlanco (1997) investigated increasing the interfacial area per unit volume of vapor and liquid mixing at the vapor-liquid interface by breaking the vapor up into small bubbles and injecting them into the liquid. Miniaturization of Absorbers Miniaturization efforts for ammonia-water absorbers have included a compact absorber consisting of a corrugated and perforated fin surface placed between rectangular coolant channels (Christensen et al. 1998). In this design, which was tested and analyzed by Garrabrant and Christensen (1997), the vapor flows upward through perforations in the corrugated fins, while absorbent solution flows downward over the corrugated fins and through the perforations. More recently, a miniaturization technology for ammonia-water absorbers (1999; Garimella 2000; Garimella 2004) that uses short lengths of very small diameter tubes placed in a square array, with several such arrays being stacked vertically, has been reported. Successive tube arrays are oriented in a transverse orientation perpendicular to the tubes in adjacent levels. Liquid solution flows in the falling-film mode counter-current to the coolant through the tube rows. Vapor flows upward through the lattice formed by the tube banks, counter-current to the falling solution. The effective vapor-solution contact minimizes heat and mass transfer resistances, the solution and vapor streams are self-distributing, and wetting problems are minimized. Coolant-side heat transfer coefficients are extremely high without any passive or active surface treatment or enhancement, due to the small tube diameter. A preliminary model using the Price and Bell (1974) approach demonstrated that a 19 kW absorber, which corresponds to a 3-ton cooling system, can be built in a very small 0.127 m x 0.127 m x 0.476 m envelope. Extensions of the compact, modular concept to other components in an absorption 4

system were also noted. Meacham and Garimella (2002a) demonstrated a prototype of this concept and achieved absorption duties of up to 16 kW. In a subsequent integrated analytical and experimental study, Meacham and Garimella (2003) developed an experimentally validated absorption heat and mass transfer model to predict the detailed temperature, concentration and mass flow rate profiles through the absorber. A surface area effectiveness ratio was defined to account for potential solution distribution and surface wetting problems, which resulted in excellent agreement between the predictions of the model and the measured data both for the overall absorber as well as the local level. Improved absorption rates were then achieved, guided by high speed visualization of the flow patterns and a revised distribution device, which substantially improved the solution distribution and wetting of the tubes (Meacham and Garimella 2004). In related research (Meacham and Garimella 2002b), a shell-and-tube heat exchanger with small diameter tubes was also demonstrated as an absorber for residential heat pumps. With a high surface/volume ratio, it was demonstrated that an extremely compact absorber (76.2 mm outside diameter, 0.508 m long) with vapor and solution flowing co-current through 253 tubes of 2.54 mm ID can transfer the absorption duty of a 10.55 kW cooling load absorption heat pump. Experiments over a wide range of solution and coolant flow rates and vapor fractions were used to deduce overall and tube-side absorption heat transfer coefficients. High performance was achieved without any surface enhancement, and with very low solution and coolant pressure drops. It was noted that improvement in the distribution of the vapor-liquid mixture would further reduce potential mal-distribution at the entrance, resulting in even larger transfer rates. Research Needs Based on the above discussion, it is clear that: absorption of ammonia-water vapor mixtures into an ammonia-water liquid solution is fundamentally different from condensation of single-component vapors due to the associated mass transfer resistances an understanding of the absorption process requires quantification of the heat and mass transfer resistances in the vapor and liquid phases and the assessment of their relative significances over a wide range of concentrations Experiments with theory-based validation are necessary to address these needs.

PROJECT SCOPE AND OBJECTIVES Scope The solicitation for this project recognized that promising absorption cycles were being severely under-utilized because of under performance and high cost of absorbers, and that better designs need an improved understanding of binary fluid heat and mass transfer. Therefore, potential contractors were asked to submit proposals with the following objectives: a) quantify ammonia absorption into water; explain/identify degradation relative to pure fluids; develop experimentally validated models. Investigation of two different configurations for absorption was solicited; for which an effort to study the following two geometries was proposed: a) fallingfilms over smooth Table 1. Originally Proposed Effort horizontal tubes; bubble Pressure Solution Notes absorption inside Concentration smooth round tubes. (% NH3) Investigations were 10 Multiple solution 150 kPa proposed for absorber Cold ambient heat pump 25 flow rates and pressures of 150 kPa mode and refrigeration 40 vapor fractions to and 500 kPa, and achieve a range of solution concentrations temperature glides of 10, 25, and 40% 10 500 kPa and flow ammonia. These Nominal cooling mode and 25 parameters conditions are warm-ambient heat pump 40 summarized in Table 1. Although the proposed effort was approved, due to available budget constraints, the scope was reduced to include only one configuration, namely falling-film absorption on a microchannel absorber configuration (Fig 4), for the originally specified conditions (Table 1). Illumination Upon project initiation, at the direction of the Project Monitoring Subcommittee (PMS), the following changes were made to project scope: 1. To address the concern that the test geometry included in the work statement might limit the applicability of the results to very small diameter tubes, the test geometry was changed to 9.525 mm (3/8") diameter horizontal tubes. Because the existing test facility and required utilities at the lab were more suited to testing microchannel absorbers, major modifications to the
Coolant Outlet Dripper Tubes Solution Inlet

Coolant Inlet

Optical Access Vapor Inlet Solution Outlet

Fig 4. Proposed and Approved Microchannel Absorber Test Geometry 6

Fig 5. Single-Pressure Absorber Heat and Mass Transfer Test Facility auxiliary equipment had to be made to allow this change in scope. 2. The test conditions were increased to include three absorber pressures instead of the originally proposed two pressures. Thus, tests at a nominal pressure of 345 kPa (50.0 psia) moderate ambient heat pump mode were added to the originally proposed scope. 3. To extend the applicability of the results to systems that depend on a wide concentration range such as GAX cycles, tests at one additional nominal desorber solution outlet concentration of 5% were added to the test matrix at each pressure under consideration. 4. It was further stipulated by the PMS that the achievement of absorber conditions representative of such a wide range of system operating conditions required a functional absorption cooling system (including rectifier, evaporator, condenser, refrigerant pre-cooler, their respective cooling/heating loops, and associated controls and instrumentation), instead of the initially proposed and approved single-pressure absorber test facility (Fig 5). In view of the above major changes to the project scope, a completely new absorber and a full-scale absorption chiller system test facility was designed and fabricated (including desorber, separator, solution heat exchanger, evaporator, condenser, rectifier, steam heating loop, evaporator heating loop, condenser, rectifier and absorber cooling loops with the corresponding heat exchangers, and the necessary instrumentation and control). Specific Objectives Based on the solicitation, the proposed effort, the approved scope and the subsequent modifications to the scope, the specific objectives for the present research are as follows:

Conduct experiments on ammonia-water absorption heat and mass transfer for a horizontal tube bank over the range of nominal desorber solution outlet concentrations 5 - 40% for three nominal absorber pressures 150, 345 and 500 kPa Analyze and model heat and mass transfer processes in falling-film ammonia-water absorption over a horizontal tube bank Use an integrated analytical and experimental approach to propose ammonia-water absorption heat and mass transfer models that can be used as experimentally validated tools for the design of absorbers

ARTI and the PMS instructed the project team to conduct the experimental tasks of the project first, followed by the modeling activity. This interim report pertains to the experimental task, in which the tests outlined above and preliminary data analyses were conducted. The following section of the report documents the test facility construction and the details of the experimental approach.

EXPERIMENTAL APPROACH The experiments for this project were conducted at the Sustainable Thermal Systems Laboratory at Georgia Institute of Technology. The following sections describe the development of the test facility and the details of the absorber geometry. The test facility development is discussed in various phases to explain the manner in which the numerous heat source, sink and fluid connection and control issues were addressed to accomplish the testing over the wide range of operating conditions. Test Facility Overview The revised test facility replicates a steam-fired absorption chiller system, and includes an absorber, desorber, condenser, evaporator, rectifier, solution heat exchanger and refrigerant pre-cooler, together with the coupling loops and heat exchangers for the externally coupled major components. The system operates at two different nominal pressure levels: high pressure at the condenser, rectifier and desorber, and low pressure at the absorber and evaporator. The different solution concentrations at different absorber pressures are obtained by controlling heat duties and temperatures of the desorber, condenser, rectifier and absorber cooling loops. It should be noted that several different versions of the test facility were designed, constructed and installed during the course of this project to address different regions of the operating condition ranges. A general description of the system layout is first described here (without schematic diagrams) followed by detailed descriptions of the components and the different system layouts. Concentrated solution is pumped from the low pressure at the absorber to the desorber at high pressure. Before entering the desorber, the solution flows through a solution heat exchanger, which recuperatively heats the concentrated solution toward the desorber inlet conditions. At the desorber, necessary heat input is supplied to generate ammonia vapor and dilute solution from the concentrated solution. The ammonia vapor is separated from the dilute solution in a vapor-liquid separator, after which, the vapor flows to the rectifier, while dilute solution flows through the solution heat exchanger. Downstream of the solution heat exchanger, the pressure of the dilute solution is reduced by an expansion device before it flows into the absorber to complete the solution circuit. The ammonia vapor is purified at the rectifier and sent to the condenser, while the reflux at the rectifier, produced during rectification, is mixed with the dilute solution exiting the separator. The ammonia vapor is condensed in the condenser and flows to the pre-cooler. Here, the ammonia enthalpy is reduced by recuperatively heating the ammonia vapor stream from the evaporator outlet. At this reduced enthalpy, the ammonia solution is expanded to the evaporator pressure using an expansion device and flows through the evaporator, where it receives the cooling load. Ammonia exiting the evaporator flows through a pre-cooler, where it serves as the coolant for the condenser outlet stream. After exiting the precooler, the ammonia vapor is introduced to the absorber, where it is absorbed in the dilute solution, while rejecting heat to a coolant stream, thus completing the refrigerant circuit. The details of these components and the various primary and secondary loops are described below. Details of Individual Components Absorber and Absorber Coolant Loop The absorber is the main component for this project in which ammonia vapor flowing from the evaporator is absorbed by the dilute solution from the desorber, rejecting its heat to a

Dilute Solution In coolant loop. Fig 6 shows a drawing of the outer shell, while Fig 7 shows a Vapor In drawing of the tube array, respectively. In this falling-film absorber, solution flows by gravity from the top to the bottom of the tube array. The heat of absorption is transferred to coolant flowing through the tubes. The absorber tube array is housed in 49.5 cm long x 35.6 cm diameter (19.5 long x 14 diameter) outer shell with a large 26.7 cm (10) port and three Concentration Solution Out additional 6.4 cm (2.5) sight ports for illumination and viewing at other angles. Fig 6. Drawing of the Absorber Outer Shell The tube array inside the shell consists of four columns of 9.5 mm (3/8) nominal OD, 0.7 mm (0.28) wall thickness, 29.2 cm (11.5) long tubes, each column containing 6 tubes for a total of 24 tubes in the bundle. The two headers at either side of the tube array are 13.0 cm (5.12) long, 12.0 cm (4.72) wide, and 2.5 cm (1) in depth. The drip tray that distributes the Dilute Solution In dilute solution from above the tube array is 34.3 cm (13.5) long x 12.0 cm (4.72) Drip Tray wide x 3.8 cm (1.5) in height. The Vapor In absorber as installed in the loop is shown Cooling Water Out in Fig 8.

The large port of the outer shell is Cooling Water In located at the front and equipped with a sight glass that facilitates visualization. Concentration Solution Out The large front port is equipped with a Bweld sight glass manufactured by Pressure Fig 7. Drawing of the Absorber Tube Array Products Company, Inc. It consists of a stainless steel body, cap, cap screws, compression ring, lens, lens packing, compression adjustment screws and two cushion gaskets. The body dimensions are 31.8 cm (12.5) OD and 8.9 cm (3.5) height (or thickness). The view port diameter is 15.2 cm (6). The 15.7 cm OD (6 3/16), 5.1 cm (2) thick tempered borosilicate lens can withstand a maximum pressure of 2068 kPa (300 psi). Two of the smaller ports are located at the top and bottom and one is located at

Fig 8. Revised Absorber Photograph (Installed Assembly, Tube Bundle)

10

the back, aligned with the large front port, and therefore used to illuminate the inside of the shell. A Bulls Eye NPT 3.8 cm (1) sight glass from Pressure Products Company, Inc. rated for 4137 kPa (600 psi), 232.2oC (450oF) service is installed at the back port. Teflon gaskets, compression rings, and O-rings are used for the sight glasses for compatibility with ammonia. The two other ports at the top and bottom are used as solution inlet and outlet, respectively. The tube array (Fig 9) is connected to one of the flanges and inserted from one end of the shell. Ammonia vapor is introduced into the chamber through this flange. The tubes are arranged in a serpentine configuration with a horizontal pitch of 3.0 cm (1.2) and vertical pitch of 2.0 cm Fig 9. Tube Array Dimensions (0.79), and a surface area of 0.2098 m2 (2.258 ft2). The tube array was initially designed for one header and U-bends at the other to achieve the serpentine flow arrangement; fabricating the U-bends with such a small bend radius using stainless steel was not feasible, which necessitated the two headers. The two headers have six coolant passes from the bottom to the top of the assembly with a total of seven ports for temperature measurement at each level. Thermocouples are placed at each level of the absorber to allow determination of row-wise heat duties, which, in turn, will assist in the estimation of the variation of heat and mass transfer coefficients within the absorber. The drip tray is placed above the tube bundles, supported by the headers and distributes dilute solution on tubes. The tray has 4 rows of 75 holes (on a 3.0 cm x 0.38 cm or 1.2 x 0.15 pitch) for the insertion of 1.5 mm (0.06) capillary tubes (a total of 300), which distribute solution on the first row of the tube bundle. The distributor is located 15.0 mm above the centerline of the first row tubes. The headers and drip tray are welded from 1.5 mm (0.06) thick 316 stainless steel plate. Absolute pressures and temperatures at the solution inlet, outlet and vapor inlet are measured using absolute pressure transducers (Rosemount Model 2088). Measurements of pressure and temperature at these three different locations enable independent estimates of thermodynamic states at these locations. A closed coolant loop using distilled water as the coolant is used to remove the heat of absorption from the absorber, which is eventually transferred to the lab chilled water-glycol coolant (cooled by a 50 RT Carrier AquaSnap Chiller) in a plate heat exchanger (Superchanger Model UX-016-UJ-21 by Tranter with a total heat transfer area of 1.65 m2 or 17.8ft2). Much of the plumbing for the closed coolant lines is fabricated using 25.4 mm (1) nominal O.D. tubing to minimize frictional pressure drops except at the inlet and outlet of the absorber. To measure the coolant pressure drop in the absorber, a differential pressure transducer by Rosemount is

11

used. The coolant is circulated using a centrifugal pump (Little Giant Model TE-7-MD-HC) with a hp motor. An expansion tank is used upstream of the pump inlet to compensate for thermal expansion of water and to provide additional positive head for the pump. The coolant flow rate is measured using a magnetic flow meter (Rosemount Model 8711). Temperatures of the absorber coolant and the chilled water are measured at the inlets and outlets of the plate heat exchanger. Chilled water-glycol flow rate is measured using a rotameter. An Fig 10. Photograph of the Desorber and Evaporator energy balance at the plate heat exchanger provides an additional means of verifying the accuracy of results over and above the energy balance between the ammonia and primary coolant in the absorber. Desorber and Steam-Heated Auxiliary Loop The desorber (Fig 10) is a 5.9 m (19.3 ft) long coiled tube-in-tube heat exchanger from Exergy, Inc. (Model 00528), with a 12.7 mm (0.5) O.D. inner and 25.4 mm (1 O.D.) outer tube, each 1.7 mm (0.065) thick for a total heat transfer area of 0.23 m2 (2.5 ft2). The maximum rated tube-side and annulus-side pressures are 31026 and 13790 kPa (4500 psi and 2000 psi), respectively at 93.3oC (200oF). Concentrated solution enters the desorber and due to the heat supplied by the steam (supplied by a boiler at up to 1379 kPa (200 psi) saturation pressure), exits as a two-phase mixture of dilute solution and ammonia vapor at the top of the desorber. Concentrated solution flows inside the inner tube while the heating fluid flows through the annulus. The desorber essentially determines the concentration of the dilute solution, although reflux from the rectifier changes the concentration of dilute solution entering the absorber to some extent. The concentration of the dilute solution is determined at the co-flow desorber outlet from pressure and temperature measurements and the two-phase exit condition. The twophase mixture flows to the separator where the two streams are separated. Solution temperatures are measured using 3-wire RTDs and pressures are measured using Rosemount (Model 2088) pressure transducers at the inlet and outlet of the desorber. The desorber pressure is determined by temperature at the condenser, which is in turn set by the coolant temperature. In the initial design, the concentrated solution was heated by a heat transfer fluid (Syltherm HF fluid, Dow Chemical) in a closed heating loop. Syltherm HF can be used over a wide temperature range of -73oC (-100oF) to 260oC (500oF). This fluid flowed on the annulus side of the tube-in-tube heat exchanger in a counter current manner with the ammonia-rich solution in the tube side of the desorber. The heat transfer fluid was heated in a shell-and-tube steam heat exchanger (American Industrial Heat Transfer, Inc., Model URCS by) using steam on the shell side (Fig 11). The heat transfer fluid was circulated in the closed loop by a centrifugal pump (Model AC4 by Finish Thompson) that has a maximum pressure limit of 689 kPa (100 psi) and a temperature limit of 149oC (300oF). The heat transfer fluid was circulated in a 25.4 mm

12

(1) nominal OD SS-316L tube between the two heat exchangers to minimize the flow velocity and frictional pressure drop. Mass flow rate of the heat transfer fluid was measured using a Coriolis mass flow meter (model RFT9739-Field Mounted, Micromotion). Fluid temperatures were measured at the inlets and outlets of the two heat exchangers. With flow rate and temperatures known, the heat duties of the two heat exchangers in the loop can be calculated. The heat transfer fluid duty in the tube-in-tube desorber provided a means to check the accuracy of the results by comparison with the solution side duty. Fig 11. Steam Heat Exchanger In addition, a balance between the heat transfer fluid duties in the two exchangers (desorber and steam-to-Syltherm heat exchanger) further validated these results. The heat transfer fluid closed heating loop was removed after several of the initial tests, because the 149oC operating temperature limit for the pump restricts testing at the 5% and 15% concentrations, as higher desorber duties and temperatures are required for these concentrations. Thus, the steam line was directly connected to the tube-in-tube desorber, which removed the high temperature operation restrictions. Condenser and Condenser Coolant Loop A shell-and-tube exchanger (Model 00677-3, Exergy, Inc.) is used as the condenser (Fig 12). Ammonia vapor flows on the tube side, while coolant flows on the shell side of this heat exchanger. The 76.2 mm (3) O.D. shell contains 253, 50.8 cm (20) long tubes of 3.2 mm (0.125) nominal O.D. and 0.3 mm (0.0125) wall thickness for a total heat transfer area of 1.24 m2 (13.4 ft2). The maximum rated tube-side pressure is 5171 kPa (750 psi) at 427oC (800oF), while the shell-side rating is 3447 kPa (500 psi) at a maximum temperature of Temperatures of the 427oC (800oF). solution at the inlet and outlet of the condenser are measured using 3-wire RTDs and the absolute pressure is measured using absolute pressure transducers (Model 3051, Rosemount). Refrigerant mass flow rate is measured using a Coriolis mass flow meter (Model C25, Micro-motion), which can measure mass flow rates from 0 to 0.023 kg/s (3 lbm/min) and can be configured to different ranges. Measurement of temperature and pressure at the inlet and Fig 12. Photograph of the Condenser outlet along with the concentration at the

13

outlet of the rectifier provide a means to estimate the ammonia-side heat duty. A closed coolant loop is used to remove the heat of condensation that is eventually transferred to the lab chilled water-glycol solution in a plate heat exchanger (Superchanger UX-016-UJ21, by Tranter) with a total heat transfer area of 1.65 m2 (17.8 ft2). Fig 13 shows a photograph of this plate heat exchanger. In the initial design, the condenser coolant loop was combined with the evaporator heating loop so that the heat load for the evaporator could be Fig 13. Plate Heat Exchangers supplied by the condenser. This configuration minimized the required electric heater duty to supply the load for the evaporator. However, after some of the tests were conducted, the condenser coolant loop was separated from the evaporator heating loop to get precise and more independent control of coolant flow rates and temperatures. Much of the plumbing for this loop also consists of 25.4 mm (1) nominal OD tubes to maintain low pressure drops. The coolant is circulated using a centrifugal pump (TE-7MD-HC, Little Giant, with a 3/4 hp motor). An expansion tank is used upstream of the pump inlet to supply the requisite positive head for the pump and also to compensate for thermal expansion of the coolant. The pressure in the coolant loop is monitored using a dial pressure gage, which helps in the detection of any severe fluctuations or overshoot of pressure and prevention of potential failures. The coolant flow rate is measured by a Coriolis mass flow meter (Model RFT9739-Field Mounted, Micromotion). Temperature measurements at the inlets and outlets of the condenser enable calculation of heat duties in the respective sides. Coolant temperatures are measured using T-type thermocouples. Heat balances between the ammonia side and the coolant side of the condenser establish the accuracy of the results. Evaporator and Evaporator Heating Loop A tube-in-tube heat exchanger similar to the desorber (Model 00528 by Exergy Inc., Fig 10) is used as the evaporator. Ammonia solution flows in the inner tube while the heating fluid flows through the annulus. As mentioned above, the evaporator heating loop was initially combined with the condenser loop, but was subsequently separated into an individual, electrically heated loop to ensure precise control of the fluid flow rate and temperature. A 5kW electric cartridge heater by Watlow (Model L12NX4B) was installed into this loop upstream of the evaporator. The coolant is circulated using a centrifugal pump (Model AC4 by Finish Thompson, with a 3/4 hp motor, similar to the heat transfer fluid pump) through the mostly 25.4 mm (1) nominal OD plumbing in this loop. An expansion tank is also used in this loop as described for the other fluid loops above. Pressure in the coolant loop is monitored using a dial pressure gage. The flow rate of the coolant in the evaporator loop is measured using a positive displacement flow meter by AW company (Model JVM-60KL) coupled to a flow transmitter (Model FEM-03 by AW Company.) Temperature measurements using T-type thermocouples at the inlets and outlets of the evaporator enabled calculation of the heat duties. Heat duty

14

calculations on the ammonia side and the coolant side in the evaporator provided a means to validate the accuracy of the results. Separator The two phase mixture exiting the desorber is separated into individual vapor and liquid streams in the separator. The dilute solution flows to the absorber through the solution heat exchanger, while the ammonia vapor is routed to the rectifier. Fig 14 shows a photograph of the separator, which consists of an assembly of meshes and tubing inside a 0.3 m long, 0.15 m O. D. stainless steel flanged cylindrical shell. Blind flanges are bolted to the flanged cylinder with a Teflon gasket interface to form the enclosure. These blind flanges have ports for the various inlet and outlet streams. The two phase mixture enters the separator at center of the top flange through a 12.7 mm (0.5) diameter tube that is extended halfway down the length of the Fig 14. Photograph of the Separator separator where it is capped and 32 holes of 1.6 mm diameter are drilled along its length to decrease the velocity of the two phase mixture. A 0.15 m long, 50 mm diameter stainless steel pipe nipple with 56 holes of 3.2 mm diameter along its length results in additional expansion and separation. Inside this separator, the vapor rises towards the top due to buoyancy and the dilute solution drains to the bottom by gravity. An assembly of varying density meshes arranged concentrically around the central tube at the top and bottom further prevent any liquid from flowing through the top of the separator. A sight glass is installed in parallel with the separator to provide an indication of the liquid level in the separator. Temperatures were measured at the dilute solution and ammonia vapor outlets, and pressure was measured at the dilute solution outlet. Rectifier The rectifier reduces the water content in the vapor generated in the desorber so that the refrigerant does not undergo an undesirable temperature rise as it evaporates in the evaporator. An externally cooled rectifier provided by Rocky Research is used in this test facility. The rectifier (Fig 15) consists of a 48.3 cm (19) long, 11.4 cm (4.5) diameter shell that encloses a helical coil of about 12.7 mm (0.5) diameter, with appropriate fill material in the space between the coil and the shell. Vapor enters at the bottom of the rectifier and the purified vapor exits at the top. The external coolant (chilled water-glycol solution) enters at the top of the rectifier and exits at the bottom. The rectifier is placed at approximately 30o tilt from vertical since the vapor is condensed by the coolant line and condensate drops fall on the cylindrical wall of the rectifier and collect at the bottom end. The reflux exit is located at lowest point of the intersection of cylindrical wall and end cap, thus preventing any vapor from exiting due to the liquid pool collected at this location. Temperatures are measured at the ammonia vapor inlet and at the 15

outlets of the purified ammonia vapor and reflux; the pressure was measured at the reflux outlet. External coolant temperatures are measured using T-type thermocouples at the inlet and outlet, with the flow rate being measured by a rotameter. The high concentration refrigerant vapor exiting the rectifier flows to the condenser and the liquid reflux is mixed with dilute solution exiting the separator. Solution Heat Exchanger The solution heat exchanger preheats the concentrated solution flowing to the desorber by cooling the dilute solution exiting the separator. This heat exchanger reduces the desorber heat duty requirement by preheating the concentrated solution, and the absorber heat duty by reducing the dilute solution temperature entering the absorber. A shell and tube exchanger (Model 00256-3 by Exergy Inc) is used as the solution heat exchanger, as shown in Fig 16. The 38.1 mm (1.5) O.D. shell contains 55, 38.1 cm (15) Fig 15. Photograph of the Rectifier long tubes of 3.2 mm (0.125) nominal OD and 0.3 mm (0.0125) wall thickness, for a total heat transfer area of 0.20 m2 (2.2 ft2). The maximum rated tube and shell side pressures are 8274 and 5516 kPa (1200 and 800 psi), respectively, at a maximum temperature of 427oC (800oF). Temperatures are measured at the inlet and outlet for both fluids and used to establish a heat balance across this component. Pre-Cooler In the pre-cooler, vapor exiting the evaporator is used to cool refrigerant liquid exiting the condenser before it enters the evaporator. The pre-cooler also recuperatively heats the vapor entering the absorber. The pre-cooler reduces the enthalpy of the condensate at the evaporator inlet, which increases the cooling capacity of the evaporator. Since the pre-cooler preheats the refrigerant vapor entering the absorber, the absorber duty is increased. The pre-cooler also causes an additional pressure drop that leads to a high evaporator pressure or lower heat rejection temperature for the absorber, or low mass flow rate of the refrigerant. However, another important Fig 16. Solution Heat Exchanger function the pre-cooler serves is that it 16

handles the evaporation in the high quality, high temperature glide region, thereby maintaining a relatively constant temperature in the evaporator, which is desirable for ensuring that adequate driving temperature differences between the cooled fluid and the refrigerant are maintained. A shell and tube exchanger (Model 00256-2 by Exergy Inc.) is used as the pre-cooler, as shown in Fig 17. The 38.1 mm (1.5) O.D. shell contains 55, 25.4 cm (10) long tubes of 3.2 mm (0.125) nominal O.D. and a 0.3 mm (0.0125) wall thickness for a heat transfer area of 0.13 m2 (1.43 ft2). The Fig 17. Photograph of the Pre-Cooler maximum rated tube side and shell side pressures are 8274 and 5516 kPa (1200 and 800 psi), respectively, at a maximum temperature of 427oC (800oF). Temperatures are measured at the inlet and outlet of the refrigerant liquid and vapor streams of this heat exchanger and energy balances established. Expansion Valve A constant superheat thermostatic expansion valve (Model #TEA 20-2 by Danfoss) was first used as an isenthalpic expander for the refrigerant ammonia between the pre-cooler and the evaporator. This valve could be used for evaporator temperatures of -20 to 30oC with up to 7 kW cooling duties, which is the anticipated range of evaporator temperatures for the pressures of interest in this study. However, this expansion valve did not work at intermediate cooling capacities, and was suspected to be oversized for the tests under consideration. Thus, to account for the variable cooling capacity at the evaporator, a manual metering needle valve (SS-4MGMH, max CV of 0.03, Swagelok) was installed in parallel with the existing thermostatic expansion valve. By installing two additional shut-off valves, it was possible to use any of these two expansion valves one at a time. Pressure Reduction Valve The system was initially designed with a pressure reduction valve to achieve the required pressure reduction from the high side pressure to the low side pressure. Two pressure reduction valves were used, one for the dilute solution between the solution heat exchanger and the absorber, and the other for the reflux flow from the rectifier. Suitable pressure reduction valves for ammonia-water solution (Model A2BO V-series, by Parker) were installed. The pressure range of the valve is between 66.7 kPa (500 mm Hg) to 800 kPa (8 bar) with 180 kPa (1.8 bar) change per turn, and a maximum operating pressure of 2070 kPa (20.7 bar). These valves are outlet-set pressure regulators that also contain stainless steel strainers to protect the valve from foreign material present in the system. The absorber pressure was expected to be maintained at the desired test condition by appropriate settings of this valve. Use of the tube-in-tube co-flow heat exchanger for the desorber prevents the reflux from entering the desorber. Therefore, using the above-mentioned pressure regulator, it was decided that the reflux could be introduced in the concentrated solution stream from the absorber before it enters the solution pump. However, these pressure reduction valves proved to be ineffective in controlling the pressure precisely and 17

were replaced with manually controlled needle valves to achieve precise control of solution flow rates and the low-side pressure. Solution Pump The solution pump is a variable-speed Tuthill magnetic gear pump with 1.4410-4 m3/s max flow rate at zero pressure difference. This pump operates between 1034 kPa (150 psi) pressure difference for continuous duty and 1724 kPa (250 psi) for intermittent duty at a maximum system pressure of 3447 kPa (500 psi). The operating temperature is between -46oC and 176oC, and the maximum operational viscosity is 2 kg/m-s (2000 cps). Instrumentation Temperature and Pressure Measurement Temperatures are measured using 3-wire RTDs and T-type thermocouples. T-type thermocouple (Omega Engineering) can be used over the range -270 - 400oC temperature range with an accuracy of 0.5oC while 3-wire RTDs (Pr-13, Omega Engineering) can be used to measure temperatures in the range -60 400oC with an accuracy of 0.8oC. Absolute pressures are measured using Rosemount model 2088 and 3051 pressure transducers. The absolute pressure transducers are capable of measuring pressures in the range 0 kPa to 27579 kPa (4,000 psia), with an accuracy of 0.075% of the calibrated span. However, for improving accuracy, the span was set to 0 kPa to 2048.5 kPa (300 psi). Differential pressures were measured using a Rosemount model 3051 transducer that is capable of measuring pressures in the range of 13790 kPa, with an accuracy of 0.075% of span for spans larger than 2758 kPa. For smaller spans, the accuracy of the transducer is [0.025 + 10/span] % of span. These pressure transducers produce 4-20 mA output signals that are converted to 1-5 V signals for use in the data acquisition system. The pressure transducers are rated for the temperature range of 40 - 121oC. Flow Meters To measure the mass flow rates of the dilute and concentrated solutions, Micromotion Coriolis flow meters (model CMF025 ELITE) are installed in the test facility. The flow rates are displayed on Micromotion RFT 9739 rack-mounted displays. These meters can be configured for the desired ranges of the parameters using the HART communicator. The flow meters can read 0-0.605 kg/s or 0- 6.30910-4 m3/s with an accuracy of 0.10% of the flow range. The operating temperature range is -240 - 204oC, with a maximum operating pressure of 10000 kPa (100 bar). The meters provide one primary and one secondary analog output and one frequency output. The flow meters generate 4-20 mA output signals that are converted to 1-5 volt signals for use by the data acquisition system. The heat transfer fluid flow rate, and subsequently, the coolant flow rate in the condenser loop, was measured using the Coriolis flow meter by Micromotion (model CMF100 Elite). The flow meter can measure flow rates of 0 - 7.56 kg/s (1000 lbm/min) with an accuracy of 0.10% and density in the range 0-5000 kg/m3 with an accuracy of 0.5 kg/m3. It can be used over a wide range of temperatures -240 204oC up to a maximum system pressure of 100 bar (10000 kPa or 1450 psi). The meter reading is transmitted using a field mount transmitter (RFT9739 by Micromotion).

18

The flow rate of the absorber coolant is measured using a magnetic flow meter by Rosemount (model 8711) coupled to a model 8712C flow transmitter, with an accuracy of 0.5%, and can be used over the range -29 149oC up to system pressures of 5102 kPa (740 psi). The flow rate of the coolant in the evaporator loop is measured using a positive displacement flow meter by AW company (model JVM-60KL) coupled to a flow transmitter (FEM-03 by AW Company). It can be used to measure flow rates in the range 0-12.61810-4 m3/s with 0.5% accuracy up to 34474 kPa (5000 psi) system pressure. Chilled water-glycol flow rates are measured using rotameters for the different cooling loops. The flow rate in the plate heat exchanger of the absorber coolant loop is measured using a VFC-143 series rotameter by Dwyer over the range 1.261810-4-12.61810-4 m3/s gpm with an accuracy of 2% of full scale. The flow rate in the plate heat exchanger of the condenser coolant loop is measured using a 7530-7C08 rotameter by King Instrument over the range 0.630910-4 - 6.30910-4 m3/s with an accuracy of 0.5% of full scale. There are two rotameters in the rectifier loop. One of the two is used depending upon the flow rate for the data point under consideration. One of the two (model VFB-85-EC) has a range of 0.12610-4-1.26210-4 m3/s, while the other (model VFB-86-EC) has a range of 0.31510-4 3.15510-4 m3/s. Both of these have an accuracy of 3% of full scale. A summary of the specifications of each instrument used in this project is provided in Table 2. Data Acquisition System A PC-based data acquisition system (supplied by IO Tech) is used to display and record data during the tests. The Tempscan/1000A with expansion unit EXP/11A interfaced with the computer through the program TempView 4.1 allows real-time display and recording of temperatures, pressures, and flow rates. Together with an expansion chassis, the Temp Scan/1100 enables the capacity to add up to 10 modules that could monitor up to 992 input channels. Data acquisition rates and durations can be programmed at desired scan rate using the TempView interface. Pressure transducers are connected to the Tempscan using a voltage scanning module (TempV/32B) that converts 4-20 mA current input to 1-5 V using 250 ohm resistors, while thermocouples and RTDs are directly connected to Tempscan through thermocouple scanning modules (TempTC/32B) and RTD scanning modules (TempRTD/16B). The thermocouple scanning module (TempTC/32B) can read 32 differential input channels, which can be configured for thermocouple types J, K, T, E, R, S, B, and N or for a 100mV input. Measurements can be designated in units of oC, oF, K, R, or volts. The RTD scanning module (TempRTD/16B) can read 16 channels of 3- or 4-wire RTDs in units of oC, oF, K, or R. The voltage scanning module (TempV/32B) can read 32 input channels with programmable ranges of 10 V, 5 V, 1 V, or 100 mV. A Windows-based TempView program provides a graphical user interface for easy configurations of hardware, acquisition, and display parameters. Real time data can be monitored by the TempView program in graphical or spread-sheet style. For the current tests, a total of 4 scanning modules, two (TempTC/32B) for T-type thermocouples, one (TempRTD/16B) for RTDs, and one (TempV/32B) for pressure transducers as well as for flow meters that have 4-20 mA output, are used. For this project, 44 thermocouples, 12 RTDs, 11 absolute and 1 differential pressure transducers, and 6 flow meters are connected to Tempscan using these 4 scanning modules. For each data point therefore, 56 temperatures, 12 pressures, 6 flow rates and 2 densities are recorded over a span of 5 minutes at 3 seconds intervals yielding 100 readings.

19

Table 2. Instrumentation Specifications Summary

Parameter/ Instrument Model Manufacturer Quantity Range Accuracy Operating Limits Flow Meters and Transmitters Specifications

Dil/Conc Solution Dil/Conc (Trans) Refrigerant Abs Coolant Abs Coolant (Trans) Evap Coolant Evap Coolant (Trans) Cond Coolant Cond Coolant (Trans)

CMF025 Elite RFT9739Rack Mounting C25 8711 8712C JVM-60KL FEM-03 CMF100 Elite RFT9739Field Mounting

MicroMotion Inc.

MicroMotion Inc. MicroMotion Inc. Rosemount Rosemount AW Company AW Company MicroMotion Inc. MicroMotion Inc.

2 1 1 1 1 1 1 1

0 - 0.6048 kg/s 0 - 6.309 10-4 m3/s 0-5000 kg/m3 -240-450oC 0 - 0.02268 kg/s 0 - 1.262 10-3 m3/s 1.262 10-4 - 1.262 10-3 m3/s

0.1% (Flow) 0.5 kg/m3 (Density)

T: -240 to 204oC P: 10000 kPa (1450psi) T: 0 to 50oC (Ambient)

0.1% 0.5% T: -29 to 149oC P: 5.1MPa (740psi) T: -29 to 60oC (Ambient) P: 34474 kPa

0.5% @ = 3 10-5 m2/s

0-7.56 kg/s 0-5000 kg/m3 -240-450oC

0.1% (Flow) 0.5 kg/m3 (Density)

T: -240 to 204oC P: 10000 kPa (1450psi) T: 0 to 50oC (Ambient)

Pressure Transducers and Dial Gauges Specifications

Absolute Absolute Absolute Differential Dial Gauge Dial Gauge

2088 3051 3051 3051C 1005P (ABS-Black) 1005P (ABS-Black)

Rosemount Rosemount Rosemount Rosemount Ashcroft Ashcroft

5 2 4 1 2 1

0-5516 kPa 0-27579 kPa 0-68948 kPa 0-13790 kPa 0-689.5 kPa 0-2068 kPa

0.25% of Calibrated Span 0.075% of Calibrated Span 0.075% of Calibrated Span 0.075% of Calibrated Span 3% of Span 3% of Span

T: -40 to121?C T: -40 to121?C T: -40 to121?C T: -40 to 121C P: 31MPa (4500psig) T: -40 to 65oC T: -40 to 65oC

Thermocouples and RTDs Specifications

Thermocouple RTD

T-type Pr-13

Omega Engineering Omega Engineering

44 12

-270-240oC -60-240oC

0.5oC 0.5-0.8oC

Rotameters Specifications (m3/s)

Absorber Condenser Rectifier Rectifier Sub-cooler

VFC-143 series 7530 7C-08 VFB -85-EC VFB -86-EC 7530 7C-06

Dwyer Instrumentation King Instrument Company Dwyer Instrumentation Dwyer Instrumentation King Instrument Company

1 1 1 1 1

1.262 10-41.262 10-3 6.309 10-56.309 10-4 1.262 10-41.262 10-3 3.155 10-53.155 10-4 2.524 10-53.155 10-4

2% of Full Scale 2% of Full Scale 3% of Full Scale 3% of Full Scale 2% of Full Scale

T: 49oC P:690kPa (100psi) T: 54oC P:862kPa(125psi) T: 65oC P:690kPa (100psi) T: 65oC P:690kPa(100psi) T: 54oC P:862kPa (125psi)

20

Tailoring the Test Facility for Wide Operating Range During this project, after the test facility became functional and a regular program of testing was initiated, the facility underwent major modifications in different stages to enable testing at vastly different test conditions. These modifications included changes to the myriad of coolant loops for each component including new heat exchangers and plumbing orientations, changes to ammonia (e.g. reflux flow rate) and coupling fluid measurement techniques, changes to control devices such as refrigerant expansion valves, changes to the heat source supply method, and numerous other minor modifications that provided the ability to progressively conduct tests and measurements over an extremely large range of high and low temperature, pressure and concentration, and flow rate conditions. It should be noted that these changes were most often not because of inadequate design or operation, but simply because no single test facility configuration can accomplish testing at all the conditions required in this project. The most significant of these changes and different test facility configurations are described briefly in the flowing sections. Test Facility, Version I (Initial Approach to Stable Operation) This section describes the first version of the test facility. Figs 18 and 19 show the schematic and photograph of the first version of the test facility, respectively. The initial design of the test facility was based on the following parameters: Evaporator cooling duties of up to 5 kW Condenser temperature in the range of 30-40oC Absorber pressures of 150 kPa (21.5 psi), 345 kPa (50.0 psi) and 500 kPa (72.5 psi) Dilute solution concentrations at the desorber of 5%, 15%, 25% and 40%.

A simulation program was developed to estimate system performance and various heat exchangers and pump sizes. The program used initial design parameters, basic principles of absorption cycle analysis such as mass, species and enthalpy balances, and assumptions regarding effectivenesses of heat exchangers. Estimates from these preliminary analyses provided guidelines for selecting and designing different components and flow rates. The test

Fig 18. Test Facility Schematic, Version I 21

Rectifier

Separator Condenser

Desorber

Evaporator Absorber

Fig 19. Photograph of the First Version of the Test Facility facility layout with dimensions of 274 x 137 x 198 cm (9 x 4.5 x 6.5 ft) was designed and fabricated to accommodate the system size and complexity of the experiments. The system fabricated in the first phase consisting of numerous heat exchangers functioning as various components of an absorption system, pumps and extensive instrumentation can be subdivided into two broad categories: solution loop and auxiliary loops. The main loop consists of the absorber, solution heat exchanger, desorber, separator, rectifier, condenser, pre-cooler and the evaporator. The rectifier is directly coupled to its cooling source, while the absorber, desorber, condenser and evaporator are coupled through intermediate loops. Steam is the ultimate heat source for the desorber. A chilled water-glycol supply from the Carrier AquaSnap 50 RT chiller is the ultimate heat sink for heat removal from the absorber and the condenser. Use of intermediate loops provides improved control over test conditions and heat duties. The components are located with consideration of their functions and ease of operation and maintenance. The desorber, separator, rectifier and condenser were mounted as high as possible on the basic frame, while the evaporator and absorber were mounted as low as possible to enhance the opportunities for providing static pressure head between the high and low pressure sides of the loop as necessary. After the test facility fabrication was completed, the system was checked for leaks. Once it was ensured that there are no leaks in the system, the main loop was evacuated and charged with the required quantities of ammonia and water to obtain the desired charge and ammoniawater concentration in the loop. Shake-down tests for each version of the facility were initiated

22

after this stage. It should be noted that for every subsequent modification of the solution side of the loop, ammonia-water solution was discharged. After the modifications, the system underwent the standard procedure of leak-checking, evacuation and re-charging. Initial tests attempted steady operation while maintaining the high and low pressure levels in the loop at the desired desorber outlet conditions. In the beginning, these objectives could not be obtained, and it was determined that the thermostatic expansion valve did not provide the required external pressure equalization. The external pressure equalizer was installed and connected to the thermostatic expansion valve from a location downstream of the evaporator. In addition, a valve was installed between the evaporator and absorber to provide isolation between different segments of the loop and flexibility for maintenance operations. A 6.4 mm (1/4) liquid line was also installed between the evaporator inlet and the pump inlet, which, if necessary, allowed removal of waterrich liquid accumulated in the evaporator to be mixed with the solution from the absorber outlet. These modifications allowed two-pressure operation. The pressure reduction valve was able to control the flow, but maintaining the pressure level posed challenges because the two pressure regulators (in the reflux line, and in the dilute solution return line to the absorber) operating between relatively similar pressures were interfering with each other, making stable operation difficult. It was also observed that at very small solution flow rates, solution returning to the absorber was flowing along the inner wall of the absorber shell. This was attributed to the previous absorber inlet design, in which the inlet line ended abruptly at the shell, allowing a portion of the solution flow along the shell wall rather than over the tube bank, particularly at the low flow rate cases. With a portion of the absorption occurring on the shell without adequate heat removal, the absorber temperature and pressure were unduly high. It was also found that often, the solution pump inlet was starved of solution, primarily due to unstable running conditions. In such instances, the pressure in the expansion tank upstream of the solution pump rose to levels higher than that in the absorber, thus preventing solution flow. These issues were addressed as follows. The 12.7 mm () inlet tubing to the absorber was extended into the absorber shell half way through the depth of the drip tray to make the solution fall on the tubes and not along the wall. To avoid pressure spikes in the expansion tank, a pressure equalization line was installed to connect the top of the sight glass and expansion tank to the absorber. Thus, when solution at the pump inlet was lost, the expansion tank could be filled soon. With these modifications implemented, two pressure levels and desired temperatures at different locations, and steady solution flow was obtained. However, since the pressure reduction valve was an outlet-set pressure regulator, whenever the pressure at the absorber, which is downstream of the regulator in the solution line, was increased, the regulator interrupted the flow. This problem was circumvented by leaving the regulator fully open, and maintaining the downstream pressure with 12.7 mm () ball valves. Mixing between the rectifier reflux and the concentrated solution line was found to cause pressure rise and unsteady operation. Therefore, for these initial tests, the reflux line was not used, essentially bypassing the rectifier function. With this test facility configuration, cooling duties of up to 5 kW (17000 Btu/h) with a dilute solution concentration of 30.5% and a concentrated solution concentration of 41.1% were achieved. A preliminary film absorption heat transfer coefficient of approximately 700 W/m2-K was obtained.

23

Test Facility (Version II) After successful and steady operation of the absorption chiller facility was accomplished, the insights obtained were used to design further modifications that would enable such operation over a much larger range of conditions with accurate energy balances for all the components around the test facility. Thus, since it was found that twopressure operation was obtained even with very coarse control offered by a 12.7 mm () ball valve, it was realized that the pressure reduction valve in the dilute solution line to the absorber was unnecessary. A manually controlled needle valve with a variable CV (flow coefficient) was therefore used to replace the pressure reduction valve and provide much better control over the flow rates and pressures. The pressure reduction valve in the reflux line, which was causing unstable operation, was also removed and the reflux was directly introduced at the separator outlet. Fig 20. Test Facility Schematic, Version II To enable this modification, the rectifier was further elevated to increase the static head available for drainage of the reflux. It was stated above that the condenser and evaporator were initially coupled to a common fluid loop. During testing it was observed that the condenser duty required coolant flow rates that were too large for the corresponding evaporator duty, and resulted in very small temperature difference in the evaporator, with correspondingly large uncertainties. To provide independent and different coupling fluid flow rates for the evaporator and condenser, a bypass line and a new flow meter were installed to control and measure the evaporator coupling fluid flow rate separately. It was also stated above that the circulation pump for the heat transfer fluid in the desorber loop had a maximum temperature limit of 149oC. This restricted high desorber temperature operation, and did not allow testing at the lower concentration cases. To mitigate this problem to some extent, a tube-in-tube heat exchanger was fabricated using 61 cm (2 ft) long 25.4 mm (1) and 12.7 () O.D. tubes and installed between the desorber heating fluid outlet and the pump inlet. This heat exchanger reduced the heating fluid temperature at the pump inlet by 10-15oC, allowing the testing of a few additional operating conditions. Thus, this second version instituted the following modifications, as summarized in Figs 20 and 21: Removal of pressure reduction valves 24

Rectifier

Separator Condenser

Desorber

Evaporator Absorber

Fig 21. Test Facility Photograph, Version II Replacing pressure regulator with a needle valve Repositioning the rectifier for additional gravity head and introduction of reflux at the separator outlet Independent condenser/evaporator coupling fluid flow rates through bypass lines Addition of heat exchanger to pre-cool the heating fluid in the desorber heating loop upstream of the pump and extend data point range

With this revised facility, tests were conducted at absorber pressures of 345 kPa (50.0 psi), 258.5 kPa (37.5 psi) and 500kPa (72.5 psi) at a variety of solution flow rates and concentrations to confirm operation over a wide range of conditions. These tests were used to establish system characteristics such as charge, time constants, controllability, and other hardware-related parameters to guide the systematic tests over the entire test matrix. For the range of flow rates tested, evaporator cooling duties from 1 to 3 kW were obtained at evaporator temperatures of -4 to -15.7oC. In general, after the long transient (several hours) after initiation of testing on any given day, it took about 2-3 hours to achieve steady state and record an individual data point. Preliminary analyses were then conducted to obtain overall heat balances and heat transfer coefficients. The various concentrations in the loop were deduced based on pressure and temperature measurements at several locations in the loop and assumptions regarding saturated 25

states at specific locations. Agreement between concentrations calculated from multiple state calculations around different components indicated the accuracy of the data for each test. Also, mass, species and energy balances at the appropriate locations provided multiple redundant methods of computing state points around the loop and for each component. These analyses indicated that while the test facility was able to run over a wide range of conditions in a stable fashion, energy balances between the different fluid streams of components were not as good as necessary to be deemed acceptable. These discrepancies were attributed to the high sensitivity of the vapor phase concentration to the measured temperatures and pressures an inherent property of ammonia-water mixtures. Also, the typical reflux flow rate generated in this system, due to the overall solution flow rates under consideration, was very small, and tended to drain intermittently, to some extent. Reflux and dilute solution mixing calculations at the desorber inlet were therefore subject to some variations due to the inherently unsteady drainage of reflux, leading to some of the discrepancy between mass, species and energy balances. Over the extremes of the test matrix, the coupling fluid flow rates chosen as required to establish either the low-side or high-side pressure were either too low, leading to inaccuracies in flow rate measurement, or too high, leading to fluid Ts too small to be measured with the desired accuracy. Once again, these are problems inherent in attempting accurate heat and mass transfer tests with the objective of determining component heat transfer coefficients while being constrained at the high side and low side by system operating condition requirements. (This is the primary reason why the original proposal specified that tests would be conducted on a singlepressure component test facility.) With stable operation controllable over a wide range of conditions achieved, the third version of the test facility focused on instituting additional measures to improve accuracies in mass, species and energy balances. Test Facility (Version III) Based on the discussion above, the heat transfer fluid closed loop for the desorber was removed because of the persistent pump operating temperature problem (T limited to < 149oC) and the associated restriction on low concentration test conditions. Steam from the lab boiler was plumbed directly to the desorber, which removed the high temperature limitation. It should be noted, however, that this resulted in the loss of one additional method to confirm energy balances around the desorber, but this was deemed acceptable because there was still a sufficient number of methods to deduce ammonia-water conditions and duties around the absorber. The ability to run the low concentration cases assumed priority in this issue. In the versions of the test facility discussed thus far, the refrigerant flow rate was not measured directly, but was obtained by subtracting the measured dilute solution flow rate from the measured concentrated solution flow rate. It was deemed that because the refrigerant flow rate was a particularly critical quantity for the measurement of absorption rates, a direct measurement of this flow rate should be conducted rather than inferring it from the differences of two flow rates or indirectly from other quantities such as the evaporator or condenser coupling side heat duty. Since measuring refrigerant flow rate in vapor form is quite challenging, based on the anticipation that the refrigerant exiting the condenser is subcooled, a Coriolis mass flow meter was installed downstream of the condenser to measure the liquid-phase refrigerant mass flow rate. This also necessitated the verification of the presence of a liquid phase at the outlet of the condenser; therefore, a sight glass was installed at the condenser outlet. In addition, the mass flow meter in the reflux line was replaced with a volumetric flow meter with a smaller flow rate 26

range, in view of the low reflux flow rates. With these modifications completed, tests were initiated again. It was now possible to make a visual confirmation of the exit state of the condenser. However it was found (by observing the sight glass) that in some cases, condenser subcooling was not being achieved, although subcooling was seen for cases with a high desorber heat input and a high vapor flow rate. Based on these observations, it was deduced that this was occurring because the refrigerant expansion valve oversized for the refrigerant flow rates under consideration. It was also clear that the wide range of refrigerant flow rates anticipated in the overall test matrix were best suited for expansion across a manually controllable variable orifice, as it was likely that a standard thermostatic expansion valve acting alone would either be undersized or over-sized for many of the test conditions. A metering needle valve (SS-4MG-MH, maximum CV of 0.03, by Swagelok) was therefore installed in parallel with the existing thermostatic expansion valve. Shut-off valves were also installed upstream of each valve to provide the flexibility to use whichever valve was appropriate (or a combination) for a given test case. Similarly, the needle valve in the dilute solution line upstream of the absorber was also found to be oversized, and was replaced with a smaller CV valve. Among other modifications, a sight glass was also installed in the reflux line to enable confirmation of the phase of the fluid in this line, and detect any potential vapor slugging. The reflux flow meter was found to be ineffective, while also adding to the flow resistance in this line, and was therefore removed. The reflux flow rate is now obtained from a mass and species balance at the rectifier. After finishing these modifications, tests were re-started, and all the tests were repeated with the modified test facility. The facility operated well, and with various amounts of charge in the loop, tests were conducted at all three pressures, 150 kPa (21.8 psi), 345 kPa (50.0 psi) and 500 kPa (72.5 psi) for three different concentrated solution flow rates, nominally 0.019, 0.026, and 0.034 kg/s (2.5, 3.5 and 4.5 lbm/min), for 5%, 15%, 25% and 40% dilute solution concentration at the desorber. These tests yielded acceptable mass, species, and energy balances for all components in the loop. It was, however, found that the 150 kPa (21.8 psi) absorber pressure cases for 25%, and 345 (50.0 psi) and 150 kPa (21.8 psi) absorber pressure cases for 40% dilute solution concentration at the desorber could not be obtained using this test configuration. For example, despite using the absorber coolant and the chilled water-glycol supply at their full capacities, it was not possible to obtain the lower pressures at the higher dilute solution concentrations. Therefore, it was decided that pre-cooling the absorber inlet solution over and above what was achievable in the solution heat exchanger would decrease the heat load that must be handled by the absorber in some of these higher dilute solution concentrations. Such a provision would offer independent control of the dilute solution temperature at the absorber inlet, and also lower the sensible cooling component of the absorber heat load, making the latent heat load dominant. High fractions of sensible heat loads in the absorber could unduly overestimate the absorption heat transfer coefficient. Based on these considerations, a 61.0 cm (2 ft) long water-cooled, tube-in-tube dilute solution subcooler (Fig 22) was designed, fabricated and installed between the solution heat exchanger and the absorber. This stainless steel heat exchanger consists of a 12.7 mm (0.5) nominal O.D. inner tube and a 25.4 mm (1) nominal O.D. outer tube. Dilute solution flows through the inner tube, while chilled water-glycol solution flows through the annulus in a counterflow orientation. It was also found that for some of the data at the 345 and 500 kPa (50 and 72.5 psi) absorber pressure cases, there were large temperature differences from the inlet to the outlet of the absorber coolant due to the relatively small absorber coolant flow rates that were required to maintain these specific conditions in the first place. Since coolant flow rate measurement errors 27

would be relatively large at such low required coolant flow rates, and also since large coolant Ts could affect absorption characteristics, it was decided to find an alternate means to reduce absorber coolant Ts while maintaining reasonably large coolant flow rates. The heat sink for the closed coolant loop of the absorber is the large plate heat exchanger, which is coupled to the common glycol-water solution coolant from the chiller. This heat sink cools the closed-loop coolant to temperatures very close to the chiller temperature, restricting the amount of control Fig 22. Dilute Solution SubCooler available over the absorber coolant inlet temperature. In view of these considerations, it was decided to use a high flow rate closed loop coolant for these test cases, but raise the temperature at the absorber inlet by introducing a 5 kW resistance heater between the plate heat exchanger and the absorber inlet. This heater therefore enabled the use of reasonably large (and measurable) coolant flow rates while reducing the coolant temperature rise within the absorber, and added considerable flexibility to the control of the absorber conditions. Figs 23 and 24 show a schematic and a photograph of the third (and current final) version of the test facility, respectively. The changes made to achieve this configuration are summarized below: Direct steam heating of the desorber without the closed heat transfer fluid loop Installation of a smaller CV needle valve as the expansion device in the refrigerant circuit Addition of refrigerant mass flow meter at the condenser outlet Addition of sight glasses at the condenser and reflux outlets Replacing needle valve in the dilute solution line with a smaller CV needle valve Installation of a dilute solution subcooler between the solution heat exchanger and absorber Installation of a 5 kW resistance heater in the absorber coolant loop Decoupling of the condenser and evaporator fluid loops to achieve independent control The next section describes the procedures used to conduct the experiments. Experimental Procedures Safety Precautions Safety issues constitute a significant priority when working with chemicals that are harmful to humans. Ammonia has toxic and corrosive characteristics and causes freeze burns from its cold temperatures, and may affect the immune system in the case of large acute exposures. Ammonia vapor is toxic, corrosive, depletes oxygen, and acts as a poison when 28

ingested. Whenever eyes are exposed to ammonia, immediate first aid is necessary. To avoid any possible mishaps, many safety and precautionary measures are taken throughout this project. Some of these important test practices are described here. A full face mask fitted with an ammonia/methylamine cartridge and gloves are worn whenever there is a possibility of direct contact with ammonia, especially during charging and discharging of the loop with ammonia. The system is located under a fume hood and also enclosed with a plastic curtain on all sides. An exhaust fan in the fume hood is constantly turned on to trap and discharge any ammonia that might be present during charging and operation, or during plumbing modifications to the outside atmosphere. An ammonia monitor that is functional all the time is installed on the wall next to the test facility. This monitor (SAM by Lumidor Safety Products) has a Fig 23. Test Facility Schematic, Version III (Current) working range of 0-99 ppm and an accuracy of 2 ppm, and is calibrated to provide a loud alarm at 25 ppm of ammonia in its vicinity. Operators also wear protective safety glasses while operating the system. Leak Testing and Charging Before charging the system, it must be ensured that there are no leaks in the system. Extensive leak tests are performed before charging the system. The system is first charged with air at up to 793 kPa (115 psi), and fittings, gaskets, welded parts and sight glasses are checked for leaks using soap detection. In several instances, in addition to soap detection, the system is charged with air and R-134a up to 1379 kPa (200 psi) and a refrigerant leak detector (CPS model L-709) is used to detect leaks. The system is then kept pressurized overnight. If there is no appreciable pressure drop, the evacuation process is started using a vacuum pump (model DV85N by DV industries) and maintained until the pressure in the system decreases to 20 Pa or lower. A vacuum gage (model 14571 by Thermal Engineering) is used to monitor the vacuum level, and pressure transducers are also used to run a pressure trace. This evacuation process removes non condensable gases, and also provides enough pressure difference to enable charging 29

Fig 24. Test Facility Photograph, Version III (Current) with water and ammonia. Distilled water is charged before charging ammonia so that when ammonia is charged, it absorbs into water and the pressure can be maintained in a controlled manner. The total required charge for the system is estimated using expected void fractions and concentrations in the various components and plumbing of the loop at nominal test conditions. Because test conditions vary over a wide range of concentrations, the amount of charge is determined based on the conditions to be tested in a given period. The mass fraction of ammonia in the loop is generally kept within 15% of the desired test concentration to keep the desorber heat duties within a reasonable range. More ammonia is added whenever higher concentration cases are to be tested. Testing A set of well established test procedures was developed and used to ensure that data were obtained with acceptable accuracies. The tests were initiated after appropriate amounts of ammonia and distilled water were charged into the system. The data acquisition system and Tempview software are started first so that the progress of the changing states in the system can be monitored continuously. Before starting a test, the system is checked for leaks, and valves that need to be in the open mode are opened. It is also ensured that there is enough liquid level at the pump inlet. The laboratory chiller is turned on and is set to the appropriate temperature. Since the desired dilute solution concentration depends on the pressure and temperature at the desorber outlet, and the desorber pressure depends on the condenser temperature, the laboratory chiller temperature is first estimated to obtain the desired condensing temperature. This

30

laboratory chiller provides chilled water-glycol solution to serve as the heat sink for the rectifier, condenser and absorber. Once the laboratory chiller is turned on, condenser, absorber, and evaporator coolant pumps are started. After all coolant pumps are turned on and it is determined that they are running at steady state, the variable speed solution pump is started at a small frequency to yield a low solution flow rate. Since there is no pressure differential across the pump at the beginning, the status of the pump is carefully monitored based on the solution level using the sight level indicator that is connected to the expansion tank upstream of the solution pump. After the solution pump is started, steam supply to the desorber is turned on. Addition of steam starts heating up the solution from the pump and raising the high side pressure. The heat duty is increased in gradual steps. Expansion valves for the dilute solution and refrigerant are kept wide open to reduce the flow restriction before some steady circulation is achieved. The solution level in the separator is also checked to ensure that there is no significant accumulation and therefore to avoid starving the pump of solution supply. The solution pump speed is then gradually increased while the steam supply is also increased according to the solution flow rate. As steam supply is increased, the desorber outlet temperature and pressure increase; however, the solution flow rate decreases since the differential pressure is now higher. It takes sometime before vapor actually starts flowing out of the rectifier and condenses in the condenser. As the amount of vapor generated at the desorber and flowing to the condenser is increased, the temperature at the evaporator starts decreasing. The heater in the evaporator heating loop is turned on when the evaporator heating fluid temperature starts decreasing. Steam supply to the desorber and the solution flow rate are adjusted to get the desired concentrations at the desorber outlet, and solution flow and rectifier coolant flow are adjusted to get the desired refrigerant concentration. When system is running in a quasi-steady state, the evaporator expansion valve is gradually closed to achieve a better pressure separation and therefore obtain a liquid phase at the condenser outlet. Sometimes, this increases the high side pressure and decreases the low side pressure, resulting in increased differential pressure and decreased solution flow rate. The decreased solution flow rate is compensated by increasing the pump speed accompanied by an adjustment of the steam supply. If the high side pressure becomes too high, first, the laboratory chiller flow rate for the condenser loop is increased and second, the rectifier flow rate is increased. The differential pressure is monitored to keep it within 1034 kPa (150 psi), which is the maximum operational differential pressure for the pump. Throughout this process, the dilute solution flow rate is controlled using the solution expansion valve in the dilute solution line, and the difference between the concentrated and dilute solution flow rates is kept close to the refrigerant flow rate measured at the condenser outlet. The absorber pressure is continuously varying during this startup process, and is controlled by varying the absorber coolant flow rates and temperatures. Once the desired concentration at the desorber outlet, the concentrated solution flow rate, and the absorber pressure are obtained, the system is observed for some time without changing any of the parameters. If the values remain fairly steady, a preliminary data point is recorded and energy balances and concentrations are checked at various positions. If any of the heat balances at the different components are unacceptable, the coolant side flow rate or temperature or both are varied to improve the agreement between the ammonia-water and coolant-side duties. If the dilute solution concentration at the desorber outlet is different from the desired concentration, the steam duty is varied accordingly. This process is repeated until the dilute solution concentration, solution flow rate, and the desired absorber pressure are achieved, combined with acceptable energy balances at the absorber, evaporator and condenser. After acceptable heat balances are obtained for each component, the system is observed until the

31

various measured quantities do not change for a significant amount of time, and the data are recorded and checked again. A full data set consisting of 100 readings over a span of 5 min is then recorded. In this study, concentration, solution flow rate and absorber pressure vary over wide ranges. Three absorber pressures of 150, 345 and 500 kPa (21.8, 50.0 and 72.5 psi), and four different concentrations of 5, 15, 25 and 40% for three different flow rates for each combination of concentration and pressure result in 36 data points. To obtain data over these wide ranges, the ultimate driving forces are the laboratory chiller that works as the heat sink and the steam that provides the necessary heat duty at the desorber. The laboratory chiller and the steam are adjusted accordingly for each data point. The laboratory chiller has cooling capacity of 50 tons and can provide coolant at temperatures as low as -9.45oC. Water-glycol solution (approximately 50% by volume) is the coolant, and serves as the heat sink for the rectifier, condenser, and absorber. The laboratory steam lines can supply steam at up to 1379 kPa (200 psi). Two pressure regulators in the range 0 to 689 kPa (100 psi) and 0 to 2068 kPa (300 psi) are used to control the steam pressure. The steam line is directly connected to the desorber and determines desorber outlet temperature and the dilute solution concentration.

32

DATA ANALYSIS The procedures followed for the detailed analysis of the data taken in this study are described in this section. As explained in the previous sections, pressure, temperature and flow rate values at various points of interest were recorded using the Data Acquisition System. For each data point, 100 readings are recorded over a duration of 5 minutes, and the average value is used to represent the conditions for that data point. The thermodynamic properties of ammoniawater solution in the liquid and vapor phases were obtained from Engineering Equation Solver (EES) V7.431-3D software (Klein 2003). Calculation Methodology The working fluid, which is a binary mixture of ammonia and water, requires three Steamin Steam out independent parameters to define its state at any location in the system. As stated above, 11 Desout absolute pressures, 1 differential pressure, and 56 temperatures are measured at various Desorber locations and used as two of the required independent parameters for directly establishing the state. For these state points, the additional Desin required parameter is typically either concentration or quality. In this study, the quality at any given state was chosen as the third Fig 25. Schematic of the Desorber independent parameter (wherever appropriate), because the uncertainties associated with obtaining samples of the vapor and liquid states of the solution at different locations around the test loop during operation is expected to be considerable. At such state points, measured temperatures and pressures, and the expected quality (e.g., saturated liquid or saturated vapor) are used to obtain the solution concentration. Other properties such as enthalpy and specific volume are typically obtained using these three known independent parameters. (In some instances, the enthalpy obtained from energy balances is used as an input to compute quality or concentration, depending on the specific state point under consideration.) Three different flow rates are also measured: dilute solution, concentrated solution, and refrigerant flow rate. The steps to analyze the processes around the test loop are described using a representative data point with the following nominal values: dilute solution concentration at the desorber outlet of 15%, absorber inlet pressure 345 kPa (50.0 psi), and a concentrated solution flow rate of 0.026 kg/s (3.5 lbm/min). The ammonia-water solution exits as a two-phase mixture from the desorber, with the liquid and vapor phases in equilibrium with each other. Therefore, the liquid and vapor phase concentrations are computed (Fig 25) based on the measured temperature and pressure at the desorber outlet as follows, where q represents the quality of the particular stream:

xDilute, Des = f (TDes ,out , PDes ,out , q = 0 )

xDilute,Des = f (135.7C , 881.8 kPa, 0 ) = 0.1452

(1) (2) (3)

xV ,Des = f (TDes,out , PDes,out , q = 1)

33

xV ,Des = f (135.7C , 881.8 kPa, 1) = 0.6689

(4)

The ammonia concentration flowing through the refrigerant circuit of the loop is obtained using an assumption of a saturated vapor condition at the measured rectifier outlet temperature and pressure (Fig 26). (The condenser inlet pressure is used as the rectifier outlet pressure, since there is no local pressure measurement at the rectifier vapor outlet.) The resulting refrigerant concentration applies for the condenser, evaporator, pre-cooler and absorber inlet. Thus,

CWin

CWout Rectifier

Concentrated Vapor out Reflux Dilute Vapor in

Fig 26. Schematic of the Rectifier

xV = f (TRec,V ,out , PCon,in , q = 1)

xV = f ( 99.1C , 895.6 kPa, 1) = 0.922

(5) (6)

The reflux concentration is estimated by assuming a saturated liquid state leaving the rectifier at the measured temperature and pressure at reflux outlet location, as follows:
xReflux = f TReflux , PReflux , q = 0

(7) (8)

xReflux = f ( 70.8C , 870.8 kPa, 0 ) = 0.4315

The vapor concentration at the rectifier inlet is the same as the concentration of the vapor leaving the separator (and therefore, also the desorber, because the separator simply accomplishes physical separation of the two phases exiting the desorber). The reflux flow rate and the rectifier inlet flow rate are calculated using mass and species balances at the rectifier using the refrigerant flow rate, and the concentrations at the rectifier inlet, rectifier vapor outlet and the reflux outlet as follows:
xV ,Des = xSep,V ,out = xRec,in = 0.6689

(9) (10) (11) (12) (13) (14) (15)

mReflux = mRec,in mRef ,Measured mReflux = mRec,in 0.003531 kg s

mRe c,in i xV ,Des = mReflux i xReflux + mRef ,measured i xV mRe c,in i0.6689 = mReflux i0.4315 + 0.003531 kg s i0.922

mRe c,in = 0.007296 kg s

mReflux = 0.003765 kg s

Mass and species balances at the separator outlet (Fig 27) are used to find the dilute solution concentration entering the absorber, and a species balance at the absorber is used to find the concentrated solution concentration, as follows:

34

mSep,out = mDilute mReflux

(16) (17) (18) (19) (20) (21)

mSep,out = 0.02199 kg s 0.003765 kg s = 0.01823 kg s

xDilute imDilute = mReflux xReflux + mSep,out xDilute,Des xDilute i0.02199 kg s = 0.003765 kg s 0.4315 + 0.01823 0.1452 = 0.1942
xConcentrated mConcentrated = mRef ,Measure xV + mDilute xDilute

xConcentrated 0.02638 kg s = 0.003531 kg s 0.922 + 0.02199 kg s 0.1942 = 0.2852

It should be noted that the dilute solution concentration entering the absorber is different from the concentration at the desorber outlet, because the reflux mixes with the solution leaving the separator before it enters the absorber. With the concentrations calculated as described above and measured temperatures and pressures at the various locations, the other properties at each location are obtained. The following sections describe the calculation of the heat duties and energy balances at each component, and the calculation of the measured absorber heat transfer coefficient.
Absorber

Separator Sepout

Reflux

Dilute

Fig 27. Separator Liquid-Phase Balance

There are three different fluid streams entering and exiting the absorber: entering dilute solution, exiting concentrated solution and the entering refrigerant vapor (Fig 28). Pressures and temperatures are measured at these three locations. All three concentrations are calculated as explained in the previous section. Therefore, these states are fully established and the enthalpies at these three locations can be obtained as follows:
hAbs,in = f (TAbs,in ,PAbs,in ,xDilute ) hAbs,in = f ( 71.9C ,340.7 kPa, 0.1942 ) =174.1 kJ kg hAbs,out = f (TAbs,out ,PAbs,out ,xConcentrated )
hAbs,out = f ( 44.86C ,347.7 kPa, 0.2852 ) = 10.32 kJ kg

(22) (23) (24) (25) (26) (27)

Abs out

Abs in Absorber CW out

hAbs,V ,in = f (TAbs,V ,in ,PAbs,V ,in ,xV )

hAbs,V ,in = f ( 2.83C ,359.4 kPa,0.922 ) = 811.1 kJ kg

Abs V,in CWin

Fig 28. Schematic of the Absorber

35

With the three enthalpies known, the solution side absorber heat duty is calculated as follows:
QAbs = mDilute hAbs,in + mRef ,Measured hAbs,V ,in mConcentrated hAbs,out

(28) (29)

QAbs = 0.02199 kg s 174.1 kJ kg + 0.003531 kg s 811.1 kJ kg 0.02638 kg s 10.32 kJ kg = 6.419 kW

The coolant-side heat duty is calculated using the absorber coolant flow rate and temperatures at the inlet and outlet:
Cp Abs,C = f (TAbs,C ,Ave ) Cp Abs,C = f ( 36.7C ) = 4.183 kJ kg C

(30) (31) (32) (33)

QAbs,C = m Abs,C Cp Abs,C (TAbs,C ,out TAbs,C ,in )

QAbs,C = 0.1559 kg s 4.183 kJ kg C ( 41.33C 32.07C ) = 6.037 kW

The coolant specific heat is calculated at the average absorber coolant temperature. After the absorber heat duty is obtained, the overall heat transfer coefficient is obtained using the log mean temperature difference. Here, the log mean temperature difference is defined using saturated solution temperatures at the inlet and outlet of the absorber, which represent the effective driving temperature difference for the absorption process better than the actual temperatures. It is assumed that absorber tube array is completely wetted. The average of the solution- and coolantside absorber heat duties is used as the representative absorber heat duty for the calculation of the overall heat transfer coefficient as follows:
LMTDAbs =

(TAbs,sat ,in TAbs,C ,out ) (TAbs,sat ,out TAbs,C ,in )

T TAbs,C ,out ln Abs,sat ,in T Abs,sat ,out TAbs,C ,in = 40.37C

(34)

LMTDAbs =

(86.93C 41.33C ) ( 67.62C 32.07C )

86.93C 41.33C ln 67.62C 32.07C

(35)

U Abs =
U Abs =

Heat Duty Abs,ave Abs Area LMTDAbs

2 = 735.5 W m 2 C

(36) (37)

( 6.419 kW + 6.037 kW )
0.2098 m 40.37C
2

Once the overall heat transfer coefficient is obtained, the solution-side heat transfer coefficient is calculated by accounting for the wall resistance and the coolant-side heat transfer resistance. The tube-wall resistance is given by:
RW = TubeOD TubeOD ln 2 kW TubeID

(38) (39)

RW =

9.525 103 m 9.525 103 m ln = 5.112 105 m 2 C W 3 2 15.07 W m C 8.103 10 m

36

For the calculation of the coolant-side heat transfer coefficient, the coolant velocity and Reynolds number are first calculated:
VAbs ,C = V / TubeCross, Area NColumn

(40)

VAbs ,C = 1.569 104 m3 s

) (5.156 10

m 2 4 = 0.7608 m / s

(41) (42)

Re Abs ,C =

Abs ,C VAbs ,C TubeID Abs ,C

= 8800

Re Abs ,C =

993.4 kg / m3 0.7608 m / s 8.103 103 m 6.959 104 kg m s

(43)

Here, the absorber coolant density, viscosity, conductivity (0.613 W/m-K), and Prandtl number (4.747) are calculated at the average absorber coolant temperature (36.7oC). The tubeside friction factor and Nusselt number are calculated using Churchills equations (1977b; 1977a) to yield f = 0.03212 and Nu = 66.27. The coolant-side heat transfer coefficient is obtained from the Nusselt number as follows:
Abs ,C =
k Abs ,C Nu Abs ,C TubeID
3

(44) (45)

Abs ,C =

0.6132 W m C 66.27 8.103 10 m

= 5015W m 2 C

With the coolant-side heat transfer coefficient known, the solution-side heat transfer coefficient is calculated as follows:
1

Abs,Film U Abs
1 = 1

TubeOD RW + Abs,C TubeID

(46)

Abs,Film

9.525 103 m 5 2 5 . 112 10 m C W + 2 2 3 735.5W m C 5015W m C 8.103 10 m

2

(47)

Abs,Film = 931.1W m C

Desorber

The desorber inlet state is specified by the concentrated solution concentration, and the measured temperature and pressure:
hDes,in = f (TDes,in ,PDes,in ,xConcentrated ) hDes,in = f ( 97.61C ,1003 kPa,0.2852 ) = 240.9 kJ kg

(48) (49)

The desorber outlet enthalpy is assumed to be the same as the separator inlet enthalpy due to the small insulated line separating them (Fig 29), and it is further assumed that the separator is adiabatic. The various enthalpies at the separator are calculated as follows:
hSep,out = f TSep,out ,PSep,out ,xDilute,Des

(50)

37

SepV,out Separator Sepout

Steamin Desout

Steam out Desorber Desin

Fig 29. Desorber and Separator Balance

hSep,out = f (132.2C ,883.2 kPa, 0.1452 ) = 464.9 kJ kg hSep,V ,out = f TSep ,V ,out ,PSep,V ,out ,xDes,V

(51) (52) (53)

hSep,V ,out = f (135.7C ,883.2 kPa, 0.6689 ) = 1940 kJ kg

Since the separator vapor and solution outlet states were fully established in the previous section, mass and energy balances at the separator are used to calculate the separator inlet enthalpy, which is in turn used to compute the desorber heat duty:
hSep ,in = hDes,out hDes,out mConcentrated = mRef ,Measured + mReflux hSep ,V ,out + mDilute mReflux hSep,out
hDes ,out 0.02638 kg s = ( 0.003531 kg s + 0.003765 kg s ) 1940 kJ kg + ( 0.02199 kg s 0.003765 kg s ) 464.9 kJ kg hDes ,out = 857.7 kJ kg

(54)

(55) (56) (57) (58)

QDes = mConcentrated ( hDes,out hDes,in )

QDes = 0.02638 kg s ( 857.7 kJ kg 240.9 kJ kg ) = 16.27 kW

Evaporator

The measured refrigerant flow rate, temperature and pressure, and previously obtained concentration specify both the inlet and outlet states (Fig 30). Enthalpies at the inlet and outlet are obtained from these measured and deduced quantities as follows:
hEva,in = f (TEva,in ,PEva,in ,x )

Evain CWout

Evaporator CWin

Evaout

Fig 30. Evaporator Energy Balance

(59) (60) (61) (62)

hEva,in = f ( 1.227C ,379.3 kPa, 0.922 ) = 60.58 kJ kg hEva,out = f (TEva,out ,PEva,out ,x )

hEva,out = f (1.188C ,348.5 kPa, 0.922 ) = 776.1 kJ kg

38

Based on these enthalpies, the evaporator duty is calculated as follows:

QEva = mRef ,Measured ( hEva,out hEva ,in ) QEva = 0.003531 kg s ( 776.1 kJ kg 60.58 kJ kg ) = 2.526 kW

(63) (64)

The evaporator heating fluid heat duty is obtained from the flow rate and inlet and outlet temperatures of the water-glycol solution (50% by volume). The evaporator heating fluid specific heat is calculated based on the water-glycol properties at the average fluid temperature:
CpEva,C = f (TEva,C ,Ave ) CpEva,C = f (16.42C ) = 3.308 kJ kg C

(65) (66) (67) (68)

QEva,C = mEva,C CpEva,C (TEva,C ,in TEva,C ,out )

QEva,C = 0.1449 kg s 3.308 kJ kg C (18.85C 13.98C ) = 2.334 kW

Condenser

The refrigerant flow rate and measured temperatures and pressures, along with the previously obtained refrigerant concentration are used to obtain the condenser heat duty. The enthalpies at the inlet and outlet of the condenser are obtained readily as both the inlet and outlet states of the condenser are fully specified:

CWin Conout

CWout Condenser Conin

Fig 31. Condenser Energy Balance

hCon,in = f (TCon,in ,PCon,in ,x ) hCon,in = f ( 93.09C ,895.6 kPa,0.922 ) = 1495 kJ kg hCon,out = f (TCon,out ,PCon,out ,x ) hCon,out = f ( 22.01C,903.2 kPa, 0.922 ) = 46.33 kJ kg

(69) (70) (71) (72)

Now the condenser duty is given by:

QCon = mRef ,Measured ( hCon,in hCon,out )
QCon = 0.003531 kg s (1495 kJ kg 46.33 kJ kg ) = 5.115 kW

(73) (74)

The condenser coolant heat duty (Fig 31) is calculated in a manner similar to that used for the evaporator. The condenser coolant is also water-glycol solution (50% by volume), and its specific heat is calculated based at the average coolant temperature. Thus,
CpCon,C = f (TCon,C ,Ave )
CpCon,C = f (19.1C ) = 3.326 kJ kg C

(75) (76)

39

QCon,C = mCon,C CpCon,C (TCon,C ,out TCon,C ,in ) QCon,C = 0.2681 kg s 3.326 kJ kg C ( 21.69C 16.42C ) = 4.699 kW

(77) (78)

Rectifier

The mass flow rates of the entering and leaving vapor and the reflux, as well as enthalpies of these streams are necessary for the calculation of the rectifier heat duty. The refrigerant vapor outlet state is fully defined based on the measured temperature and pressure and the saturated vapor state. The mass flow rates of the reflux and the vapor entering the rectifier are calculated using mass and species balances as explained in a previous section. The enthalpies at the rectifier vapor inlet and outlet, and for the reflux are obtained as follows:
hRec,in = f (TRec,in ,PRec,in ,xRec,in )
hRec,in = f (134.4C ,883.2 kPa, 0.6689 ) = 1884 kJ kg

(79) (80) (81) (82) (83) (84)

hRec,V ,out = f (TRec,V ,out ,PRec,V ,out ,xRec,V ,out )

hRec,V ,out = f ( 99.13C ,895.6 kPa, 0.922 ) = 1574 kJ kg hReflux = f TReflux ,PReflux ,xReflux

hReflux = f ( 70.76C,870.8 kPa, 0.4315 ) = 81.56 kJ kg

Now, rectifier duty is calculated as follows:

QRec = mRef ,Measured + mReflux hRec,in mReflux hReflux + mRef ,Measured hRec,V ,out QRec = ( 0.003531 kg s + 0.003765 kg s )i1884 kJ kg ( 0.003765 kg s 81.56 kJ kg + 0.003531 kg s 1574kJ kg ) = 7.881 kW

(85) (86)

The rectifier coolant heat duty is calculated using the rectifier coolant (50% water-glycol mixture) flow rate and inlet and outlet temperatures. The coolant specific heat is calculated at the average coolant temperature, yielding the coolant-side duty as follows:
CpRec,C = f (TRec,C ,Ave )
CpRec,C = f ( 29.89C ) = 3.494 kJ kg C QRec,C = mRec,C CpRec,C (TRec,C ,out TRec,C ,in ) QRec,C = 0.05951 kg s 3.494 kJ kg C ( 45.97C 13.81C ) = 6.687 kW

(87) (88) (89) (90)

40

RESULTS AND DISCUSSION Experimental results obtained in this project are discussed in this section, along with a discussion of the trends observed from the data. Absorber heat duty and overall and solution heat transfer coefficients are presented first as a function of concentrated solution flow rate, with absorber pressure and dilute solution concentration at the desorber outlet as parameters. Parametric effects of absorber pressure and dilute solution concentration at the desorber outlet on the solution heat transfer coefficient are discussed separately. Nominally, three solution flow rates, 0.019, 0.026, and 0.034 kg/s (2.5, 3.5, and 4.5 lbm/min) designated as low, medium and high flow rate are tested at each of the nominal absorber pressures of 500, 345 and 150 kPa (72.5, 50.0 and 21.8 psi) representing a wide range of heat source/sink combinations. Nominal concentrations at the desorber outlet of 5, 15, 25, and 40% are tested for the above combinations of absorber pressures and concentrated solution flow rates. Tests were conducted at conditions as close to the nominal values as possible with the test facility described in previous sections. Test Matrix Table 3 shows the matrix of data points obtained in this study. Table 4 outlines the primary challenges in obtaining the nominal conditions, along with the techniques used to maintain conditions close to nominal values. The cells without shading represent test conditions obtained without the heater in the absorber coolant loop and also without the subcooler upstream of the dilute solution inlet to the absorber. The cells with shading represent tests conditions obtained with the help of either the absorber coolant heater or the subcooler The overall test matrix can be subdivided into twelve blocks based on absorber inlet pressures and dilute solution concentrations at the desorber. Tests for four of these blocks (no shading) are conducted without the absorber coolant heater and the subcooler. The absorber coolant heater is used for test conditions in three of these blocks, while the subcooler is used for test conditions in seven of these blocks. Note that the darker cells represent conditions that were progressively difficult to obtain for the reasons shown.

Tests conducted without the use of the heater or the subcooler were typically at absorber pressures only slightly higher than the ammonia-water solution saturation pressures at the charged concentrations in the loop and room temperature. For example, to obtain the 5% and 15% dilute solution concentration conditions at the desorber outlet, the loop is charged with ammonia-water solution yielding approximately a maximum of 25% ammonia mass fraction that yields a saturation pressure of 57.2 kPa (8.3 psi) at 25oC room temperature. Likewise, to obtain the 25% dilute solution concentration, the loop is charged with about 35% ammonia mass fraction, which yields a solution saturation pressure of 122.1 kPa (17.7 psi) at 25oC room temperature. To obtain 40% dilute solution concentration, the loop is charged with about 50% ammonia mass fraction, which yields a solution saturation pressure of 303 kPa (44 psi) at 25oC room temperature. These are the saturation pressures in the system shut-down state. When the system is operating, the absorber temperature increases because of the high temperature of the solution coming from the desorber, and due to the heat of absorption in the absorber. Therefore, when system is running, the absorber pressures are slightly higher than the system shut down state pressures. Tests could be conducted with relative ease for the conditions where the nominal absorber pressure is relatively close to the system shut down state pressure, e.g., 150 kPa (21.8

41

Table 3. Test Conditions for the Present Study

150kPa Flow Rate (kg/s) 0.0189 0.0265 0.0340 Flow Rate (kg/s) 5% Concentration (%) Pressure (kPa) Flow Rate (kg/s) 15% Concentration (%) Pressure (kPa) Flow Rate (kg/s) 25% Concentration (%) Pressure (kPa) Flow Rate (kg/s) 40% Concentration (%) Pressure (kPa) 345 kPa Flow Rate (kg/s) 0.0189 0.0265 0.0340 500 kPa Flow Rate (kg/s) 0.0189 0.0265 0.0340

0.0186 9.8 174.7 0.0193 13.5 172.9 0.019 24.0 183.6 0.0195 38.34 247.3

0.0265 9.3 177.7 0.0264 15.2 167.5 0.0263 24.1 175.8 0.0263 40.16 260.1

0.0342 10.9 179.5 0.0341 15.6 175.9 0.0338 25.9 180.5 0.0329 39.39 277.4

0.0189 9.1 341.6 0.0189 15.5 354.2 0.0188 23.9 349.3 0.0186 36.2 358.9

0.0263 10.1 358.3 0.0264 14.5 340.7 0.0265 23.9 351.8 0.0266 39.8 352

0.0301 10.3 358.5 0.0339 15.4 336.0 0.0337 25.2 350.9 0.0343 40.8 343.7

0.0223 9.88 469.9 0.0191 14.78 476.1 0.0186 26.8 512.4 0.019 34.2 509

0.0262 10.4 480.8 0.0265 14.65 513.1 0.0264 24.9 511.9 0.0264 37.4 501.4

0.0337 11.5 496.6 0.0343 15.64 480.8 0.0341 25.6 515.7 0.0341 38.7 510.4

psi) cases at 5% dilute solution concentration, which were obtained without the heater or subcooler. Similarly, 345 and 500 kPa (50.0 and 72.5 psi) cases at 25%, and 500 kPa (72.5 psi) cases at 40% could be obtained without the heater or the subcooler. If the nominal absorber pressures are either too low or too high compared to the system shut down state pressure for the respective charged concentrations, it was necessary to modify the test facility to overcome the large pressure difference, as the absorber coolant also poses limitations. When the nominal absorber pressures are much higher than the system shut down state pressure at charged concentrations, it is necessary to use the heater in the absorber coolant loop for increasing the solution temperature, which in turn increases the absorber pressure to the required value. It should be noted that if the heater is not used, the absorber pressure can be increased by flowing the coolant at an extremely small flow rate, which would raise the heat rejection temperature and therefore the absorber pressure. However, the resulting temperature rise in the coolant would be unrealistically high, and would confound the absorber solution heat transfer coefficient measurement. By raising the coolant inlet temperature through the use of the heater, a higher coolant flow rate can be maintained (leading to lower uncertainties in the solution-side heat transfer coefficient) and also the coolant temperature difference can be maintained within 42

Table 4. Primary Challenges and Techniques Used

150kPa Flow Rate (kg/s) 0.0189 0.0265 0.0340 Flow Rate (kg/s) 5% Concentration (%) Pressure (kPa) Flow Rate (kg/s) 15% Concentration (%) Pressure (kPa) Flow Rate (kg/s) 25% Concentration (%) Pressure (kPa) Flow Rate (kg/s) 40% Concentration (%) Pressure (kPa)
High shut down pressure, difficult to get and maintain low pressure. Large coolant flow rate, very small coolant T.

345 kPa Flow Rate (kg/s) 0.0189 0.0265 0.0340

Very small coolant flow rate, large coolant T. Electric heater is used at absorber coolant loop

500 kPa Flow Rate (kg/s) 0.0189 0.0265 0.0340

Low shut down pressure, difficult to maintain high pressure while running. Very small coolant flow rate, large coolant T. Electric heater is used at absorber coolant loop Low shut down pressure, difficult to maintain high pressure while running. Very small coolant flow rate, large coolant T. Electric heater is used at absorber coolant loop

0.0186 9.8 174.7

0.0265 9.3 177.7

0.0342 10.9 179.5 0.0189

Sub-cooler is used to obtain low absorber pressure.

15.5 354.2 0.0188

Sub-cooler is used to obtain low absorber pressure.

0.0265 23.9 351.8

0.0337 25.2 350.9

0.0186 26.8 512.4 0.019

0.0264 24.9 511.9 0.0264 37.4 501.4

0.0341 25.6 515.7 0.0341 38.7 510.4

Sub-cooler is used to obtain low absorber pressure.

23.9 349.3

Sub-cooler is used to obtain low absorber pressure.

Sub-cooler is used to obtain low absorber pressure.

34.2 509

reasonable limits. When the nominal absorber pressure is much less than the system shut down state pressure at the charged concentration, it is necessary to use the subcooler to decrease the temperature of the solution flowing to the absorber to values below what is possible with the solution heat exchanger. The subcooler therefore helps in decreasing the sensible cooling load on the absorber and the absorber pressure. Test conditions with the light gray shaded blocks in Table 4 are obtained with the use of either the absorber coolant heater or the subcooler. At the absolute extremes of the test matrix, despite using the heater in the absorber coolant loop and the subcooler upstream of the absorber, it is very difficult to obtain test conditions where the absorber pressure is substantially different from the system shut down state pressure at the charged concentration. For example, for the 500 kPa (72.5 psi) test conditions, the absorber coolant heater is used for the 5% and 15% concentration cases. However, it is still quite difficult to increase the absorber pressure up to 500 kPa (72.5 psi) while maintaining reasonable coolant flow rates. Even if the 500 kPa (72.5 psi) pressure is obtained, the absorber coolant flow rates required are extremely small, and in some cases, result in very small coolant heat transfer coefficients that in turn result in unreasonable solution-side heat transfer

43

coefficients. Similarly, even if the subcooler is used at its fullest capacity, at the 40% concentration cases, the lowest absorber pressure achievable while still retaining acceptable uncertainties in heat transfer coefficient determination was 241 kPa (35 psi) lower pressures could in fact be obtained; however, this would be at the expense of heat transfer coefficient accuracies. The actual conditions obtained with reasonable coolant temperature rises and reasonable energy balances are shown in Table 3.
Energy Balance Quantification and Preliminary Uncertainties Table 5 shows the number of data within acceptable energy balances for the absorber, condenser and evaporator. For each of these components, energy balances were calculated between the two sides of the exchangers. The criterion for acceptability was to ensure that the calculated energy balances are within 15% for all three components. As Table 5 shows, a significant number of the data points have energy balances within 5% or 10%, with all the data having absorber energy balances within 15%. The average difference between the heat duties on the two sides of the absorber for the whole text matrix is 4.77%. (For 3 data points in the matrix, the condenser and/or evaporator balances are outside this limit.)

A rigorous uncertainty propagation analysis was also conducted on the data collected in this study. The results are summarized in Table 6. As shown in this table, only two data points show uncertainties > 20% in the film heat transfer coefficient. Among all variables, the absorber vapor inlet pressure contributes the greatest to the total uncertainty in the film heat transfer coefficient. There are 5 data points showing uncertainties > 25% for the reflux; this is to be expected because of the relatively low flow rates of the reflux, and also the high sensitivity of the vapor phase properties to the vapor concentration in the vicinity of the rectifier. The highest uncertainties in the reflux calculation are for data at 345 kPa (50 psi) and 40% desorber outlet concentration this is because at such high solution concentrations, the reflux flow rate is very small. It should be noted once again that the above comments are based on single-point, preliminary analyses of the data. Due to the gradients in the absorber, segmented analyses within the component are necessary to more clearly establish the heat and mass transfer coefficients and their respective uncertainties. This will be done in Phase II of this project. The currently reported uncertainties are primarily to establish the validity of the experiments. Also, for these analyses, complete wetting of the tubes was assumed to obtain overall and solution-side heat transfer coefficients in the absorber. These assumptions will be revisited in Phase II of the project. Figs 32-34 show concentration ranges for each data point obtained in the study. Thus, each data point is represented by three symbols, representing the solution concentration at the desorber outlet, the absorber inlet and the absorber outlet, with the symbols being plotted as a function of the concentrated solution flow rate. It should be noted that the absorber inlet
Table 5. Energy Balances for Absorber, Condenser and Evaporator
Component (0 - 5)% Absorber Condenser Evaporator Number of Data within (5 - 10)% (10 - 15)% (15 - 18)% Average Absolute Difference (%) 4.77 7.05 7.31

20 12 12

14 17 17

2 5 4

2 3

44

Table 6. Uncertainties in Key Parameters

Parameter Range Uncertainty Range (Absolute) Uncertainty Range (%) Average Uncertainty (%)

Xdes,out Xabs,in Xabs,out Xref,vap Qabs (kW) Qabs_C (kW) Qabs_avg (kW) C (W/m2-K) U (W/m2-K) Film (W/m2-K) Reflux (lbm/min)

0.091 0.408 0.152 0.418 0.207 0.442 0.807 0.999 3.26 10.75 2.96 9.79 3.11 10.20 2634 11408 476.4 941.9 558.8 1140 0.0025 1.073

0.0028 0.0046 0.0025 0.0185 0.0022 0.0159 0.0001 0.0047 0.06 1.54 0.17 1.63 0.10 0.99 14.99 114.3 15.15 153.8 24.71 219.4 0.0003-0.1793

0.85 3.48 0.92 9.52 0.76 6.13 0.01 0.57 1.31 20.51 3.32 30.27 1.84 15.89 0.42 1.46 2.79 18.46 3.69 23.83 5.55 108.01

1.86 2.88 1.97 0.20 6.18 13.31 7.84 0.63 9.84

11.90

18.13

concentration is different from the desorber outlet concentration because the reflux is mixed with the dilute solution stream before it flows to the absorber. The difference in these two concentrations, therefore, depends on the reflux concentration and flow rate, which in turn change as the test conditions are varied. For the nominal 150 kPa (21.8 psi) case (Fig 32), for a given dilute solution concentration, the concentration range decreases slightly with increasing solution flow rate. As the dilute solution concentration increases, the concentration range decreases at any given flow rate. Concentration ranges are small for the 40% and 25% concentration cases, while the ranges are large for the 5% and 15% concentration cases. The smaller

Fig 32. Concentration Ranges at Nominal 150 kPa (21.8 psi)

45

concentration ranges at the higher dilute solution concentration cases are primarily due to the lower refrigerant flow rates in these cases. Fig 33 shows the variation of the concentration ranges for various dilute solution concentrations at the desorber outlet at a nominal absorber pressure of 345 kPa (50.0 psi). The concentration range shows similar trends as those for the 150 kPa (21.8 psi) cases. However, concentration ranges are larger at this pressure than for the 150 kPa (21.8 psi) nominal cases for 40% and 25% concentration, while the concentration Fig 33. Concentration Ranges at Nominal ranges for the 5% and 15% concentration 345 kPa (50.0 psi) cases are similar to those of the 150 kPa (21.8 psi) nominal cases. Fig 34 shows the variation of concentration ranges for various dilute solution concentrations at the desorber outlet at a nominal absorber pressure of 500 kPa (72.5 psi). As with the 150 kPa (21.8 psi) and 345 kPa (50 psi) cases, the concentration ranges decrease slightly as the solution flow rate increases. The concentration ranges for the various flow rates for 40% concentration are much larger than those at 150 and 345 kPa (21.8 psi and 50.0 psi), while for 5% and 15%, the concentration ranges are smaller than those at 150 and 345 kPa (21.8 psi and 50.0 psi). The various concentration ranges observed at these conditions can also be taken as indicator of difficulties in establishing the particular test condition. In general, the more challenging test conditions result in smaller concentration ranges. In the presentation of the results that follows, the concentrated solution flow rates are shown in terms of the linear mass flux based on the tube length. Thus, the concentrated solution flow rate is divided by the total length of the four tubes over which the solution flows, and further divided by a factor of two to account for the fact that the solution flows around two sides of a tube. This yields a concentrated solution mass flux = m ( 2 Ltube N tube ) . Therefore, a nominal flow rate of 0.019 kg/s (2.5 lbm/min) corresponds to 0.0081 kg/m-s, 0.026 kg/s (3.5 lbm/min) corresponds to 0.0113 kg/m-s and 0.034 kg/s (4.5 lbm/min) corresponds to 0.0146 kg/m-s. In the discussion of the trends, however, the flow rate in lbm/min or kg/s is used for ease of comprehension.

Fig 34. Concentration Ranges for Nominal 500 kPa (72.5 psi)

46

Absorber Heat Duty, Overall U, and Solution-Side Heat Transfer COefficient Fig 35 shows the absorber heat duties measured in this study as a function of concentrated solution flow rate, which has a significant effect on the heat duty. At any pressure or solution concentration, as expected, the absorber heat duty increases with increasing solution flow rate, ranging from 3.1 to 10.2 kW. No other significant conclusions should be drawn about absorption heat and mass transfer from these heat duty graphs, because the duty depends to a large extent on the coolant flow rates and the log mean temperature difference provided. Fig 35. Absorber Duty vs. Solution Flow Rate Fig 36 shows the overall heat transfer coefficient U as a function of the concentrated solution flow rate. The U varies from 476 to 942 W/m2-K over the entire test matrix depending on the test conditions, in general increasing as the concentrated solution flow rate is increased. Once again, the overall U is a function of tube-side and solution-side phenomena, and therefore, no significant conclusions should be drawn about absorption phenomena from these graphs.

Figure 37 shows the solution heat transfer coefficient for all the data points obtained in the current experiments. It should be noted that the solution heat transfer coefficients are based on the entire tube surface area (assuming complete wetting). The solution heat transfer coefficient ranges from 558 to 1140 W/m2-K, depending on the test condition. An important aspect can be observed by comparing the graphs in Fig 36 and 37: the trend in the solution heat transfer coefficient is similar to the trend in the overall U. This is due to the large absorber coolant heat transfer coefficients for most of the test conditions. By keeping the coolant heat transfer coefficient high, the solution side resistance becomes dominant and the effect of the coolant-side heat transfer coefficient is minimized, which yields low uncertainties in the absorption heat transfer coefficients. This phenomenon is further illustrated in Appendix A, where the overall U, coolant h, and solution-side h for each test are plotted together. The graphs in this appendix clearly demonstrate that the test procedures used successfully maintained the solution-side resistance as the dominant resistance for most of the data points over this wide range of solution flow rates, pressures and concentrations. As expected, Figure 37 shows that the solution heat transfer coefficient increases Fig 36. Overall U vs. Solution Flow Rate 47

with the solution flow rate. The effects of the dilute solution concentration and absorber pressure on the solution heat transfer coefficient are discussed in detail in the following sections.
Effect of Dilute Solution Concentration Figs 38-40 show the effect of dilute solution concentration on the solution heat transfer coefficient at nominal pressures of 150, 345, and 500 kPa (21.8 psi, 50.0 psi, and 72.5 psi), respectively. The solution heat transfer coefficient increases almost linearly with increasing Fig 37. Solution-Side Heat Transfer Coefficient vs. Flow Rate concentrated solution flow rate regardless of the dilute solution concentration at all three pressures. However, there are no general trends that demonstrate a strong influence of dilute solution concentration at a given absorber pressure. At a nominal pressure of 150 kPa or 21.8 psi (Fig 38), at a given concentrated solution flow rate, the solution heat transfer coefficient decreases with increasing dilute solution concentration, although the effect is not very pronounced at a flow rate of 0.019 kg/s (2.5 lbm/min). Solution heat transfer coefficients are the highest for the 5% nominal concentration case, and increase significantly when the flow rate is increased from 0.019 to 0.026 kg/s (2.5 to 3.5 lbm/min). At nominal concentrations of 25% and 40%, the heat transfer coefficients are lower than those for 5% and 15% for all flow rates. Also, there does not appear to be an appreciable difference in the heat transfer coefficients at nominal concentrations of 25% and 40%.

The low absorber pressures of 150 kPa (21.8 psi) necessitate the use of the subcooler described in the previous section for solution concentrations of 15%, 25% and 40%. This results in some of the sensible heat being removed prior to the solution entering the absorber, which reduces absorber heat duties and could cause a decrease in the solution heat transfer coefficient. In other words, the sensible heat fraction in the 5% nominal solution concentration case is larger (because the subcooler is not used) compared to the data at the other concentrations, leading to larger absorber heat duties. It should also be noted that due to the challenges described in the previous section, the nominal combination of low pressure (150 kPa or 21.8 psi) and high Fig 38. Solution Heat Transfer Coefficient vs. concentration (40%) was in reality Flow Rate at 150 kPa (21.8 psi) 48

tested at somewhat higher absorber inlet pressures (241 276 kPa or 35 40 psi, depending upon the flow rate). These different actual pressures for the same nominal pressure could also contribute to come of the differences in heat transfer coefficients, and therefore the trends seen in Fig 39 represent the combined influence of changes in concentration as well as operating pressure, to some extent. Fig 39 shows the solution heat transfer coefficient as a function of the concentrated solution flow rate with the dilute solution concentration as a Fig 39. Solution Heat Transfer Coefficient vs. parameter at a nominal pressure of 345 Flow Rate at 345 kPa (50.0 psi) kPa (50.0 psi). The solution heat transfer coefficient increases as the concentrated solution flow rate increases for each concentration, although there is no monotonous, systematic effect of concentration. For example, the 15% concentration cases show higher heat transfer coefficients than those at a nominal concentration of 5%, mostly due to the larger concentration range at the absorber. The lowest solution heat transfer coefficients are observed for the 40% dilute solution concentration case. The data at the 5% nominal concentration show an increase in solution heat transfer coefficient between a solution flow rate of 0.019 kg/s (2.5 lbm/min) and 0.026 kg/s (3.5 lbm/min), and a less significant influence at the higher flow rates. This may be partly because the actual measured inlet concentrations at the 5% and 15% nominal concentrations are somewhat close to each other at the higher flow rates of 0.026 kg/s (3.5 lbm/min) and 0.030 kg/s (4.0 lbm/min). The use of the subcooler at 345 kPa (50.0 psi) and 40% may also have caused lower heat transfer coefficients. The 15% and 25% nominal concentration cases show very similar heat transfer coefficients values at any given flow rate, primarily due to the use of the subcooler for the 15% nominal concentration case. Fig 40 shows the solution heat transfer coefficient as a function of the concentrated solution flow rate with the dilute solution concentration as a parameter at a nominal absorber pressure of 500 kPa (72.5 psi). At low flow rates, the solution heat transfer coefficients for the 15% and 40% are lower than those of 5% and 25%, however, the trend reverses itself somewhat as the flow rate increases. Also, the rise in heat transfer coefficient with flow rate is higher for

Fig 40. Solution Heat Transfer Coefficient vs. Flow Rate at 500 kPa (72.5 psi)

49

the 15% and 40% concentration cases as compared to the 5% and 25% cases. The 5% and 15% nominal concentration cases required very low coolant flow rates to maintain the higher absorber pressures, which resulted in very large coolant Ts. To keep the coolant temperature difference within reasonable limits, the coolant inlet temperature is increased by using a cartridge heater in the absorber coolant loop, which resulted in different driving temperature difference profiles along the absorber. This may have affected the absorption Fig 41. Solution Heat Transfer Coefficient vs. rates and heat transfer coefficients. Flow Rate (xdes,out = 5% Nominal) The heat transfer coefficients for 40% concentration are in general lower than those at 25% except at the highest flow rate, which may be because of the smaller driving concentration gradients available for absorption into the solution richer in ammonia.
Effect of Absorber Pressure Figs 41-44 show the effect of absorber inlet pressure on the solution heat transfer coefficient for nominal dilute solution concentrations at the desorber outlet of 5%, 15%, 25%, and 40%, respectively. As discussed in the previous section on the effect of dilute solution concentration, the solution heat transfer coefficient increases almost linearly as the flow rate increases for any given concentration. Fig 41 shows the effect of the absorber inlet pressure on the solution heat transfer coefficient for the nominal 5% concentration cases. The heat transfer coefficient increases substantially as when the flow rate is increased from 0.019 to 0.026 kg/s (2.5 to 3.5 lbm/min) for the 150 and 345 kPa (21.8 and 50.0 psi) nominal pressure cases, while the increase is more gradual at a nominal pressure of 500 kPa (72.5 psi). The heat transfer coefficient decreases as the absorber pressure increases (except at 0.019 kg/s or 2.5 lbm/min). Fig 42 shows the effect of the absorber inlet pressure on the solution heat transfer coefficient at a nominal concentration of 15%. The solution heat transfer coefficient increases almost linearly with increasing solution flow rate; however, the increase is more pronounced at 500 kPa (72.5 psi) than at 150 and 345 kPa (21.8 and 50.0 psi). The heat transfer Fig 42. Solution Heat Transfer Coefficient vs. coefficients at 345 kPa (50 psi) are the Flow Rate (xdes,out = 15% nominal)

50

highest at all solution flow rates, followed by the heat transfer coefficients at 150 and 500 kPa (21.8 psi and 72.5 psi). (The lower solution heat transfer coefficient at 150 kPa (21.8 psi) may again be due to the use of the subcooler to obtain these data.) Fig 43 shows the effect of the absorber inlet pressure on the solution heat transfer coefficient at a nominal concentration of 25%. Heat transfer coefficients at 345 kPa (50.0 psi) are generally higher than those at 500 kPa (72.5 psi) except at 0.026 kg/s (3.5 lbm/min), where the heat transfer coefficients are very close to each other. The heat transfer coefficients at 150 kPa (21.8 psi) are lower than those at 345 and 500 kPa (50.0 and 72.5 psi) for all the solution flow rates. Fig 44 shows the effect of absorber inlet pressure on the solution heat transfer coefficient for a nominal dilute solution concentration of 40%. Solution heat transfer coefficients at 500 kPa (72.5 psi) are higher than those at 150 and 345 kPa (21.8 and 50.0 psi) for all the solution flow rates. In addition, the heat transfer coefficients at 150 kPa (21.8 psi) are slightly larger than those at 345 kPa (50 psi) at all the solution flow rates. In general, the slopes of these curves at all the pressures are higher for the 40% concentration case compared to those at other concentrations.

Fig 43. Solution Heat Transfer Coefficient vs. Flow Rate (xdes,out = 25% nominal)

Fig 44. Solution Heat Transfer Coefficient vs. Flow Rate (xdes,out = 40% nominal)

51

CONCLUSIONS Experiments on ammonia-water absorption in horizontal tube banks were conducted. A large test facility was designed and constructed with numerous design features and control and plumbing options to enable the measurement of solution heat transfer coefficients over a large range of flow rates (0.019 0.034 kg/s or 2.5 4.5 lbm/min), concentrations (5 - 40%) and pressures (150 500 kPa or 21.8 72.5 psi). Care was taken to develop a facility that would allow measurements of fundamental phenomena such as heat transfer coefficients in a specific component (the absorber) while a full scale absorption system was operating over a wide range of operating conditions. Numerous challenges presented by the requirement to conduct tests over this wide range of conditions were addressed by developing supplemental heating and cooling loops, auxiliary heat sources and sinks, bypass lines and multiple expansion valves, redundant flow rate measurements, and multiple energy balances to validate the computation of each individual state point and process. These features were included in the test facility as needed, particularly when testing was conducted at the extreme conditions of the test matrix. Care was also taken throughout the study to not only establish the desired conditions, but also to maintain the solution-side thermal resistance as the governing resistance so that absorption heat and mass transfer phenomena could be measured accurately. For some of the desired conditions, when the exact nominal conditions could not be obtained, conditions as close to the nominal values as possible were established. Measured quantities such as temperatures, pressures and flow rates at numerous locations around the test loop were analyzed to obtain absorber heat duties, overall heat transfer coefficients and solution heat transfer coefficients for the various test conditions. The trends in heat transfer coefficients were plotted and discussed, highlighting the effect of solution flow rate, with preliminary interpretations of the effects of solution concentration and absorber pressure on the solution heat transfer coefficients.

For the range of experiments conducted, the absorber heat duty varied from 3.11 to 10.2 kW, the overall heat transfer coefficient varied from 476 to 942 W/m2-K, and the solution heat transfer coefficient varied from 558 to 1140 W/m2-K, depending on the test condition. The absorber heat duty, the overall heat transfer coefficient and the solution heat transfer coefficient increased with increasing solution flow rate. The effects of absorber pressure and solution concentration are less pronounced and harder to discern through simple single-point analyses such as those discussed in this report.

52

WORK PLANNED FOR PHASE II This report documents the experimental phase of the overall project on ammonia-water absorption heat and mass transfer. All the 36 test conditions in the work plan have been tested at values as close to the nominal conditions as possible, and the data analyzed. During Phase II of the project, attention will turn toward the development of a heat and mass transfer model for falling-film ammonia-water absorption around horizontal tube banks. The models developed during the second phase of the study will be validated using the data presented in the present report. The work reported here and that planned in Phase II will result in experimentally validated models for the absorption process for the geometry under consideration, which will assist the design of falling-film ammonia-water absorbers.

53

REFERENCES

Abdelmessih, A. N., Rabas, T. J. and Panchal, C. B. (1997), "Reflux Condensation of Pure Vapors With and Without a Noncondensable Gas Inside Plain and Enhanced Tubes, p. 227. Beutler, A., Ziegler, F. and Alefeld, G. (1996), "Falling Film Absorption with Solutions of a Hydroxide Mixture," International Ab-Sorption Heat Pump Conference, Montreal, Canada, pp. 303-309. Christensen, R. N., Garimella, S., Kang, Y. T. and Garrabrant, M. A. (1998). Perforated-Fin Heat and Mass Transfer Device. USA. 5,704,417. Churchill, S. W. (1977a), "Comprehensive Correlating Equations for Heat, Mass and Momentum Transfer in Fully Developed Flow in Smooth Tubes," Industrial & Engineering Chemistry, Fundamentals, 16 (1): 109-116. Churchill, S. W. (1977b), "Friction-Factor Equation Spans all Fluid-Flow Regimes," Chemical Engineering Progress, 84 (24): 91-92. Colburn, A. P. and Drew, T. B. (1937), "The Condensation of Mixed Vapours," AIChE Transactions, 33: 197-212. Dence, A. E., Nowak, C. C. and Perez-Blanco, H. (1996), "A Novel GAX Heat Exchanger for Cooling Applications," International Ab-Sorption Heat Pump Conference, Montreal, Canada, pp. 595-600. DeVault, R. C. and Marsala, J. (1990), "Ammonia-Water Triple-Effect Absorption Cycle," ASHRAE Transactions, 96 (1). Engler, M., Grossman, G. and Hellmann, H. M. (1996), "Comparative Simulation and Investigation of Ammonia-Water Absorption Cycles for Heat Pump Applications," International Ab-sorption Heat Pump Conference, Montreal, Canada, pp. 209-219. Garimella, S. (1999), "Miniaturized Heat and Mass Transfer Technology for Absorption Heat Pumps," Proceedings of the International Sorption Heat Pump Conference, Munich, Germany, pp. 661-670. Garimella, S. (2000), "Microchannel Components for Absorption Space-conditioning Systems," ASHRAE Transactions, 106 (1): 453-462. Garimella, S. (2004). Method and means for miniaturization of binary-fluid heat and mass exchangers. USA. 6,802,364. Garimella, S., Christensen, R. N. and Lacy, D. (1996), "Performance evaluation of a generatorabsorber heat-exchange heat pump," Applied Thermal Engineering, 16 (7): 591-604. Garimella, S., Christensen, R. N. and Petty, S. E. (1992), "Cycle description and performance simulation of a gas-fired hydronically coupled double-effect absorption heat pump system," Winter Annual Meeting of the American Society of Mechanical Engineers, Anaheim, CA, USA, ASME, New York, NY, USA, pp. 7-14. Garimella, S., Lacy, D. and Stout, R. E. (1997), "Space-conditioning Using Triple-effect Absorption Heat Pumps," Applied Thermal Engineering, 17 (12): 1183-1197. Garrabrant, M. A. and Christensen, R. N. (1997), "Modeling and experimental verification of a Perforated Plate-Fin Absorber for aqua-ammonia absorption systems," Proceedings of the 1997 ASME International Mechanical Engineering Congress and Exposition, Nov 16-21 1997, Dallas, TX, USA, ASME, Fairfield, NJ, USA, pp. 337-347. Gommed, K. and Grossman, G. (1990), "Performance Analysis of Staged Absorption Heat Pumps Water-Lithium Bromide Systems," ASHRAE Transactions, 96 (1): 1590 - 1598. 54

Grossman, G., Wilk, M. and DeVault, R. C. (1994), "Simulation and performance analysis of triple-effect absorption cycles," Proceedings of the ASHRAE Winter Meeting, New Orleans, LA, USA, ASHRAE, Atlanta, GA, USA, pp. 452-462. Grossman, G., Zaltash, A. and DeVault, R. C. (1995), "Simulation and performance analysis of a four-effect lithium bromide-water absorption chiller," Proceedings of the 1995 ASHRAE Annual Meeting, Chicago, IL, USA, ASHRAE, Atlanta, GA, USA, pp. 1302-1312. Herbine, G. S. and Perez-Blanco, H. (1995), "Model of an ammonia-water bubble absorber," Proceedings of the 1995 ASHRAE Annual Meeting, Jan 29-Feb 1 1995, Chicago, IL, USA, ASHRAE, Atlanta, GA, USA, pp. 1324-1332. Higbie (1935), "The Rate of Absorption of a Pure Gas Into a Still Liquid During Short Periods of Exposure," Transactions of the AICHE, 31: 365-389. IPCC (1995). Climate Change 1995: Second Assessment Report of the Intergovernmental Panel on Climate Change, Intergovernmental Panel on Climate Change. Ivester, D. N. and Shelton, S. V. (1994), "Varying heat exchanger parameters in the triple-effect absorption cycle," Proceedings of the International Absorption Heat Pump Conference, New Orleans, LA, USA, ASME, New York, NY, USA, pp. 243-250. Jeong, S., Koo, K.-K. and Lee, S. K. (1998), "Heat transfer performance of a coiled tube absorber with working fluid of ammonia/water," Proceedings of the 1998 ASHRAE Winter Meeting. Part 2 (of 2), Jan 18-21 1998, San Francisco, CA, USA, ASHRAE, Atlanta, GA, USA, pp. 1577-1583. Kang, Y. T., Chen, W. and Christensen, R. N. (1997), "A Generalized Component Design Model by Combined Heat and Mass Transfer Analysis in NH3-H2O Absorption Heat Pump Systems," ASHRAE Transactions: Symposia: 444-453. Kang, Y. T. and Christensen, R. N. (1994), "Development of a counter-current model for a vertical fluted tube GAX absorber," Proceedings of the International Absorption Heat Pump Conference, Jan 19-21 1994, New Orleans, LA, USA, ASME, New York, NY, USA, pp. 7-16. Kang, Y. T. and Christensen, R. N. (1995), "Combined heat and mass transfer analysis for absorption in a fluted tube with a porous medium in confined cross flow," Proceedings of the 1995 ASME/JSME Thermal Engineering Joint Conference. Part 1 (of 4), Mar 19-24 1995, Maui, HI, USA, ASME, New York, NY, USA, pp. 251-260. Kang, Y. T., Kashiwagi, T. and Christensen, R. N. (1998), "Ammonia-water bubble absorber with a plate heat exchanger," Proceedings of the 1998 ASHRAE Winter Meeting. Part 2 (of 2), Jan 18-21 1998, San Francisco, CA, USA, ASHRAE, Atlanta, GA, USA, pp. 1565-1575. Killion, J. D. and Garimella, S. (2001), "A critical review of models of coupled heat and mass transfer in falling-film absorption," International Journal of Refrigeration, 24 (8): 755797. Killion, J. D. and Garimella, S. (2003), "A review of experimental investigations of absorption of water vapor in liquid films falling over horizontal tubes," HVAC and R Research, 9 (2): 111-136. Klein, S. A. (2003). Engineering Equation Solver. Middleton, WI, F-Chart Software. McGahey, K. R., Garimella, S., Cook, F. B. and Christensen, R. N. (1994), "Enhancement of the ORNL absorption model and simulation of a double-effect absorption heat pump," Proceedings of the International Absorption Heat Pump Conference, New Orleans, LA, USA, ASME, New York, NY, USA, pp. 141-148.

55

Meacham, J. M. and Garimella, S. (2002a), "Experimental Demonstration of a Prototype Microchannel Absorber for Space-Conditioning Systems," International Sorption Heat Pump Conference, Shanghai, China, pp. 270-276. Meacham, J. M. and Garimella, S. (2002b), "Miniaturized Shell-and-Tube Heat and Mass Exchangers for Absorption Heat Pumps," 12th International Heat Transfer Conference, Grenoble, France Meacham, J. M. and Garimella, S. (2003), "Modeling of local measured heat and mass transfer variations in a microchannel ammonia-water absorber," ASHRAE Transactions, 109 (1): 412-422. Meacham, J. M. and Garimella, S. (2004), "Ammonia-water absorption heat and mass transfer in microchannel absorbers with visual confirmation," ASHRAE Transactions, 110 (1): 525532. Merrill, T., Setoguchi, T. and Perez-Blanco, H. (1994), "Compact bubble absorber design and analysis," Proceedings of the International Absorption Heat Pump Conference, Jan 1921 1994, New Orleans, LA, USA, ASME, New York, NY, USA, pp. 217-223. Merrill, T. L. and Perez-Blanco, H. (1997), "Combined heat and mass transfer during bubble absorption in binary solutions," International Journal of Heat and Mass Transfer, 40 (3): 589-603. Merrill, T. L., Setoguchi, T. and Perez-Blanco, H. (1995), "Passive heat transfer enhancement techniques applied to compact bubble absorber design," Journal of Enhanced Heat Transfer, 2 (3): 199-208. Perez-Blanco, H. (1988), "A Model of an Ammonia-Water Falling Film Absorber," ASHRAE Transactions, 94 (1): 467-483. Potnis, S. V., Anand, G., Gomezplata, A., Erickson, D. C. and Papar, R. A. (1997), "GAX component simulation and validation," Proceedings of the 1997 ASHRAE Winter Meeting, Jan 26-29 1997, Philadelphia, PA, USA, ASHRAE, Atlanta, GA, USA, pp. 454-459. Price, B. C. and Bell, K. J. (1974), "Design of Binary Vapor Condensers Using the ColburnDrew Equations," AIChE Symposium Series - Heat Transfer - Research and Design, 70 (138): 163-171. Ruckenstein, E. and Berkente, C. (1968), "Mass Transfer to Falling Liquid Films at Low Reynolds Numbers," International Journal of Heat and Mass Transfer, 11: 743-753. Ryan, W. (2002), "New Developments in Gas Cooling," ASHRAE Journal: 23-28. Takuma, M., Yamada, A. and Matsuo, T. (1993), "Condensation Heat Transfer Characteristics of Ammonia-Water Vapor Mixture on Tube Bundles," Condensation and Condenser Design, ASME, pp. 207-217. Ziegler, F. and Alefeld, G. (1994), "Comparison of multi-effect absorption cycles," Proceedings of the International Absorption Heat Pump Conference, New Orleans, LA, USA, ASME, New York, NY, USA, pp. 257-264.

56

APPENDIX A OVERALL, COOLANT, AND SOLUTION HEAT TRANSFER COEFFICIENTS

57

Figure A.1. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 5%, 150 kPa (21.8 psi)

58

Figure A.2. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 5%, 345 kPa (50.0 psi)

59

Figure A.3. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 5%, 500 kPa (72.5 psi)

60

Figure A.4. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 15%, 150 kPa (21.8 psi)

61

Figure A.5. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 15%, 345 kPa (50.0 psi)

62

Figure A.6. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 15%, 500 kPa (72.5 psi)

63

Figure A.7. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 25%, 150 kPa (21.8 psi)

64

Figure A.8. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 25%, 345 kPa (50.0 psi)

65

Figure A.9. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 25%, 500 kPa (72.5 psi)

66

Figure A.10. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 40%, 150 kPa (21.8 psi)

67

Figure A.11. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 40%, 345 kPa (50.0 psi)

68

Figure A.12. Overall, Coolant, and Film Heat Transfer Coefficients @ Nominal Conditions of 40%, 500 kPa (72.5 psi)

69