- Email: [email protected]
Horizontal tube bundle falling film distiller for ammonia–water mixtures
international journal of refrigeration 59 (2015) 304–316
Available online at www.sciencedirect.com
ScienceDirect j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / i j r e f r i g
Horizontal tube bundle falling film distiller for ammonia–water mixtures E.W. Zavaleta-Aguilar a,b, J.R. Simões-Moreira a,b,* a b
IEE – Institute of Energy and Environment, University of São Paulo, São Paulo, SP, Brazil SISEA- Alternative Energy Systems Lab., Escola Politécnica, University of São Paulo, SP, Brazil
A R T I C L E
I N F O
A B S T R A C T
Article history:
This work presents an experimental study of an important component of an ammonia–
Received 23 April 2015
water absorption refrigeration cycle based on the falling film technology – the distiller, which
Received in revised form 4 July 2015
is formed by two parts: the generator and the rectifier. It was carried out a parametrical
Accepted 15 July 2015
study on the operation of the distiller by testing several parameters. The experimental results
Available online 23 July 2015
showed that the distilled ammonia vapor concentration always degraded as rectifier, generator, and concentrated solution temperatures increased and improved as concentrated
Keywords:
solution mass flow rates and concentrated solution concentrations increased. The highest
Ammonia–water
distilled ammonia vapor concentration obtained was of 0.9974. Photographic documenta-
Falling film
tion of flow pattern over the generator tube bundle was also obtained resulting in both jet
Distillation
column and sheet patterns. Finally, a Nusselt number correlation for the ammonia–water
Absorption
falling film over the generator tube bundle was proposed based on the experimental results and compared with other authors’ correlations. © 2015 Elsevier Ltd and International Institute of Refrigeration. All rights reserved.
Distillateur à faisceau de tubes horizontaux à film tombant pour des mélanges ammoniac-eau Mots clés : Ammoniac-eau ; Film tombant ; Distillation ; Absorption
1.
Introduction
Absorption refrigeration cycles (ARC) are promising candidates for diminishing the dominating dependence on regular mechanical vapor compression refrigeration cycles (CRC) in the air conditioning and refrigeration industries. Their distinct advantage over CRC relies on the fact that they can be powered by either waste and recovered heat sources or be directly
powered by solar energy at none or low electrical energy consumption. They can also operate in a combined heat and power (CHP) or cogeneration configuration increasing the final usage of the fuel chemical energy content used to power a thermal machine, i.e., increasing the energy utilization factor (EUF). On the other hand, absorption refrigeration cycles are usually heavier, larger, and have a higher capital expenditure than standard compression refrigeration cycles at the same nominal capacity. They are larger and heavier due to the high number
* Corresponding author. Escola Politécnica at University of São Paulo, Mechl. Eng. Dept, SISEA – Alternative Energy Systems Lab., Av. Prof. Mello Moraes, 2231, 05508-900 São Paulo, SP, Brazil. Tel.: +55 11 30915684; Fax: +55 11 30915684. E-mail address: [email protected] (J.R. Simões-Moreira). http://dx.doi.org/10.1016/j.ijrefrig.2015.07.022 0140-7007/© 2015 Elsevier Ltd and International Institute of Refrigeration. All rights reserved.
international journal of refrigeration 59 (2015) 304–316
305
Nomenclature a A b COP cp d g Ga F k LMTD m n Nu Pr q Re SF T TR U x y
constant [–] area [m2] constant [–] coefficient of performance [–] specific heat at constant pressure [J kg−1 °C−1] diameter [m] acceleration of gravity [m s−2] modified Galileo number, Ga = ργ3/µ4g [–] correction factor [–] thermal conductivity [W m−1 °C−1] logarithmic mean temperature difference [°C] mass flow rate [kg s−1] constant [–] Nusselt number, Nu = α(v2/k3g)1/3 [–] Prandtl number, Pr = cpµ/k [–] heat transfer rate [W] Reynolds number, Re = 4Γ/µ [–] shape factor [–] temperature [°C] tons of refrigeration [3517 W] overall heat transfer coefficient [W m−2 °C−1] ammonia–water concentration – liquid [kgammonia kgsolution−1] ammonia–water concentration – vapor [kgammonia kgsolution−1]
Greek symbols α heat transfer coefficient [W m−2 °C−1] γ surface tension [N m−1]
of heat and mass exchangers employed to increase its efficiency. ARC are commercially available in two technologies. The first one is based on lithium bromide–water mixtures, and the second one on ammonia–water mixtures. The former can be applied only to air conditioning systems, as the fluid refrigerant is plain water. However, that cycle has a higher coefficient of performance (COP). Simple effect ARC have typical COP of 0.6–0.7 for lithium bromide cycles and 0.5 for ammonia– water cycles (ASHRAE, 1994), higher COP can be obtained in multi-effects cycles (Srikhirin et al., 2001). The second one can be used to either subzero or above zero degree Celsius applications, as the fluid refrigerant is ammonia. Furthermore, the ammonia–water cycle demands smaller heat exchangers, it operates at above atmospheric pressure and there is no any crystallization concern (Herold et al., 1996). In the evaporation process of the ammonia–water mixture, a small fraction of water is always present in the vapor phase dominated by ammonia. Thus, the wet vapor ammonia produced in the generator section must be purified to get rid of its water content, since liquid water can accumulate and increase the evaporator temperature, a phenomenon known as temperature glide. In accordance with such an operational problem, Fernádez-Seara and Sieres (2006) studied theoretically the consequences of exiting high amounts of water from the distiller in ammonia–water absorption refrigeration cycles and concluded that water content in wet ammonia is accountable for the low COP in experimental and industrial AARC.
Γ
Δ ε μ v ρ σ
mass flow rate per axial unit length of tube (each side) [kg m−1 s−1] difference [–] emissivity [–] dynamic viscosity [kg m−1 s−1] kinematic viscosity [m2 s−1] density [kg m−3] Stefan–Boltzmann constant [W m−2 K−4]
Subscripts c correlation conv convection cs concentrated solution d distilled ds diluted solution f falling film g generator in inside m average o oil out outside r rectifier rad radiation s saturation v vapor vess vessel w wall
The two key components of ammonia–water absorption refrigeration cycles (AARC) design and operation are the absorber and the distiller, wherein the former one has been the most studied (Determan and Garimella, 2011). The distiller must provide pure or at a very high purity degree of ammonia vapor to operate the cycle. Small commercial ammonia–water absorption refrigeration cycles (5 TR) have flooded generators (also known as desorbers), in which wet ammonia vapor is produced from a liquid solution heated by hot flue gases. The wet ammonia vapor is driven to a tray filling type analyzer, also known as rectification column or fractionation column (Herold et al., 1996; Merrick, 1972). After that, the ammonia vapor is directed to the rectifier to, finally, reach the condenser where a high degree of purity has been reached. According to Merrick (1972), the tray type distiller is generally the largest component in an AARC. This is due to the use of the analyzer and a gas burner placed below the generator, in his work the gas burner was placed at side of the vertical generator to decrease its size. Previous works on the distiller are Anand and Erickson (1999), who analyzed theoretically a distillation column for an 8TR AARC. Zavaleta-Aguilar and Simões-Moreira (2012) carried out a thermal modeling analysis and design of a holed tray distillation column for a 5 TR AARC. Those authors carried out a parametric analysis of the liquid and vapor ammonia–water mass flow rates, concentration, and temperature along with a study over geometrical parameters influence such as plate hole and column
306
international journal of refrigeration 59 (2015) 304–316
Fig. 1 – (a) Packed column (adapted from Sieres et al., 2009), (b) spray column (adapted from Mendes, 2008).
diameters. The authors do not acknowledge previous experimental studies of tray type distillers in open literature. Sieres et al. (2009) compared experimentally commercial random and structured filling types (Fig. 1a) such as: Pall® rings 10 mm (stainless steel), Berl® saddles 1/2” (ceramic), Raschig® rings 15 mm (glass), Novalox® saddles 1/2” (ceramic), Mellapakplus® 752.Y (stainless steel) and SulzerBX® and found that the first (Pall® rings) needed less column height to produce the same degree of ammonia purification than the other fillings. The generator used was a coil heated by thermal oil. On the other hand, according to Starace and De Pascalis (2012) in diffusion absorption refrigerator (DAR) cycles, the vapor leaving the rectifier is rarely pure ammonia, moreover, the COP in these systems increases as a consequence of the decreasing con-
centration (and the increasing temperature) of the ammonia vapor leaving the rectifier. Ahachad et al. (1992) compared analytically the performance of two simple effect ARC. In the first one, wet ammonia vapor came out from the generator and it was driven directly to the condenser without any rectification column. In the second study, it used a bubble column rectifier to purify the wet ammonia produced by the generator. The results showed that the COP was improved by 35% by the presence of the bubble column rectifier. However, even with that system the COP obtained was low, around 0.4, as the highest value. That happen most probably due to the low ammonia concentration (yd = 0.95) leaving the bubble column system. No further geometry detail of the generator was explained. Mendes (2008) studied experimentally two wet ammonia purification systems for an ammonia–water absorption refrigeration cycle. The first system (also studied by Mendes et al., 2007) was a spray column type (Fig. 1b), where the concentrated solution was sprayed in the wet ammonia vapor by a 45° angle spray nozzle.The descending liquid spray droplets came into contact with the counterflow wet ammonia vapor produced in the generator. The second purification system was an analyzer packed with Novalox® 1/2” saddle filling.The author obtained ammonia vapor concentration from 0.9754 to 0.9903 for the spray column and from 0.9825 to 0.9912 for the packed column. It used a plate heat exchanger as generator in their work. Determan and Garimella (2011) studied experimentally a falling film generator for an ammonia–water absorption refrigeration cycle with capacity up to 5 TR. The generator was formed by 304 stainless steel horizontal tube bundle of 1.067 mm i.d., 1.575 mm o.d., and 140 mm length. Their generator was divided into five passes, each pass had 16 rows, each row was oriented at 90° of the other and each row had 27 columns. The generator had an overall geometry of 178 mm × 178 mm × 508 mm. The authors obtained an average wet area of 45%. It used a water–glycol mixture as the heating fluid in the generator operating at an inlet temperature around 100 °C.The two tested concentrations of the concentrated solution were 0.44 and 0.48.The pressure in the generator was 6.7 bar. Finally, they obtained 0.9873 wet ammonia concentration at the generator outlet. A similar concentration was obtained by Zavaleta-Aguilar and Simões-Moreira (2012) at the enrichment section output. This would indicate that it is not necessary to use the analyzer (stripping and enrichment sections) above the generator, when the falling film technology is applied for generators. On the other hand, the COP is higher when partial condensation rectifiers were used than the ones based on complete condensation (Fernádez-Seara et al., 2003). Therefore, in this work we chose generator/partial condensation rectifier configuration. The present work aims at the novel study and analysis of the generator and the rectifier both of them combined in one single component (the distiller) using the falling film technology for an ammonia–water absorption refrigeration cycle. In order to achieve a distiller parametric analysis, the main controlling magnitudes were independently analyzed, namely: the tubing heating temperature, the rectifier temperature, the concentrated solution temperature, the solution mass flow rate, and the ammonia–water concentration of the liquid solution. By independently controlling those magnitudes it was possible to simulate actual operation conditions and
international journal of refrigeration 59 (2015) 304–316
evaluate their influence on the concentration and mass flow rate of the distilled ammonia vapor. The parametric study consisted in changing the control variables such as solution temperature, generator, rectifier, etc. to verify their influence on the vapor leaving the distiller. Detailed discussion on the distiller operation and its main components is carried out. Particularly, the liquid solution distributor is discussed, as its operation should properly homogenize the falling film over the first tube row and the ones underneath. Finally a heat transfer analysis to evaluate the Nusselt number for the liquid ammonia–water mixture is presented and compared with other authors’ correlations.
2.
Experimental setup
The experimental setup is composed of a main ammonia– water solution circuit, whose ammonia concentration can be properly altered according to a test run. Two other auxiliary circuits were also built. The first one is formed by a heating oil circuit to heat up the ammonia–water solution to the preset testing temperature. The other circuit is formed by a water cooling circuit to rectify the wet ammonia vapor. Each circuit is explained in detail in the following subsections.
2.1.
Ammonia–water testing rig
A schematic diagram of the experimental setup is depicted in Fig. 2. The ammonia–water distilled saturated vapor at high
T p M V
Ammonia-water circuit Heating circuit Water cooling circuit
Termocouple Pressure transmitter Coriolis flow meter Oil flow meter
Water tank
Oil tank
Water air cooler
Water heater
Distilled vapor p
T
Tp
Condenser
RECTIFIER T
Oil heater T
Sub-cooler
T
T T V Oil flow meter
degree of purity exits the distiller and flows to the condenser where it is condensed to in sequence which undergoes a subcooling process in the sub-cooler. The condensed ammonia coming from the sub-cooler is mixed up with the diluted solution exiting the distiller (after passing through the solution cooler). After the mixing process, the liquid (concentrated solution) flows to the tank solution, where the liquid level could be checked by observing a mounted borosilicate glass sight. Next, the concentrated solution is pumped to the distiller by a magnetic driven gear pump controlled by a frequency inverter. Between the distiller and the pump there is an electrical heater (heater solution) PID controlled at the set solution temperature for the test run. An electric resistance was considered for having a precise control of the concentrated solution temperature entering to the distiller. In real systems the concentrated solution is pre-heated in a heat exchanger with the diluted solution leaving the generator. The measurements of the ammonia distilled, concentrated solution, and diluted solution mass flow rates were attained by three Coriolis flow meters (Krohne model: Optimass MFM 33000 C) indicated by the letter “M” in Fig. 2. Besides giving the precise mass flow rates, the flow meters also indicated the liquid density. Thermocouples (T type) were installed at the inlet and outlet of each heat exchanger. Pressure readings were obtained by pressure transmitters (Salcas, 316 stainless steel diaphragm, model: SA-1000), one installed at the distiller top wall and the other on the condenser wall. Furthermore, three other thermocouples (T type) and three pressure transmitters (Salcas, 316 stainless steel diaphragm, model: SA1000) were installed in the measuring stations of the mass flow meters (see Fig. 2). Thus, a complete set of thermodynamic properties (density, pressure, and temperature) of the distilled ammonia, concentrated solution, and diluted solution was directly measured, which allowed one to obtain the corresponding ammonia concentration at the measuring station by using a solving thermodynamic properties routine, such as the one implemented in the Engineering Equation Solver software (Klein, 2014). At the distilled ammonia vapor line, just after the distiller top cover (see Figs. 3 and 4c) it was installed a safety pressure relief valve set to open at internal pressure of 17 bar or higher. The relief valve piping vents to a water tank installed outside the laboratory. All elements of the ammonia– water circuit work at high pressure. The whole circuit was tested at static pressures up to 21 bar for assuring operational safety and a leak free circuit. Two important elements of that circuit, the distiller and the liquid distribution system, are detailed in the following two subsections. The whole test rig was made of stainless steel.
p,T DISTILLER
GENERATOR
M Coriolis flow meter Oil pump
307
Diluted solution
Solution cooler
Concentrated Solution Solution T heater
M
p,T
M
p,T
Mixer p T Solution tank
Fig. 2 – Experimental setup schematics.
2.1.1.
The distiller
Figure 3 shows the schematics of the distiller set. In the lower part, it is illustrated the tube bundle generator and on the top it is seen the tube bundle rectifier. The whole set is placed inside a pressure vessel. We call this distiller as the falling film type, because the concentrated solution of ammonia–water flows around the generator horizontal tube bundle giving rise to a thin liquid film configuration on each tube. Also, the rectifier is built following the same falling film technology, as seen. Figure 4 shows still pictures of parts of the distiller prior to installation. The generator (Fig. 4a) is heated by a heating oil that flows inside the horizontal tube bundle. As the
308
international journal of refrigeration 59 (2015) 304–316
Distilled ammonia vapor
Coolant fluid Concentrated solution Heating fluid
TOP COVER VESSEL
RECTIFIER
Ammonia-water falling film Termocouple Distributor
Dry wall
GENERATOR
concentrated solution flows down around the external heated tube wall, wet ammonia–water vapor is produced. There are 70 tubes of 8 mm o.d. and 1 mm wall thickness. The ammonia– water liquid solution that remains after evaporation is accumulated at the bottom part and it is called diluted solution, because it has less amount of ammonia than that of the initial concentrated solution. The diluted solution exits at the pressure vessel from the bottom. The wet ammonia vapor produced in the generator moves upwards to reach the rectifier section (Fig. 4b). In the rectifier section, the excess of water vapor is removed from the vapor by a cooling process, thereby obtaining distilled ammonia vapor. The rectifier was built having two lateral vertical plates in order to force to wet ammonia vapor produced in the generator section to be put in direct contact with the water cooled rectifier tube bundle. Consequently, the wet ammonia vapor condenses partially also forming a condensing falling film over the rectifier tubing.
2.1.2. Diluted solution
Fig. 3 – Distiller schematics.
The liquid distributor
The liquid distributor is the component which ensures an even distribution of the liquid around the generator tubes both spanwise and circumferentially each tube. A thorough comparison of types of distributors used in a horizontal tube falling film technology is better described in Narvaez-Romo et al. (2014) and Zavaleta-Aguilar (2015). Those authors carried out several tests with plain water to verify the distribution and wettability
Fig. 4 – Still pictures of distiller parts.
international journal of refrigeration 59 (2015) 304–316
309
electrical resistance set at the testing temperature. Next, the heated oil was pumped to the generator and distributed as uniform as possible to the horizontal tube bundle. In order to carry out an energy balance, both the inlet and outlet temperatures were also measured and the oil flow readings were obtained from a volumetric flow turbine meter (“V” in Fig. 2). To avoid any piping stress because of the decrease of the oil density, an atmospheric expansion tank was also installed in the oil circuit (see Fig. 2). A still picture of the experimental setup built showing the circuits and the cycle components described above is shown in Fig. 6.
3.
Fig. 5 – Distributor schematics. of water on the generator tubes and to find out the better geometrical distributor system. In this work, it was implemented the distributor shown in Fig. 5. The distribution system consists of a 3/4” tube branched into 7 tubes of 8 mm o.d. and 6 mm i.d. which have holes at 0° and 60° from the horizontal plane (Fig. 5a), to allow the exit of ammonia–water solution. A 12 mm o.d. tube with 1 mm wall thickness positioned outside the 8 mm o.d. tube directs the solution to its bottom part to a thin blade V-shaped that was welded at the bottom of the 8 mm tube. To ensure flow homogeneity it was placed a tray with nozzles, as shown in Fig. 5b. Water droplet diameter falling on a sanded stainless steel plate from a height of 6 mm corresponds to 23 mm average diameter spreading region as seen in experiments. Therefore, in order to ensure a good tube wall wettability it was adopted a 17 mm nozzle pitch. The distribution system is shown in Fig. 5c.
2.2.
Water cooling circuit
Water was used as the cooling liquid for the rectifier, for the condenser, for the sub-cooler, and for the solution cooler. After cooling those heat exchangers, the water was pumped to a forced air cooler placed outside the laboratory (Fig. 2). Each water sub-circuit had an independent centrifugal pump and the water flow was controlled by needle valves. The cooling water circuit had an electrical heater mounted just preceding the rectifier so that the ammonia vapor purification could be tested at different cooling temperatures.
2.3.
Heating circuit
Synthetic thermal oil was used as the heating fluid in the generator. Oil heating system was attained by a PID controlled
Experimental procedure
Table 1 shows the tested experimental parameters. Two liquid solution concentrations (xcs) were tested whose average values were 0.37 and 0.49. For each concentration it was evaluated the influence of the solution mass flow rate, mcs, spanning from 0.0160 to 0.0276 kg s−1 as well as the influence of two average rectifier temperature, Tr,m, at 62 and 34 °C, taken as the average between the inlet and outlet cooling water temperatures. For the average concentration of 0.37 it was analyzed the influence of the oil temperature, To,m, from 110 to 136 °C taken as the average between the inlet and outlet heating oil temperatures. For the solution concentration of 0.49 the oil temperature range was between 97 and 125 °C. For the concentration of 0.37 it was analyzed the influence of the concentrated solution temperature, Tcs, from 87 to 103 °C and for the concentration of 0.49 the concentrated solution temperature ranged from 69 to 82 °C. The oil and concentrated solution temperature ranges tested for the two concentrations (0.37 and 0.49) were different because their saturation temperatures were also distinct. The total heat flux supplied to the generator (qg) was calculated from the First Law of Thermodynamics applied to the oil flow as given by Eq. (1). The thermal energy flux is transferred to the falling film (qf) and to the generator dry wall (qw) (Eq. 2). The generator dry wall (see Fig. 3) transfers energy by radiation and by convection as indicated in Eq. (3).
qg = mo c p, o ( To,in − To, out )
(1)
qg = q f + qw
(2)
qw = qw, rad + qw, conv
(3)
The energy transferred by radiation (qw,rad) is the energy transferred from the generator dry wall to the inner wall of the vessel bottom section and is calculated by Eq. (4). On the other hand, it was assumed natural convection heat transfer from the dry wall generator to the vapor (qw,conv) (Eq. 5), since we observed that the vapor was mainly still around the generator dry wall. The convective heat transfer coefficient ( α v ) in Eq. (5) was calculated from Cengel and Ghajar (2011) correlation. The generator dry wall (q w ) transferred on average 8% of the total of energy (qg), where 0.3% was by radiation and 7.7% was by convection.
310
international journal of refrigeration 59 (2015) 304–316
Fig. 6 – Still picture of experimental setup built.
qw, rad =
4 σ (Tw4 − Tvess ) 1−ε 1 1−ε + + ε Aw AwSFw − vess ε Avess
qw, conv = α v Aw (Tw − Tv )
(4)
(5)
By leakage concern it was not possible to measure the oil temperature exactly at the tube bundle inlet and outlet. Thus, the logarithmic mean temperature difference (LMTD) evaluation was approached with the generator oil temperature inlet and outlet. This approach seems reasonable since the generator oil temperature inlet and outlet averaged 3.1 °C. The LMTD is calculated by Eq. (6). Since the ammonia–water and oil flows were in cross flow configuration it used the F-factor for correcting the calculation performed as if it were a counterflow configuration. The overall heat transfer coefficient (Uf) concerning to the outer area of the generator tubes (Af) was obtained by Eq. (7). The average heat transfer coefficient ( α f ) and the Nusselt number ( Nu f ) of the falling film were evaluated by Eqs. (8) and (9), respectively. To evaluate the heat transfer coefficient inside the tube ( α in ) it was considered the thermal entrance length in laminar flow (Bejan, 2004). In order to determine the ammonia–water thermophysical properties the relations in M. Conde Engineering (2006) were used.
LMTD =
Uf =
(To,in − Tds ) − (To,out − Tcs ) (T − Tds ) ln o,in (To,out − Tcs )
(6)
qf FA f LMTD
{
1 1 1 ⎛ dout ⎞ dout ⎛ dout ⎞ ln ⎜ = − ⎟ ⎜ ⎟+ αf Uf α in ⎝ din ⎠ 2kw ⎝ din ⎠ ⎛ v2f ⎞ Nu f = α f ⎜ 3 ⎟ ⎝ k f g⎠
(7)
}
(8)
13
4.
Results and discussions
4.1.
Hydraulic analysis
(9)
Figure 7 shows a still picture of the ammonia–water liquid solution distribution. The nozzles of the tray distributor eliminated the liquid spanwise contraction between the distributor and the top generator tubes. The liquid spanwise contraction causes dry areas that can decrease the heat exchange efficiency of a
international journal of refrigeration 59 (2015) 304–316
311
Table 1 – Testing parameter range. Parameter
Unit
Concentrated — solution concentration, xcs Concentrated solution mass flow rate, mcs
kg s−1
°C Average rectifier temperature, Tr,m
°C Average oil temperature, To,m °C Concentrated solution temperature, Tcs
Value 0.37
Observation
0.49
Correspond to evaporation temperature of −13 and 5 °C respectively (Ortigosa et al., 2008) 0.0160–0.0276 Suitable for a 1200 W cooling capacity AARC and a concentrated and dilute concentration difference of about 3% (Ortigosa et al., 2008) 34 and 62 34 °C corresponds approximately to the temperature of a cooling tower in summer season and 64 °C was used just to verify the a hot condition (use of an internal fluid of the cycle) 110–136 97–125 Use of low temperature sources (solar, cogeneration) 87–103 69–82 In AARC the diluted solution can be pre-heated before entering to the generator by using a heat exchanger between the absorber and the generator (Herold et al., 1996). In this work the diluted solution was warmed close to its saturation temperature (see Fig. 12).
heat exchanger (Ribatski, 2006). Those nozzles also allowed to stabilize and to evenly distribute the liquid film. There was no visible spanwise contraction on the tubes and between the tubes because of the small distance between tubes (2 mm). The generator tubes were blasted with glass balls and it was achieved an average roughness of 196 µm. It was noted an increase of the wettability after the blasting. In previous tests without the vessel (see Figs. 3 and 4d) using plain water rather than an ammonia–water mixture,
Fig. 7 – Ammonia–water distribution at: mcs = 0.0276 kg s−1, xcs = 0.3658, Re = 222, 15.4 bar, Tcs = 100.1 °C.
Fig. 8 – Flow pattern between the generator tubes, for Re from 108 to 246.
Fig. 9 – Flow pattern between the tubes compared with that of Hu and Jacobi (1996).
each time that the external tube array surface was wet, it was observed that the inside tube arrays also were wet. Consequently, in actual tests using ammonia–water mixtures it was presumed that the same phenomenon would happen as direct observation would not be possible. However, the conclusion came after observing the two opposite tube arrays of the tube bank (see Fig. 4d). The wettability obtained in this study was 100% and overcame that of Lazcano-Véliz et al. (2014) who obtained 80% of its wet area for similar conditions of mass flow rate (Γ = 0.0080 kg m−1s−1). The distance between tubes used for those authors was 5 mm. It was noticed in the present work that the first dry areas appeared at Re = 140 for xcs = 0.37 and at Re = 100 for xcs = 0.49. The sheet and column flow patterns were observed in all tests, as shown in Fig. 8, where the Reynolds number varied in the range Re = 108–246 (and Galileo number Ga1/4 from 834 to 1249). This result is in opposition to Hu and Jacobi’s (1996), since, according to them, the flow pattern must be of droplets type, as shown in Fig. 9. The sheet and column flow patterns obtained in this work were mainly due to the spacing of 2 mm imposed between the tubes, smaller than the droplet size estimated in 4.2 mm (Zavaleta-Aguilar, 2015).
312
international journal of refrigeration 59 (2015) 304–316
Fig. 10 – A general distillation process.
4.2.
Parameter influence
The temperature-concentration (T–x) diagram in Fig. 10 shows the thermodynamic paths of a distillation process for the ammonia–water solution. There can be seen that the concentrated solution enters the distiller at the generator at liquid sub-cooled condition with temperature Tcs = 100.1 °C and concentration xcs = 0.3777, labeled “1” in the diagram. This solution is heated in the generator tube bundle until its bubble point is reached, when the evaporation process starts off. As the liquid solution receives heat from the heated tubes as it flows around them, the evaporation process goes on, which causes a decrease in liquid concentration and an increase in its temperature. Near the distiller outlet, the concentrated solution has become a diluted solution (at the saturated condition), at temperature of T ds = 108.6 °C and at concentration of xds = 0.3500, labeled as number “2” in the T–x diagram. The wet ammonia vapor produced in the generator (label “3” in the T–x diagram) moves upward into the rectifier section and it is put in contact with the cooling tube bundle, where a part of its vapor is condensed and sent back to the bottom of the generator. It can be noticed in T–x diagram in Fig. 10 that the vapor concentration that leaves the rectifier (label “4”) is higher than the one at the inlet; this means that the vapor condensed in the rectifier contained a portion of water. The distilled vapor obtained is condensed in the condenser/sub-cooler, leaving as sub-cooled liquid (label “5”). Parametric analysis results are shown in Figs. 11, 12, and 13. Figure 11a shows the influence of the distilled ammonia concentration as a function of the solution mass flow rate (mcs) for two average rectifier temperatures (Tr,m = 34 and 62 °C) at two concentrated liquid solution concentrations (xcs = 0.37 and 0.49). In this test the average oil temperature (To,m) was 132 °C and the concentrated solution temperature (Tcs) was 100 °C for the concentrated solution concentration of 0.37; and To,m = 109 and Tcs = 76 °C for the concentrated solution concen-
tration of 0.49. Figure 11a indicates that when the concentrated solution mass flow rate increased, the concentration of ammonia distillate (yd) increased as well. It happened because the larger the concentrated solution mass flow rate, the smaller the thermal heat transfer of the falling film, because the film thickness is thicker. Thus, the amount of the wet ammonia vapor produced in the generator is smaller, nevertheless with higher concentration. The graphics also shows that for xcs = 0.37 at Tr,m = 62 °C the ammonia distilled concentration ranged from 0.9750 to 0.9842. On the other hand, when the rectifier average temperature was 34 °C the concentration yd ranged from 0.9879 to 0.9901. This means that the lower the rectifier temperature the higher the ammonia distilled concentration. Other important conclusion is that the ammonia distilled concentration (yd) is higher for the higher concentrated solution concentration, i.e. xcs = 0.49 reaching up to 0.9974. The distilled mass flow rate (md) decreases as the concentrated mass flow rate increases, m cs , as indicated in Fig. 11b, as a general rule. The distiller vessel pressure varied from 14.1 to 15.4 bar for Tr,m = 62 °C, and from 12.4 to 13.9 bar for Tr,m = 34 °C. The next series of graphs in Fig. 12 shows the influence of the average oil temperature (To,m). The concentrated solution mass flow rate was constant (mcs = 0.0230 kg s−1) as well as the rectifier temperature (Tr,m = 34 °C). The concentrated solution inlet temperature was 100 °C for xcs = 0.37, and 76 °C for xcs = 0.49. In Fig. 12a, one can see that the distilled ammonia vapor concentration (y d ) decreases as the heating oil temperature increases for both concentration values of the concentrated solutions. On the other hand, the distilled ammonia vapor mass flow rate displays the reverse trend regarding the oil heating as seen in Fig. 12b. Finally, it is worthy to mention that the minimum average heating oil temperature to produce a reasonable amount of ammonia vapor was 110 °C for the 0.37 concentrated solution concentration, and 98 °C for xcs = 0.49. Weber et al. (2014) used water at temperatures between 160
international journal of refrigeration 59 (2015) 304–316
Fig. 11 – Influence of the concentrated solution mass flow rate (mcs), the average rectifier temperature (Tr,m), the concentration of the concentrated solution (xcs) on the distilled concentration (yd) and distilled mass flow rate (md). Full line (___) is for xcs = 0.37 and dashed line (--) is for xcs = 0.49. and 200 °C to warm up the generator of a commercial ammonia–water absorption cycle. Tests were carried out to analyze the influence of the concentrated solution inlet temperature considering the degree of sub-cooling, ΔTs , which can be defined as the difference between the solution bubble temperature and the actual concentrated solution inlet temperature. Those results are shown in Fig. 13a and 13b. For these experiments, the average rectifier temperature (Tr,m) was kept at 34 °C and the average heating oil temperature (To,m) was also constant for a given concentrated solution concentration, namely To,m = 109 °C for xcs = 0.49 and To,m = 132 °C for xcs = 0.37. In any situation the concentrated solution mass flow rate was 0.0232 kg s−1. For all cases, as the degree of sub-cooling decreases the distilled ammonia vapor concentration decreases (Fig. 13a), and the distilled ammonia vapor mass flow rate increases (Fig. 13b). The oil mass flow oil remained practically constant in all the tests. In general, when the total heat flux supplied to the
313
Fig. 12 – Influence of the average oil temperature (To,m) and the concentration of the concentrated solution (xcs) on the distilled concentration (yd) and distilled mass flow rate (md). Full line (___) is for xcs = 0.37 and dashed line (--) is for xcs = 0.49.
generator increases the diluted solution concentration decreases as well and its temperature increases and, furthermore, the distilled ammonia vapor mass flow rate increases and its concentration decreases.The system pressure remained constant in each test. This pressure depends basically on the condensator pressure that was controlled by the condenser cooling temperature. Small variations in the system pressure are also achieved by varying the rectifier and generator temperatures. The vapor pressure drop into the distiller was negligible since the vapor pressure drop from the distiller to the sub-cooler (longer and narrower path into the distiller, see Fig. 2) was only 0.8%.
4.3.
Nusselt correlation
It was possible to experimentally obtain average Nusselt numbers (Nuf) for this study data basis as indicated in Fig. 14. Experimental Nusselt numbers were calculated according to
314
international journal of refrigeration 59 (2015) 304–316
Fig. 14 – Experimental results of the ammonia–water falling film average Nusselt number divided by Pr0.4.
Fig. 13 – Influence of the concentrated solution temperature (Tcs) and the concentration of the concentrated solution (xcs) on the distilled concentration (yd) and distilled mass flow rate (md). Full line (___) is for xcs = 0.37 and dashed line (--) is for xcs = 0.49.
Fig. 15 – Comparison of the Nusselt number correlation obtained in this work with many other literature correlations.
possible to correlate the Nuf/Pr0.4 values against aReb, resulting in a = 0.75 and b = −0.27, as indicated by a full line in Fig. 14. The Nusselt correlation final equation is given by Eq. (9).
Nuc = 0.75Re−0.27 Pr 0.4 Eq. (9). Ammonia–water liquid and vapor mixing rules for property calculations can be found in Zavaleta-Aguilar (2015). From the experimental Nusselt numbers obtained (Nuf) it was evaluated a correlation of Nusselt number (Nuc) as a function of Reynolds (Re) and Prandtl (Pr) numbers, as usual from theory. Chun and Seban (1971) proposed that the Nusselt number in falling film can be expressed in a general form according to Eq. (8).
Nu = aReb Pr n
(8)
In this work the Reynolds number ranged from 108 to 246 and the Prandtl number ranged from 1.68 to 2.65, both evaluated for the concentrated solution inlet state at the generator (Tcs, xcs, p). Because the data are for a heating process, it was adopted n = 0.4 as the Prandtl exponent, as usually recommended. By the least squares curve fitting method it was
(9)
Figure 14 also shows the overall uncertainty bars. The maximum deviation of this correlation (Nuc) with experimental data was 46.9%. High uncertainty values were obtained because of the propagation of uncertainty values associated with the heat transfer overall coefficient (Uf). Figure 15 shows a comparison of the Nusselt number correlated from this work (Eq. 9) compared with many other correlations from literature proposed by Wilke (1962), Chun and Seban (1972), Owens (1978), Mitrovic (1986), and Fujita and Tsutsui (1998). Whenever necessary, it was used the average Prandtl number which equals 2.0 (the average value obtained in this work). All the literature curves, except the correlation of Mitrovic (1986), showed a decreasing dependence of the Nusselt number with the Reynolds number until about Re = 300–400, and became nearly independent of Re afterwards (up to 1000).
international journal of refrigeration 59 (2015) 304–316
5.
Conclusions
In this work it was successfully built and tested a stainless steel test rig for carrying out an experimental study on falling films technology applied to an ammonia–water distiller (generator + rectifier). The generator consisted of an oil heated horizontal tube bundle having the liquid solution flowing over the tubes giving rise to a falling film pattern. Experimental tests were carried out with two concentrated solution concentrations xcs at nominal values of 0.37 and 0.49. A 100% surface wettability was achieved in this work for Re above 100 up to nearly 250. The flow pattern observed was sheet and column for Re in the range 108–246 and Ga1/4 in the range 834–1249. It was not observed the formation of droplets between the tube spacing probably due to the fact that the distance (2 mm) was smaller than the ammonia–water mixture average droplet size of 4.2 mm. There was no spanwise film contraction on the tubes and between the tubes probably because the small distance between the tubes (2 mm). There was no nucleation of the film in all the experiments and only film evaporation regime dominated. It was obtained a distilled ammonia concentration up to 0.9974 at the best operational condition. It was possible to obtain distilled ammonia concentrations above 0.9750 in all tested cases. A parametric sensitivity analysis showed that distilled ammonia vapor purity enhances with: (1) increasing of the concentrated solution mass flow rate; (2) increasing of the concentration of the concentrated solution; (3) lowering of the rectifier temperature; (4) lowering of the average heating oil temperature (down to a certain value); (5) increasing the subcooling of the concentrated solution. The distilled ammonia mass flow rate (m d ) enhances with: (1) increasing of the average heating oil temperature; (2) increasing of the concentrated solution temperature; (3) decreasing of the concentrated solution mass flow rate; (4) decreasing the average temperature of the rectifier. The thermal analysis of the generator allowed reducing the experimental Nusselt number. Reynolds number ranged from Re = 108 to 246 and the Prandtl number that ranged from Pr = 1.68 to 2.64. The obtained correlation was Nu = 0.75Re−0.27Pr0.4, which compared well with previous researchers, except Mitrovic (1986).
Acknowledgments The authors thank to São Paulo Research Foundation (FAPESP), grant no. 2010/10858-4 for the project financial support and the first author thanks to FAPESP, grant no. 2010/16304-0 for the personal support. REFERENCES
Ahachad, M., Charia, M., Bernatchou, B., 1992. Study of an improved NH3/H2O solar absorption refrigerating machine in Rabat (Morocco). Sol. Energ. Mat. Sol. C. 28, 71–79. Anand, G., Erickson, D., 1999. Compact sieve-tray distillation column for ammonia-water absorption heat pump: part I – design methodology. ASHRAE Trans. 105, 796–803.
315
ASHRAE, 1994. Handbook: Refrigeration: Systems and Applications. Atlanta. Bejan, A., 2004. Forced convection: internal flows. Heat Transfer Handbook. John Wiley & Sons, NJ, USA (Chapter 5). Cengel, Y.A., Ghajar, A.J., 2011. Heat and Mass Transfer: Fundamentals and Applications, fourth ed. McGraw-Hill, New York. Chun, K.R., Seban, R., 1971. Heat transfer to evaporating liquids films. J. Heat Trans. 93, 391–396. Chun, K.R., Seban, R., 1972. Performance prediction of fallingfilms evaporators. J. Heat Trans. 94, 432–436. Determan, M.D., Garimella, S., 2011. Ammonia-water desorption heat and mass transfer in microchannel devices. Int. J. Refrigeration 34, 1197–1208. Fernádez-Seara, J., Sieres, J., 2006. Ammonia-water absorption refrigeration systems with flooded evaporators. Appl. Therm. Eng. 26, 2236–2246. Fernádez-Seara, J., Sieres, J., Vázquez, M., 2003. Distillation column configurations in ammonia-water absorption refrigeration systems. Int. J. Refrigeration 26, 28–34. Fujita, Y., Tsutsui, M., 1998. Experimental investigation of falling film evaporation on horizontal tubes. Heat Transf. Jpn. Res. 27, 609–618. Herold, K.E., Radermacher, R., Klein, S.A., 1996. Absorption Chillers and Heat Pumps. CRC Press, New York. Hu, X., Jacobi, A.M., 1996. The intertube falling-film: part 1 – flow characteristics, mode transitions, and hysteresis. J. Heat Trans. 118, 616–625. Klein, S.A., 2014. Engineering Equation Solver, Academic Professional version V9.698-3D, F-Chart Software. Lazcano-Véliz, Y., Siqueiros, J., Juárez-Romero, D., Morales, L.I., Torres-Merino, J., 2014. Analysis of effective wetting area of a horizontal generator for absorption heat transformer. Appl. Therm. Eng. 62, 845–849. Mendes, G.A., 2008, Study of vapor purification systems for a small absorption machine driven by solar energy (Estudo de Sistemas de Refinação de Vapor numa Máquina de Absorção de Pequena Potência Alimentada por Energia Solar. In Portuguese). Master Thesis, Technical University of Lisbon, Lisbon. Mendes, L.F., Collares-Pereira, M., Ziegler, F., 2007. A rich solution spray as a refining method in a small capacity, single effect, solar assisted absorption machine with the pair NH3/H2O: experimental results. Energ. Convers. Manage 48, 2996–3000. Merrick, M.H., 1972. Horizontal firing of generator in absorption refrigeration. US Patent 3750421. Mitrovic, J., 1986. Influence of tube spacing and flow rate on heat transfer from a horizontal tube to a falling liquid film. In: 8th International Heat Transfer Conference. San Francisco, 1949– 1956. Narvaez-Romo, B., Zavaleta-Aguilar, E.W., Simões-Moreira, J.R., 2014. An intrusive method for film thickness measurement on smooth horizontal tubes for sub-cooled water. In: HEFAT 10th International Conference on Heat Transfer, Fluid Mechanics and Thermodynamics, Orlando, USA. Ortigosa, A.S.P., Preter, F.C., Labozetto, R.L., Zavaleta-Aguilar, E.W., Simões-Moreira, J.R., 2008. Modeling, and simulation of a commercial ammonia – water absorption refrigeration cycle for production of chilled water. In ENCIT – 12th Brazilian Congress of Thermal Sciences and Engineering, Belo Horizonte, Brazil. Owens, W.L., 1978. Correlation of thin film evaporation heat transfer coefficients for horizontal tubes. In: 5th OTEC Conference, Miami Beach, 3, VI-71-89. Ribatski, G., 2006. An experimental study on flow characteristics of a liquid film falling on a vertical column of horizontal tubes. In: ENCIT – 11th Brazilian Congress of Thermal Sciences and Engineering, Curitiba, Brazil.
316
international journal of refrigeration 59 (2015) 304–316
Sieres, J., Fernández-Seara, J., Uhía, F.J., 2009. Experimental characterization of the rectification process in ammoniawater absorption systems with a large-specific-area corrugated sheet structured packing. Int. J. Refrigeration 32, 1230–1240. Srikhirin, P., Aphornratana, S., Chungpaibulpatana, S., 2001. A review of absorption refrigeration technologies. Renew. Sust. Energ. Rev. 5, 343–372. Starace, G., De Pascalis, L., 2012. An advanced analytical model of the diffusion absorption refrigerator cycle. Int. J. Refrigeration 35, 605–612. Weber, C., Berger, M., Mehling, F., Heinrich, A., Núñez, T., 2014. Solar cooling with water-ammonia absorption chillers and concentrating solar collector – operational experience. Int. J. Refrigeration 39, 57–76.
Wilke, W., 1962. Heat transfer in falling films (Wärmeübergang an Rieselfilme. In German). VDI – Forschungsheft 490, 4– 36. Zavaleta-Aguilar, E.W., 2015. Experimental study of a falling film distiller for an ammonia-water absorption refrigeration cycle in a horizontal tube bundle (Estudo experimental de um destilador por filme descendente para um ciclo de refrigeração por absorção de amônia-água em um banco de tubos horizontais. In Portuguese). PhD Thesis, University of São Paulo, Brazil. Zavaleta-Aguilar, E.W., Simões-Moreira, J.R., 2012. Thermal design of a tray type distillation column of an ammonia/ water absorption refrigeration cycle. Appl. Therm. Eng. 41, 52–60.
Available online at www.sciencedirect.com
ScienceDirect j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / i j r e f r i g
Horizontal tube bundle falling film distiller for ammonia–water mixtures E.W. Zavaleta-Aguilar a,b, J.R. Simões-Moreira a,b,* a b
IEE – Institute of Energy and Environment, University of São Paulo, São Paulo, SP, Brazil SISEA- Alternative Energy Systems Lab., Escola Politécnica, University of São Paulo, SP, Brazil
A R T I C L E
I N F O
A B S T R A C T
Article history:
This work presents an experimental study of an important component of an ammonia–
Received 23 April 2015
water absorption refrigeration cycle based on the falling film technology – the distiller, which
Received in revised form 4 July 2015
is formed by two parts: the generator and the rectifier. It was carried out a parametrical
Accepted 15 July 2015
study on the operation of the distiller by testing several parameters. The experimental results
Available online 23 July 2015
showed that the distilled ammonia vapor concentration always degraded as rectifier, generator, and concentrated solution temperatures increased and improved as concentrated
Keywords:
solution mass flow rates and concentrated solution concentrations increased. The highest
Ammonia–water
distilled ammonia vapor concentration obtained was of 0.9974. Photographic documenta-
Falling film
tion of flow pattern over the generator tube bundle was also obtained resulting in both jet
Distillation
column and sheet patterns. Finally, a Nusselt number correlation for the ammonia–water
Absorption
falling film over the generator tube bundle was proposed based on the experimental results and compared with other authors’ correlations. © 2015 Elsevier Ltd and International Institute of Refrigeration. All rights reserved.
Distillateur à faisceau de tubes horizontaux à film tombant pour des mélanges ammoniac-eau Mots clés : Ammoniac-eau ; Film tombant ; Distillation ; Absorption
1.
Introduction
Absorption refrigeration cycles (ARC) are promising candidates for diminishing the dominating dependence on regular mechanical vapor compression refrigeration cycles (CRC) in the air conditioning and refrigeration industries. Their distinct advantage over CRC relies on the fact that they can be powered by either waste and recovered heat sources or be directly
powered by solar energy at none or low electrical energy consumption. They can also operate in a combined heat and power (CHP) or cogeneration configuration increasing the final usage of the fuel chemical energy content used to power a thermal machine, i.e., increasing the energy utilization factor (EUF). On the other hand, absorption refrigeration cycles are usually heavier, larger, and have a higher capital expenditure than standard compression refrigeration cycles at the same nominal capacity. They are larger and heavier due to the high number
* Corresponding author. Escola Politécnica at University of São Paulo, Mechl. Eng. Dept, SISEA – Alternative Energy Systems Lab., Av. Prof. Mello Moraes, 2231, 05508-900 São Paulo, SP, Brazil. Tel.: +55 11 30915684; Fax: +55 11 30915684. E-mail address: [email protected] (J.R. Simões-Moreira). http://dx.doi.org/10.1016/j.ijrefrig.2015.07.022 0140-7007/© 2015 Elsevier Ltd and International Institute of Refrigeration. All rights reserved.
international journal of refrigeration 59 (2015) 304–316
305
Nomenclature a A b COP cp d g Ga F k LMTD m n Nu Pr q Re SF T TR U x y
constant [–] area [m2] constant [–] coefficient of performance [–] specific heat at constant pressure [J kg−1 °C−1] diameter [m] acceleration of gravity [m s−2] modified Galileo number, Ga = ργ3/µ4g [–] correction factor [–] thermal conductivity [W m−1 °C−1] logarithmic mean temperature difference [°C] mass flow rate [kg s−1] constant [–] Nusselt number, Nu = α(v2/k3g)1/3 [–] Prandtl number, Pr = cpµ/k [–] heat transfer rate [W] Reynolds number, Re = 4Γ/µ [–] shape factor [–] temperature [°C] tons of refrigeration [3517 W] overall heat transfer coefficient [W m−2 °C−1] ammonia–water concentration – liquid [kgammonia kgsolution−1] ammonia–water concentration – vapor [kgammonia kgsolution−1]
Greek symbols α heat transfer coefficient [W m−2 °C−1] γ surface tension [N m−1]
of heat and mass exchangers employed to increase its efficiency. ARC are commercially available in two technologies. The first one is based on lithium bromide–water mixtures, and the second one on ammonia–water mixtures. The former can be applied only to air conditioning systems, as the fluid refrigerant is plain water. However, that cycle has a higher coefficient of performance (COP). Simple effect ARC have typical COP of 0.6–0.7 for lithium bromide cycles and 0.5 for ammonia– water cycles (ASHRAE, 1994), higher COP can be obtained in multi-effects cycles (Srikhirin et al., 2001). The second one can be used to either subzero or above zero degree Celsius applications, as the fluid refrigerant is ammonia. Furthermore, the ammonia–water cycle demands smaller heat exchangers, it operates at above atmospheric pressure and there is no any crystallization concern (Herold et al., 1996). In the evaporation process of the ammonia–water mixture, a small fraction of water is always present in the vapor phase dominated by ammonia. Thus, the wet vapor ammonia produced in the generator section must be purified to get rid of its water content, since liquid water can accumulate and increase the evaporator temperature, a phenomenon known as temperature glide. In accordance with such an operational problem, Fernádez-Seara and Sieres (2006) studied theoretically the consequences of exiting high amounts of water from the distiller in ammonia–water absorption refrigeration cycles and concluded that water content in wet ammonia is accountable for the low COP in experimental and industrial AARC.
Γ
Δ ε μ v ρ σ
mass flow rate per axial unit length of tube (each side) [kg m−1 s−1] difference [–] emissivity [–] dynamic viscosity [kg m−1 s−1] kinematic viscosity [m2 s−1] density [kg m−3] Stefan–Boltzmann constant [W m−2 K−4]
Subscripts c correlation conv convection cs concentrated solution d distilled ds diluted solution f falling film g generator in inside m average o oil out outside r rectifier rad radiation s saturation v vapor vess vessel w wall
The two key components of ammonia–water absorption refrigeration cycles (AARC) design and operation are the absorber and the distiller, wherein the former one has been the most studied (Determan and Garimella, 2011). The distiller must provide pure or at a very high purity degree of ammonia vapor to operate the cycle. Small commercial ammonia–water absorption refrigeration cycles (5 TR) have flooded generators (also known as desorbers), in which wet ammonia vapor is produced from a liquid solution heated by hot flue gases. The wet ammonia vapor is driven to a tray filling type analyzer, also known as rectification column or fractionation column (Herold et al., 1996; Merrick, 1972). After that, the ammonia vapor is directed to the rectifier to, finally, reach the condenser where a high degree of purity has been reached. According to Merrick (1972), the tray type distiller is generally the largest component in an AARC. This is due to the use of the analyzer and a gas burner placed below the generator, in his work the gas burner was placed at side of the vertical generator to decrease its size. Previous works on the distiller are Anand and Erickson (1999), who analyzed theoretically a distillation column for an 8TR AARC. Zavaleta-Aguilar and Simões-Moreira (2012) carried out a thermal modeling analysis and design of a holed tray distillation column for a 5 TR AARC. Those authors carried out a parametric analysis of the liquid and vapor ammonia–water mass flow rates, concentration, and temperature along with a study over geometrical parameters influence such as plate hole and column
306
international journal of refrigeration 59 (2015) 304–316
Fig. 1 – (a) Packed column (adapted from Sieres et al., 2009), (b) spray column (adapted from Mendes, 2008).
diameters. The authors do not acknowledge previous experimental studies of tray type distillers in open literature. Sieres et al. (2009) compared experimentally commercial random and structured filling types (Fig. 1a) such as: Pall® rings 10 mm (stainless steel), Berl® saddles 1/2” (ceramic), Raschig® rings 15 mm (glass), Novalox® saddles 1/2” (ceramic), Mellapakplus® 752.Y (stainless steel) and SulzerBX® and found that the first (Pall® rings) needed less column height to produce the same degree of ammonia purification than the other fillings. The generator used was a coil heated by thermal oil. On the other hand, according to Starace and De Pascalis (2012) in diffusion absorption refrigerator (DAR) cycles, the vapor leaving the rectifier is rarely pure ammonia, moreover, the COP in these systems increases as a consequence of the decreasing con-
centration (and the increasing temperature) of the ammonia vapor leaving the rectifier. Ahachad et al. (1992) compared analytically the performance of two simple effect ARC. In the first one, wet ammonia vapor came out from the generator and it was driven directly to the condenser without any rectification column. In the second study, it used a bubble column rectifier to purify the wet ammonia produced by the generator. The results showed that the COP was improved by 35% by the presence of the bubble column rectifier. However, even with that system the COP obtained was low, around 0.4, as the highest value. That happen most probably due to the low ammonia concentration (yd = 0.95) leaving the bubble column system. No further geometry detail of the generator was explained. Mendes (2008) studied experimentally two wet ammonia purification systems for an ammonia–water absorption refrigeration cycle. The first system (also studied by Mendes et al., 2007) was a spray column type (Fig. 1b), where the concentrated solution was sprayed in the wet ammonia vapor by a 45° angle spray nozzle.The descending liquid spray droplets came into contact with the counterflow wet ammonia vapor produced in the generator. The second purification system was an analyzer packed with Novalox® 1/2” saddle filling.The author obtained ammonia vapor concentration from 0.9754 to 0.9903 for the spray column and from 0.9825 to 0.9912 for the packed column. It used a plate heat exchanger as generator in their work. Determan and Garimella (2011) studied experimentally a falling film generator for an ammonia–water absorption refrigeration cycle with capacity up to 5 TR. The generator was formed by 304 stainless steel horizontal tube bundle of 1.067 mm i.d., 1.575 mm o.d., and 140 mm length. Their generator was divided into five passes, each pass had 16 rows, each row was oriented at 90° of the other and each row had 27 columns. The generator had an overall geometry of 178 mm × 178 mm × 508 mm. The authors obtained an average wet area of 45%. It used a water–glycol mixture as the heating fluid in the generator operating at an inlet temperature around 100 °C.The two tested concentrations of the concentrated solution were 0.44 and 0.48.The pressure in the generator was 6.7 bar. Finally, they obtained 0.9873 wet ammonia concentration at the generator outlet. A similar concentration was obtained by Zavaleta-Aguilar and Simões-Moreira (2012) at the enrichment section output. This would indicate that it is not necessary to use the analyzer (stripping and enrichment sections) above the generator, when the falling film technology is applied for generators. On the other hand, the COP is higher when partial condensation rectifiers were used than the ones based on complete condensation (Fernádez-Seara et al., 2003). Therefore, in this work we chose generator/partial condensation rectifier configuration. The present work aims at the novel study and analysis of the generator and the rectifier both of them combined in one single component (the distiller) using the falling film technology for an ammonia–water absorption refrigeration cycle. In order to achieve a distiller parametric analysis, the main controlling magnitudes were independently analyzed, namely: the tubing heating temperature, the rectifier temperature, the concentrated solution temperature, the solution mass flow rate, and the ammonia–water concentration of the liquid solution. By independently controlling those magnitudes it was possible to simulate actual operation conditions and
international journal of refrigeration 59 (2015) 304–316
evaluate their influence on the concentration and mass flow rate of the distilled ammonia vapor. The parametric study consisted in changing the control variables such as solution temperature, generator, rectifier, etc. to verify their influence on the vapor leaving the distiller. Detailed discussion on the distiller operation and its main components is carried out. Particularly, the liquid solution distributor is discussed, as its operation should properly homogenize the falling film over the first tube row and the ones underneath. Finally a heat transfer analysis to evaluate the Nusselt number for the liquid ammonia–water mixture is presented and compared with other authors’ correlations.
2.
Experimental setup
The experimental setup is composed of a main ammonia– water solution circuit, whose ammonia concentration can be properly altered according to a test run. Two other auxiliary circuits were also built. The first one is formed by a heating oil circuit to heat up the ammonia–water solution to the preset testing temperature. The other circuit is formed by a water cooling circuit to rectify the wet ammonia vapor. Each circuit is explained in detail in the following subsections.
2.1.
Ammonia–water testing rig
A schematic diagram of the experimental setup is depicted in Fig. 2. The ammonia–water distilled saturated vapor at high
T p M V
Ammonia-water circuit Heating circuit Water cooling circuit
Termocouple Pressure transmitter Coriolis flow meter Oil flow meter
Water tank
Oil tank
Water air cooler
Water heater
Distilled vapor p
T
Tp
Condenser
RECTIFIER T
Oil heater T
Sub-cooler
T
T T V Oil flow meter
degree of purity exits the distiller and flows to the condenser where it is condensed to in sequence which undergoes a subcooling process in the sub-cooler. The condensed ammonia coming from the sub-cooler is mixed up with the diluted solution exiting the distiller (after passing through the solution cooler). After the mixing process, the liquid (concentrated solution) flows to the tank solution, where the liquid level could be checked by observing a mounted borosilicate glass sight. Next, the concentrated solution is pumped to the distiller by a magnetic driven gear pump controlled by a frequency inverter. Between the distiller and the pump there is an electrical heater (heater solution) PID controlled at the set solution temperature for the test run. An electric resistance was considered for having a precise control of the concentrated solution temperature entering to the distiller. In real systems the concentrated solution is pre-heated in a heat exchanger with the diluted solution leaving the generator. The measurements of the ammonia distilled, concentrated solution, and diluted solution mass flow rates were attained by three Coriolis flow meters (Krohne model: Optimass MFM 33000 C) indicated by the letter “M” in Fig. 2. Besides giving the precise mass flow rates, the flow meters also indicated the liquid density. Thermocouples (T type) were installed at the inlet and outlet of each heat exchanger. Pressure readings were obtained by pressure transmitters (Salcas, 316 stainless steel diaphragm, model: SA-1000), one installed at the distiller top wall and the other on the condenser wall. Furthermore, three other thermocouples (T type) and three pressure transmitters (Salcas, 316 stainless steel diaphragm, model: SA1000) were installed in the measuring stations of the mass flow meters (see Fig. 2). Thus, a complete set of thermodynamic properties (density, pressure, and temperature) of the distilled ammonia, concentrated solution, and diluted solution was directly measured, which allowed one to obtain the corresponding ammonia concentration at the measuring station by using a solving thermodynamic properties routine, such as the one implemented in the Engineering Equation Solver software (Klein, 2014). At the distilled ammonia vapor line, just after the distiller top cover (see Figs. 3 and 4c) it was installed a safety pressure relief valve set to open at internal pressure of 17 bar or higher. The relief valve piping vents to a water tank installed outside the laboratory. All elements of the ammonia– water circuit work at high pressure. The whole circuit was tested at static pressures up to 21 bar for assuring operational safety and a leak free circuit. Two important elements of that circuit, the distiller and the liquid distribution system, are detailed in the following two subsections. The whole test rig was made of stainless steel.
p,T DISTILLER
GENERATOR
M Coriolis flow meter Oil pump
307
Diluted solution
Solution cooler
Concentrated Solution Solution T heater
M
p,T
M
p,T
Mixer p T Solution tank
Fig. 2 – Experimental setup schematics.
2.1.1.
The distiller
Figure 3 shows the schematics of the distiller set. In the lower part, it is illustrated the tube bundle generator and on the top it is seen the tube bundle rectifier. The whole set is placed inside a pressure vessel. We call this distiller as the falling film type, because the concentrated solution of ammonia–water flows around the generator horizontal tube bundle giving rise to a thin liquid film configuration on each tube. Also, the rectifier is built following the same falling film technology, as seen. Figure 4 shows still pictures of parts of the distiller prior to installation. The generator (Fig. 4a) is heated by a heating oil that flows inside the horizontal tube bundle. As the
308
international journal of refrigeration 59 (2015) 304–316
Distilled ammonia vapor
Coolant fluid Concentrated solution Heating fluid
TOP COVER VESSEL
RECTIFIER
Ammonia-water falling film Termocouple Distributor
Dry wall
GENERATOR
concentrated solution flows down around the external heated tube wall, wet ammonia–water vapor is produced. There are 70 tubes of 8 mm o.d. and 1 mm wall thickness. The ammonia– water liquid solution that remains after evaporation is accumulated at the bottom part and it is called diluted solution, because it has less amount of ammonia than that of the initial concentrated solution. The diluted solution exits at the pressure vessel from the bottom. The wet ammonia vapor produced in the generator moves upwards to reach the rectifier section (Fig. 4b). In the rectifier section, the excess of water vapor is removed from the vapor by a cooling process, thereby obtaining distilled ammonia vapor. The rectifier was built having two lateral vertical plates in order to force to wet ammonia vapor produced in the generator section to be put in direct contact with the water cooled rectifier tube bundle. Consequently, the wet ammonia vapor condenses partially also forming a condensing falling film over the rectifier tubing.
2.1.2. Diluted solution
Fig. 3 – Distiller schematics.
The liquid distributor
The liquid distributor is the component which ensures an even distribution of the liquid around the generator tubes both spanwise and circumferentially each tube. A thorough comparison of types of distributors used in a horizontal tube falling film technology is better described in Narvaez-Romo et al. (2014) and Zavaleta-Aguilar (2015). Those authors carried out several tests with plain water to verify the distribution and wettability
Fig. 4 – Still pictures of distiller parts.
international journal of refrigeration 59 (2015) 304–316
309
electrical resistance set at the testing temperature. Next, the heated oil was pumped to the generator and distributed as uniform as possible to the horizontal tube bundle. In order to carry out an energy balance, both the inlet and outlet temperatures were also measured and the oil flow readings were obtained from a volumetric flow turbine meter (“V” in Fig. 2). To avoid any piping stress because of the decrease of the oil density, an atmospheric expansion tank was also installed in the oil circuit (see Fig. 2). A still picture of the experimental setup built showing the circuits and the cycle components described above is shown in Fig. 6.
3.
Fig. 5 – Distributor schematics. of water on the generator tubes and to find out the better geometrical distributor system. In this work, it was implemented the distributor shown in Fig. 5. The distribution system consists of a 3/4” tube branched into 7 tubes of 8 mm o.d. and 6 mm i.d. which have holes at 0° and 60° from the horizontal plane (Fig. 5a), to allow the exit of ammonia–water solution. A 12 mm o.d. tube with 1 mm wall thickness positioned outside the 8 mm o.d. tube directs the solution to its bottom part to a thin blade V-shaped that was welded at the bottom of the 8 mm tube. To ensure flow homogeneity it was placed a tray with nozzles, as shown in Fig. 5b. Water droplet diameter falling on a sanded stainless steel plate from a height of 6 mm corresponds to 23 mm average diameter spreading region as seen in experiments. Therefore, in order to ensure a good tube wall wettability it was adopted a 17 mm nozzle pitch. The distribution system is shown in Fig. 5c.
2.2.
Water cooling circuit
Water was used as the cooling liquid for the rectifier, for the condenser, for the sub-cooler, and for the solution cooler. After cooling those heat exchangers, the water was pumped to a forced air cooler placed outside the laboratory (Fig. 2). Each water sub-circuit had an independent centrifugal pump and the water flow was controlled by needle valves. The cooling water circuit had an electrical heater mounted just preceding the rectifier so that the ammonia vapor purification could be tested at different cooling temperatures.
2.3.
Heating circuit
Synthetic thermal oil was used as the heating fluid in the generator. Oil heating system was attained by a PID controlled
Experimental procedure
Table 1 shows the tested experimental parameters. Two liquid solution concentrations (xcs) were tested whose average values were 0.37 and 0.49. For each concentration it was evaluated the influence of the solution mass flow rate, mcs, spanning from 0.0160 to 0.0276 kg s−1 as well as the influence of two average rectifier temperature, Tr,m, at 62 and 34 °C, taken as the average between the inlet and outlet cooling water temperatures. For the average concentration of 0.37 it was analyzed the influence of the oil temperature, To,m, from 110 to 136 °C taken as the average between the inlet and outlet heating oil temperatures. For the solution concentration of 0.49 the oil temperature range was between 97 and 125 °C. For the concentration of 0.37 it was analyzed the influence of the concentrated solution temperature, Tcs, from 87 to 103 °C and for the concentration of 0.49 the concentrated solution temperature ranged from 69 to 82 °C. The oil and concentrated solution temperature ranges tested for the two concentrations (0.37 and 0.49) were different because their saturation temperatures were also distinct. The total heat flux supplied to the generator (qg) was calculated from the First Law of Thermodynamics applied to the oil flow as given by Eq. (1). The thermal energy flux is transferred to the falling film (qf) and to the generator dry wall (qw) (Eq. 2). The generator dry wall (see Fig. 3) transfers energy by radiation and by convection as indicated in Eq. (3).
qg = mo c p, o ( To,in − To, out )
(1)
qg = q f + qw
(2)
qw = qw, rad + qw, conv
(3)
The energy transferred by radiation (qw,rad) is the energy transferred from the generator dry wall to the inner wall of the vessel bottom section and is calculated by Eq. (4). On the other hand, it was assumed natural convection heat transfer from the dry wall generator to the vapor (qw,conv) (Eq. 5), since we observed that the vapor was mainly still around the generator dry wall. The convective heat transfer coefficient ( α v ) in Eq. (5) was calculated from Cengel and Ghajar (2011) correlation. The generator dry wall (q w ) transferred on average 8% of the total of energy (qg), where 0.3% was by radiation and 7.7% was by convection.
310
international journal of refrigeration 59 (2015) 304–316
Fig. 6 – Still picture of experimental setup built.
qw, rad =
4 σ (Tw4 − Tvess ) 1−ε 1 1−ε + + ε Aw AwSFw − vess ε Avess
qw, conv = α v Aw (Tw − Tv )
(4)
(5)
By leakage concern it was not possible to measure the oil temperature exactly at the tube bundle inlet and outlet. Thus, the logarithmic mean temperature difference (LMTD) evaluation was approached with the generator oil temperature inlet and outlet. This approach seems reasonable since the generator oil temperature inlet and outlet averaged 3.1 °C. The LMTD is calculated by Eq. (6). Since the ammonia–water and oil flows were in cross flow configuration it used the F-factor for correcting the calculation performed as if it were a counterflow configuration. The overall heat transfer coefficient (Uf) concerning to the outer area of the generator tubes (Af) was obtained by Eq. (7). The average heat transfer coefficient ( α f ) and the Nusselt number ( Nu f ) of the falling film were evaluated by Eqs. (8) and (9), respectively. To evaluate the heat transfer coefficient inside the tube ( α in ) it was considered the thermal entrance length in laminar flow (Bejan, 2004). In order to determine the ammonia–water thermophysical properties the relations in M. Conde Engineering (2006) were used.
LMTD =
Uf =
(To,in − Tds ) − (To,out − Tcs ) (T − Tds ) ln o,in (To,out − Tcs )
(6)
qf FA f LMTD
{
1 1 1 ⎛ dout ⎞ dout ⎛ dout ⎞ ln ⎜ = − ⎟ ⎜ ⎟+ αf Uf α in ⎝ din ⎠ 2kw ⎝ din ⎠ ⎛ v2f ⎞ Nu f = α f ⎜ 3 ⎟ ⎝ k f g⎠
(7)
}
(8)
13
4.
Results and discussions
4.1.
Hydraulic analysis
(9)
Figure 7 shows a still picture of the ammonia–water liquid solution distribution. The nozzles of the tray distributor eliminated the liquid spanwise contraction between the distributor and the top generator tubes. The liquid spanwise contraction causes dry areas that can decrease the heat exchange efficiency of a
international journal of refrigeration 59 (2015) 304–316
311
Table 1 – Testing parameter range. Parameter
Unit
Concentrated — solution concentration, xcs Concentrated solution mass flow rate, mcs
kg s−1
°C Average rectifier temperature, Tr,m
°C Average oil temperature, To,m °C Concentrated solution temperature, Tcs
Value 0.37
Observation
0.49
Correspond to evaporation temperature of −13 and 5 °C respectively (Ortigosa et al., 2008) 0.0160–0.0276 Suitable for a 1200 W cooling capacity AARC and a concentrated and dilute concentration difference of about 3% (Ortigosa et al., 2008) 34 and 62 34 °C corresponds approximately to the temperature of a cooling tower in summer season and 64 °C was used just to verify the a hot condition (use of an internal fluid of the cycle) 110–136 97–125 Use of low temperature sources (solar, cogeneration) 87–103 69–82 In AARC the diluted solution can be pre-heated before entering to the generator by using a heat exchanger between the absorber and the generator (Herold et al., 1996). In this work the diluted solution was warmed close to its saturation temperature (see Fig. 12).
heat exchanger (Ribatski, 2006). Those nozzles also allowed to stabilize and to evenly distribute the liquid film. There was no visible spanwise contraction on the tubes and between the tubes because of the small distance between tubes (2 mm). The generator tubes were blasted with glass balls and it was achieved an average roughness of 196 µm. It was noted an increase of the wettability after the blasting. In previous tests without the vessel (see Figs. 3 and 4d) using plain water rather than an ammonia–water mixture,
Fig. 7 – Ammonia–water distribution at: mcs = 0.0276 kg s−1, xcs = 0.3658, Re = 222, 15.4 bar, Tcs = 100.1 °C.
Fig. 8 – Flow pattern between the generator tubes, for Re from 108 to 246.
Fig. 9 – Flow pattern between the tubes compared with that of Hu and Jacobi (1996).
each time that the external tube array surface was wet, it was observed that the inside tube arrays also were wet. Consequently, in actual tests using ammonia–water mixtures it was presumed that the same phenomenon would happen as direct observation would not be possible. However, the conclusion came after observing the two opposite tube arrays of the tube bank (see Fig. 4d). The wettability obtained in this study was 100% and overcame that of Lazcano-Véliz et al. (2014) who obtained 80% of its wet area for similar conditions of mass flow rate (Γ = 0.0080 kg m−1s−1). The distance between tubes used for those authors was 5 mm. It was noticed in the present work that the first dry areas appeared at Re = 140 for xcs = 0.37 and at Re = 100 for xcs = 0.49. The sheet and column flow patterns were observed in all tests, as shown in Fig. 8, where the Reynolds number varied in the range Re = 108–246 (and Galileo number Ga1/4 from 834 to 1249). This result is in opposition to Hu and Jacobi’s (1996), since, according to them, the flow pattern must be of droplets type, as shown in Fig. 9. The sheet and column flow patterns obtained in this work were mainly due to the spacing of 2 mm imposed between the tubes, smaller than the droplet size estimated in 4.2 mm (Zavaleta-Aguilar, 2015).
312
international journal of refrigeration 59 (2015) 304–316
Fig. 10 – A general distillation process.
4.2.
Parameter influence
The temperature-concentration (T–x) diagram in Fig. 10 shows the thermodynamic paths of a distillation process for the ammonia–water solution. There can be seen that the concentrated solution enters the distiller at the generator at liquid sub-cooled condition with temperature Tcs = 100.1 °C and concentration xcs = 0.3777, labeled “1” in the diagram. This solution is heated in the generator tube bundle until its bubble point is reached, when the evaporation process starts off. As the liquid solution receives heat from the heated tubes as it flows around them, the evaporation process goes on, which causes a decrease in liquid concentration and an increase in its temperature. Near the distiller outlet, the concentrated solution has become a diluted solution (at the saturated condition), at temperature of T ds = 108.6 °C and at concentration of xds = 0.3500, labeled as number “2” in the T–x diagram. The wet ammonia vapor produced in the generator (label “3” in the T–x diagram) moves upward into the rectifier section and it is put in contact with the cooling tube bundle, where a part of its vapor is condensed and sent back to the bottom of the generator. It can be noticed in T–x diagram in Fig. 10 that the vapor concentration that leaves the rectifier (label “4”) is higher than the one at the inlet; this means that the vapor condensed in the rectifier contained a portion of water. The distilled vapor obtained is condensed in the condenser/sub-cooler, leaving as sub-cooled liquid (label “5”). Parametric analysis results are shown in Figs. 11, 12, and 13. Figure 11a shows the influence of the distilled ammonia concentration as a function of the solution mass flow rate (mcs) for two average rectifier temperatures (Tr,m = 34 and 62 °C) at two concentrated liquid solution concentrations (xcs = 0.37 and 0.49). In this test the average oil temperature (To,m) was 132 °C and the concentrated solution temperature (Tcs) was 100 °C for the concentrated solution concentration of 0.37; and To,m = 109 and Tcs = 76 °C for the concentrated solution concen-
tration of 0.49. Figure 11a indicates that when the concentrated solution mass flow rate increased, the concentration of ammonia distillate (yd) increased as well. It happened because the larger the concentrated solution mass flow rate, the smaller the thermal heat transfer of the falling film, because the film thickness is thicker. Thus, the amount of the wet ammonia vapor produced in the generator is smaller, nevertheless with higher concentration. The graphics also shows that for xcs = 0.37 at Tr,m = 62 °C the ammonia distilled concentration ranged from 0.9750 to 0.9842. On the other hand, when the rectifier average temperature was 34 °C the concentration yd ranged from 0.9879 to 0.9901. This means that the lower the rectifier temperature the higher the ammonia distilled concentration. Other important conclusion is that the ammonia distilled concentration (yd) is higher for the higher concentrated solution concentration, i.e. xcs = 0.49 reaching up to 0.9974. The distilled mass flow rate (md) decreases as the concentrated mass flow rate increases, m cs , as indicated in Fig. 11b, as a general rule. The distiller vessel pressure varied from 14.1 to 15.4 bar for Tr,m = 62 °C, and from 12.4 to 13.9 bar for Tr,m = 34 °C. The next series of graphs in Fig. 12 shows the influence of the average oil temperature (To,m). The concentrated solution mass flow rate was constant (mcs = 0.0230 kg s−1) as well as the rectifier temperature (Tr,m = 34 °C). The concentrated solution inlet temperature was 100 °C for xcs = 0.37, and 76 °C for xcs = 0.49. In Fig. 12a, one can see that the distilled ammonia vapor concentration (y d ) decreases as the heating oil temperature increases for both concentration values of the concentrated solutions. On the other hand, the distilled ammonia vapor mass flow rate displays the reverse trend regarding the oil heating as seen in Fig. 12b. Finally, it is worthy to mention that the minimum average heating oil temperature to produce a reasonable amount of ammonia vapor was 110 °C for the 0.37 concentrated solution concentration, and 98 °C for xcs = 0.49. Weber et al. (2014) used water at temperatures between 160
international journal of refrigeration 59 (2015) 304–316
Fig. 11 – Influence of the concentrated solution mass flow rate (mcs), the average rectifier temperature (Tr,m), the concentration of the concentrated solution (xcs) on the distilled concentration (yd) and distilled mass flow rate (md). Full line (___) is for xcs = 0.37 and dashed line (--) is for xcs = 0.49. and 200 °C to warm up the generator of a commercial ammonia–water absorption cycle. Tests were carried out to analyze the influence of the concentrated solution inlet temperature considering the degree of sub-cooling, ΔTs , which can be defined as the difference between the solution bubble temperature and the actual concentrated solution inlet temperature. Those results are shown in Fig. 13a and 13b. For these experiments, the average rectifier temperature (Tr,m) was kept at 34 °C and the average heating oil temperature (To,m) was also constant for a given concentrated solution concentration, namely To,m = 109 °C for xcs = 0.49 and To,m = 132 °C for xcs = 0.37. In any situation the concentrated solution mass flow rate was 0.0232 kg s−1. For all cases, as the degree of sub-cooling decreases the distilled ammonia vapor concentration decreases (Fig. 13a), and the distilled ammonia vapor mass flow rate increases (Fig. 13b). The oil mass flow oil remained practically constant in all the tests. In general, when the total heat flux supplied to the
313
Fig. 12 – Influence of the average oil temperature (To,m) and the concentration of the concentrated solution (xcs) on the distilled concentration (yd) and distilled mass flow rate (md). Full line (___) is for xcs = 0.37 and dashed line (--) is for xcs = 0.49.
generator increases the diluted solution concentration decreases as well and its temperature increases and, furthermore, the distilled ammonia vapor mass flow rate increases and its concentration decreases.The system pressure remained constant in each test. This pressure depends basically on the condensator pressure that was controlled by the condenser cooling temperature. Small variations in the system pressure are also achieved by varying the rectifier and generator temperatures. The vapor pressure drop into the distiller was negligible since the vapor pressure drop from the distiller to the sub-cooler (longer and narrower path into the distiller, see Fig. 2) was only 0.8%.
4.3.
Nusselt correlation
It was possible to experimentally obtain average Nusselt numbers (Nuf) for this study data basis as indicated in Fig. 14. Experimental Nusselt numbers were calculated according to
314
international journal of refrigeration 59 (2015) 304–316
Fig. 14 – Experimental results of the ammonia–water falling film average Nusselt number divided by Pr0.4.
Fig. 13 – Influence of the concentrated solution temperature (Tcs) and the concentration of the concentrated solution (xcs) on the distilled concentration (yd) and distilled mass flow rate (md). Full line (___) is for xcs = 0.37 and dashed line (--) is for xcs = 0.49.
Fig. 15 – Comparison of the Nusselt number correlation obtained in this work with many other literature correlations.
possible to correlate the Nuf/Pr0.4 values against aReb, resulting in a = 0.75 and b = −0.27, as indicated by a full line in Fig. 14. The Nusselt correlation final equation is given by Eq. (9).
Nuc = 0.75Re−0.27 Pr 0.4 Eq. (9). Ammonia–water liquid and vapor mixing rules for property calculations can be found in Zavaleta-Aguilar (2015). From the experimental Nusselt numbers obtained (Nuf) it was evaluated a correlation of Nusselt number (Nuc) as a function of Reynolds (Re) and Prandtl (Pr) numbers, as usual from theory. Chun and Seban (1971) proposed that the Nusselt number in falling film can be expressed in a general form according to Eq. (8).
Nu = aReb Pr n
(8)
In this work the Reynolds number ranged from 108 to 246 and the Prandtl number ranged from 1.68 to 2.65, both evaluated for the concentrated solution inlet state at the generator (Tcs, xcs, p). Because the data are for a heating process, it was adopted n = 0.4 as the Prandtl exponent, as usually recommended. By the least squares curve fitting method it was
(9)
Figure 14 also shows the overall uncertainty bars. The maximum deviation of this correlation (Nuc) with experimental data was 46.9%. High uncertainty values were obtained because of the propagation of uncertainty values associated with the heat transfer overall coefficient (Uf). Figure 15 shows a comparison of the Nusselt number correlated from this work (Eq. 9) compared with many other correlations from literature proposed by Wilke (1962), Chun and Seban (1972), Owens (1978), Mitrovic (1986), and Fujita and Tsutsui (1998). Whenever necessary, it was used the average Prandtl number which equals 2.0 (the average value obtained in this work). All the literature curves, except the correlation of Mitrovic (1986), showed a decreasing dependence of the Nusselt number with the Reynolds number until about Re = 300–400, and became nearly independent of Re afterwards (up to 1000).
international journal of refrigeration 59 (2015) 304–316
5.
Conclusions
In this work it was successfully built and tested a stainless steel test rig for carrying out an experimental study on falling films technology applied to an ammonia–water distiller (generator + rectifier). The generator consisted of an oil heated horizontal tube bundle having the liquid solution flowing over the tubes giving rise to a falling film pattern. Experimental tests were carried out with two concentrated solution concentrations xcs at nominal values of 0.37 and 0.49. A 100% surface wettability was achieved in this work for Re above 100 up to nearly 250. The flow pattern observed was sheet and column for Re in the range 108–246 and Ga1/4 in the range 834–1249. It was not observed the formation of droplets between the tube spacing probably due to the fact that the distance (2 mm) was smaller than the ammonia–water mixture average droplet size of 4.2 mm. There was no spanwise film contraction on the tubes and between the tubes probably because the small distance between the tubes (2 mm). There was no nucleation of the film in all the experiments and only film evaporation regime dominated. It was obtained a distilled ammonia concentration up to 0.9974 at the best operational condition. It was possible to obtain distilled ammonia concentrations above 0.9750 in all tested cases. A parametric sensitivity analysis showed that distilled ammonia vapor purity enhances with: (1) increasing of the concentrated solution mass flow rate; (2) increasing of the concentration of the concentrated solution; (3) lowering of the rectifier temperature; (4) lowering of the average heating oil temperature (down to a certain value); (5) increasing the subcooling of the concentrated solution. The distilled ammonia mass flow rate (m d ) enhances with: (1) increasing of the average heating oil temperature; (2) increasing of the concentrated solution temperature; (3) decreasing of the concentrated solution mass flow rate; (4) decreasing the average temperature of the rectifier. The thermal analysis of the generator allowed reducing the experimental Nusselt number. Reynolds number ranged from Re = 108 to 246 and the Prandtl number that ranged from Pr = 1.68 to 2.64. The obtained correlation was Nu = 0.75Re−0.27Pr0.4, which compared well with previous researchers, except Mitrovic (1986).
Acknowledgments The authors thank to São Paulo Research Foundation (FAPESP), grant no. 2010/10858-4 for the project financial support and the first author thanks to FAPESP, grant no. 2010/16304-0 for the personal support. REFERENCES
Ahachad, M., Charia, M., Bernatchou, B., 1992. Study of an improved NH3/H2O solar absorption refrigerating machine in Rabat (Morocco). Sol. Energ. Mat. Sol. C. 28, 71–79. Anand, G., Erickson, D., 1999. Compact sieve-tray distillation column for ammonia-water absorption heat pump: part I – design methodology. ASHRAE Trans. 105, 796–803.
315
ASHRAE, 1994. Handbook: Refrigeration: Systems and Applications. Atlanta. Bejan, A., 2004. Forced convection: internal flows. Heat Transfer Handbook. John Wiley & Sons, NJ, USA (Chapter 5). Cengel, Y.A., Ghajar, A.J., 2011. Heat and Mass Transfer: Fundamentals and Applications, fourth ed. McGraw-Hill, New York. Chun, K.R., Seban, R., 1971. Heat transfer to evaporating liquids films. J. Heat Trans. 93, 391–396. Chun, K.R., Seban, R., 1972. Performance prediction of fallingfilms evaporators. J. Heat Trans. 94, 432–436. Determan, M.D., Garimella, S., 2011. Ammonia-water desorption heat and mass transfer in microchannel devices. Int. J. Refrigeration 34, 1197–1208. Fernádez-Seara, J., Sieres, J., 2006. Ammonia-water absorption refrigeration systems with flooded evaporators. Appl. Therm. Eng. 26, 2236–2246. Fernádez-Seara, J., Sieres, J., Vázquez, M., 2003. Distillation column configurations in ammonia-water absorption refrigeration systems. Int. J. Refrigeration 26, 28–34. Fujita, Y., Tsutsui, M., 1998. Experimental investigation of falling film evaporation on horizontal tubes. Heat Transf. Jpn. Res. 27, 609–618. Herold, K.E., Radermacher, R., Klein, S.A., 1996. Absorption Chillers and Heat Pumps. CRC Press, New York. Hu, X., Jacobi, A.M., 1996. The intertube falling-film: part 1 – flow characteristics, mode transitions, and hysteresis. J. Heat Trans. 118, 616–625. Klein, S.A., 2014. Engineering Equation Solver, Academic Professional version V9.698-3D, F-Chart Software. Lazcano-Véliz, Y., Siqueiros, J., Juárez-Romero, D., Morales, L.I., Torres-Merino, J., 2014. Analysis of effective wetting area of a horizontal generator for absorption heat transformer. Appl. Therm. Eng. 62, 845–849. Mendes, G.A., 2008, Study of vapor purification systems for a small absorption machine driven by solar energy (Estudo de Sistemas de Refinação de Vapor numa Máquina de Absorção de Pequena Potência Alimentada por Energia Solar. In Portuguese). Master Thesis, Technical University of Lisbon, Lisbon. Mendes, L.F., Collares-Pereira, M., Ziegler, F., 2007. A rich solution spray as a refining method in a small capacity, single effect, solar assisted absorption machine with the pair NH3/H2O: experimental results. Energ. Convers. Manage 48, 2996–3000. Merrick, M.H., 1972. Horizontal firing of generator in absorption refrigeration. US Patent 3750421. Mitrovic, J., 1986. Influence of tube spacing and flow rate on heat transfer from a horizontal tube to a falling liquid film. In: 8th International Heat Transfer Conference. San Francisco, 1949– 1956. Narvaez-Romo, B., Zavaleta-Aguilar, E.W., Simões-Moreira, J.R., 2014. An intrusive method for film thickness measurement on smooth horizontal tubes for sub-cooled water. In: HEFAT 10th International Conference on Heat Transfer, Fluid Mechanics and Thermodynamics, Orlando, USA. Ortigosa, A.S.P., Preter, F.C., Labozetto, R.L., Zavaleta-Aguilar, E.W., Simões-Moreira, J.R., 2008. Modeling, and simulation of a commercial ammonia – water absorption refrigeration cycle for production of chilled water. In ENCIT – 12th Brazilian Congress of Thermal Sciences and Engineering, Belo Horizonte, Brazil. Owens, W.L., 1978. Correlation of thin film evaporation heat transfer coefficients for horizontal tubes. In: 5th OTEC Conference, Miami Beach, 3, VI-71-89. Ribatski, G., 2006. An experimental study on flow characteristics of a liquid film falling on a vertical column of horizontal tubes. In: ENCIT – 11th Brazilian Congress of Thermal Sciences and Engineering, Curitiba, Brazil.
316
international journal of refrigeration 59 (2015) 304–316
Sieres, J., Fernández-Seara, J., Uhía, F.J., 2009. Experimental characterization of the rectification process in ammoniawater absorption systems with a large-specific-area corrugated sheet structured packing. Int. J. Refrigeration 32, 1230–1240. Srikhirin, P., Aphornratana, S., Chungpaibulpatana, S., 2001. A review of absorption refrigeration technologies. Renew. Sust. Energ. Rev. 5, 343–372. Starace, G., De Pascalis, L., 2012. An advanced analytical model of the diffusion absorption refrigerator cycle. Int. J. Refrigeration 35, 605–612. Weber, C., Berger, M., Mehling, F., Heinrich, A., Núñez, T., 2014. Solar cooling with water-ammonia absorption chillers and concentrating solar collector – operational experience. Int. J. Refrigeration 39, 57–76.
Wilke, W., 1962. Heat transfer in falling films (Wärmeübergang an Rieselfilme. In German). VDI – Forschungsheft 490, 4– 36. Zavaleta-Aguilar, E.W., 2015. Experimental study of a falling film distiller for an ammonia-water absorption refrigeration cycle in a horizontal tube bundle (Estudo experimental de um destilador por filme descendente para um ciclo de refrigeração por absorção de amônia-água em um banco de tubos horizontais. In Portuguese). PhD Thesis, University of São Paulo, Brazil. Zavaleta-Aguilar, E.W., Simões-Moreira, J.R., 2012. Thermal design of a tray type distillation column of an ammonia/ water absorption refrigeration cycle. Appl. Therm. Eng. 41, 52–60.