1、AN-04-7-4 Ammonia-Water Absorption Heat and Mass Transfer in Microchannel Absorbers with Visual Confirmation J. Mark Meacham Student Member ASHRAE ABSTRACT An experimental investigation of an ammonia-water absorber that utilizes microchannel tube arrays was conducted. Liquid ammonia-water solution f
2、lows in the fall- ing-jlm mode around an array of small diameter coolant tubes, while vapor$ows upward through the tube array coun- tercurrent to the falling Jilm. Previous investigations of an absorber designedfor use in a 10.55 kW cooling load (19.28 kW absorber load) residential heat pump demonst
3、rated the potential of this technology for achieving high heat and mass transfer rates with no surface treatment or enhancement and with relatively low solution and coolant pressure drops. Howevel; these previous integrated analytical and experimen- tal studies indicated that solution distribution p
4、roblems might be resulting in incomplete utilization of the provided surface area. In the present study, an absorber with optical access was constructed so that various improvements to the original absorber, including particularly the solution distribution mechanism, could be evaluated. Experiments
5、covering a wide range of solution and coolant flow rates and vapor fractions were used to determine the overall and solution-side heat and mass transfer coeficients. It was found that although the surface area of this improved absorber was only 0.456 m2, approximately 30% of the surface area of the
6、original proto- type absorber, it was able to transfer duties as high as 15.1 kit: almost equaling the load of the original larger absorber. This signijicant increase in performance is attributed to the substantially improved flow distribution. Visual documenta- tion of theflow in this absorber also
7、 conjrmed the signiJcantly improved flow distribution and higher participation of the surface in the heat and mass transferprocess. Srinivas Garimella, Ph.D. Member ASHRAE INTRODUCTION Absorption space-conditioning systems are environmen- tally benign alternatives to vapor compression systems and ha
8、ve the potential for high coefficients of performance (COPs) through the use of increasingly complex thermodynamic cycles. However, the increased complexity necessary to achieve high theoretical COPs requires that numerous heat exchangers and control systems be incorporated in the system. A lack of
9、practically feasible and economically viable compact heat exchangers prevents the realization of the performance potential of these advanced cycles. This limita- tion becomes increasingly significant while attempting to implement absorption technology in the small capacity resi- dential and light-co
10、mmercial market where compact heat exchangers are essential. The adoption of absorption technol- ogy as a practical space-conditioning alternative at these low capacities continues to be hindered by the challenges facing the development of such heat and mass exchangers. The success of the entire abs
11、orption cycle depends on the design of the absorber, which has been referred to as the “bottleneck“ in the heat pump (Beutler et al. 1996a). However, available designs for absorbers have either high thermal resis- tances on the coolant side, poor solution distribution, mass flux limitations due to f
12、looding concerns in counterflow, or, in the case of forced-convective absorption inside tubes, large ammonia-water pressure drops that reduce saturation temper- atures and, thus, the driving temperature differences. An understanding of heat and mass transfer in absorbers with volatile absorbents, es
13、pecially with experimental validation, has also been lacking. A comprehensive survey of falling-film absorption processes and models appears in the recent paper by Killion and Garimella (2001), although much of the atten- J. Mark Meacham is a research assistant and Srinivas Garimella is an assistant
14、 professor at the Georgia Institute of Technology, Atlanta, Georgia. Q2004 ASHRAE. 525 tion of such efforts seems to be directed toward LiBr/H20 systems. Beutler et al. (1996a, 1996b) and Hoffmann and Ziegler (1996) conducted experimental studies on NaOH/ KOH mixtures, LiBr/H20, and NH,/H,O films ov
15、er horizon- tal tube banks and vertical tubes to show that the solution does not fully cover the tube surface of smooth tubes and tends to flow as rivulets corresponding to distributor locations. Jeong fin inserts. Some of the concepts presented above have yielded high heat and mass transfer rates i
16、n commercial applications, but while designs for use in residential systems must have favorable heat and mass transfer rates, they must necessarily attain these rates with simple and compact geometries. ABSORBER CONFIGURATION AND DESIGN and Garimella (2002) addressed this lack of complete wetting by
17、 introducing a wetting ratio for LiBr/H,O absorbers that successfully predicted data from other investigators and also provided insights into the relative significance of absorption in the falling-film and droplet formation and fall modes, respec- tively. For horizontal-tube, falling-film ammonia-wa
18、ter absorbers, Perez-Blanco (1988) found that at typical operating conditions, the absorption rate is controlled by the mass trans- fer process in the falling film, with all other factors having negligible effects. Herbine and Perez-Blanco (1 995) modeled the absorption process in an ammonia-water v
19、ertical tube bubble absorber. Unlike previous models, their analysis was able to account for water desorption from the solution in some portions of the absorber due to the prevailing concentration gradients. Potnis et al. (1 997) used a generalized approach for GAX component and system simulation an
20、d reported that although the mass transfer resistance resides primarily in the vapor phase, the liquid-phase mass transfer resistance should not be considered negligible for an ammonia-water system. An experimental study of a falling film over a coiled tube ammonia-water absorber was conducted by Je
21、ong et al. (1 998), who found that the heat transfer coefficient of the fall- ing film increased linearly with the solution flow rate both with and without absorption. Attempts at obtaining compact ammonia-water absorber geometries have included countercurrent vertical fluted-tube absorbers (Kang an
22、d Christensen 1994) and bubble absorbers for GAX systems (Merrill et al. 1994, 1995). Merrill et al. (1 994,1995) used numerous passive enhancement techniques, such as repeated roughness elements, internal spacers, and increased thermal conductivity metal, to improve heat trans- fer, and mass transf
23、er improvement was achieved through the use of static mixers, variable cross-sectional flow areas, and numerous vapor injector designs. Merrill and Perez-Blanco (1 997) investigated ammonia-water bubble absorption in a compact absorber in which the interfacial area per unit volume of vapor and the l
24、iquid mixing at the vapor-liquid interface were increased by breaking the vapor up into small bubbles and injecting them into the liquid. Garrabrant and Christensen (1 997) analyzed a corrugated and perforated fin surface placed between rectangular coolant channels (Christensen et al. 1998) for ammo
25、nia-water absorbers. In this design, the vapor flows upward through perforations in the corrugated fins, while absorbent solution flows downward over the corrugated fins and through the perforations. Kang et al. (1 998) evaluated the heat and mass transfer resistances in both the liquid and vapor re
26、gions in a countercurrent ammonia-water bubble absorber composed of a plate heat exchanger with offset strip This paper presents the results of the experimental inves- tigation of an ammonia-water absorber utilizing a miniatur- ization technology for heat and mass exchangers. The design and construc
27、tion of this absorber are based upon a concept originally developed by Garimella (1999). The advantages of this concept are detailed in that paper. Experiments conducted on the first prototype for this concept were reported by Meacham and Garimella (2002), and a detailed analysis of the local variat
28、ions in heat and mass transfer coefficients along the absorber was presented in Meacham and Garimella (2003). Garimella (1 999) provides a detailed illustration of the concept for the novel absorber, including the hydronic fluid and solution flow orientations. The original design consists of short l
29、engths of microchannel tubes placed in multiple square arrays, with successive arrays being oriented transversely perpendicular to the adjacent arrays. The number of tube arrays in the absorber is determined by design duty require- ments. Hydronic fluid flows in parallel through individual tubes in
30、the arrays from the bottom to the top of the absorber countercurrent to the falling solution. Due to the small diam- eter of the tubes, this configuration allows for extremely high coolant-side heat transfer coefficients approaching 2500 W/ m2-K even when the flow is laminar. Coolant-side pressure d
31、rops are maintained within the required limits by dividing the coolant flow through multiple tube arrays. Vapor introduced at the bottom of the absorber flows upward countercurrent to the gravity-driven falling solution. This flow arrangement leads to effective vapor-solution contact, minimizing hea
32、t and mass transfer resistances. The low solution-side heat transfer resis- tance coupled with the high tube-side heat transfer coeficient allows the heat of absorption to be transferred to the coolant with low thermal resistances. The heat and mass transfer model developed by Garimella (1 999) was
33、used to estimate the size of the absorber required for a single-effect 10.55 kW (36,000 Bhuh) cooling load resi- dential heat pump, i.e., an absorption load of about 19.28 kW (66,000 Bhuh), and a prototype absorber was fabricated based on these results. While the overall experimental results present
34、ed by Meacham and Garimella (2002) indicate this concept?s capability to transfer heat duties representative of residential heat pumps in a small envelope, the performance was somewhat lower than that predicted by the preliminary model of Garimella (1 999). In addition, modeling of the local measure
35、d heat and mass transfer variations within the proto- type absorber (Meacham and Garimella 2003) indicated that only 20% to 30% of the absorber surface area might have been effectively utilized. From these results, it was inferred that solution distribution problems at the top of the array might be
36、526 ASH RAE Transactions: Symposia Figure I Visualization absorber, (a) external view, (b) view of the internal tube array, and (c) dripper tray. Tube inner diameter, mm Tube wall material responsible for the potential solution maldistribution. However, it was clear that the absorber performance was
37、 solu- tion-side limited, due to the relatively small influence of the coolant flow rate on absorber performance. An improved configuration of the absorber was developed based on these inferences, primarily with the goal of addressing solution-side flow distribution and surface wetting issues. In ad
38、dition, opti- cal access to the absorption process was provided to enable confirmation of the improvement in flow distribution. Testing and analysis ofthe performance of the revised absorber, shown in Figure 1, forms the subject of the current work. The absorber investigated in this study is housed
39、in a flanged stainless steel 0.264 m ID schedule 10 shell with a large sight port as shown in Figure la. This sight port is the primary access for the documentation of the solution flow. Three additional smaller sight ports at 90“ intervals are provided in the shell to enable viewing from different
40、orien- tations (axial and transverse) as well as for the illumination of the absorber. The microchannel tube array is placed in this housing. The array (Figure lb) consists of 1.575 mm OD, 0.2 rnm wall tubes, sandblasted to facilitate surface wetting, with the straight section of each tube segment b
41、eing O. 137 m long. There are 20 rows of tubes, with each row consisting of 33 tubes. Other details about the absorber geometry are provided in Table 1. It should be noted that in the first prototype absorber investigated by Meacham and Garimella (2002, 2003), successive tube arrays were oriented tr
42、ansversely perpendicular to the adjacent arrays. However, in the absorber under consideration in the present work, that arrangement would prevent optical access due to the headers required for coolant inlet and outlet in both perpendicular directions. Therefore, in this absorber, the tubes in adjace
43、nt arrays are oriented in the same direction as shown in Figure Ib. With such an arrangement, the solution flow and fluid distribution along the tube length can be viewed through the large sight port without obstruction by headers. Series flow between passes of the absorber is achieved by bending th
44、e tubes instead of using headers. (With the U-bends at the ends of the tubes, 1 .O67 Stainless Steel Table 1. Visualization Absorber Geometry Number of tubes per row Number of rows per pass Number of oasses I Tube outer diameter. mm I 1.575 I 33 2 10 Tube transverse pitch, 111111 Row vertical pitch,
45、 mm 1 Tube length, m I 0.137 I 4.76 7.94 1 Total surface area, m2 0.456 Absorber height, m I 0.150 I ASHRAE Transactions: Symposia 527 these two streams to enter the absorber at different locations. The dilute solution is cooled as necessary to ensure that a single-phase liquid flows through the dil
46、ute solution mass flow meter. This stream enters the absorber and is distributed over the coolant tubes through the drip tray described above. The vapor stream enters the absorber at the bottom of the tube array and flows upward countercurrent to the dilute solution. The heat of absorption is remove
47、d by the coolant flowing upward within the tube array. The exiting concentrated solu- tion is further subcooled as necessary in a small shell-and-tube heat exchanger before returning to the solution pump. The coolant loop consists of a circulation pump and a plate heat exchanger, where the heat of a
48、bsorption removed at the absorber is rejected to a city water stream. City water and cool- ing water flow rate variations provide the requisite control of the absorber heat sink. Temperature, pressure, and mass flow rate measurement locations are also shown in Figure 2. Coriolis mass flow meters (*O
49、. 15% uncertainty) measure the concentrated and dilute solution flow rates. The concentrated solution flow rate is controlled by a variable-speed solution pump, while the vapor generation rate is varied by increasing or decreasing the saturation pressure of the steam in the desorber. This vapor generation rate is calculated by taking the difference between the measured concentrated and dilute solution flow rates. The closed-loop cooling water flow rate is measured using a magnetic flow meter with an uncertainty of *OS%. Solution and vapor inlet and outlet pressures are measured using abs
