- Email: [email protected]
Experimental study of a graphite disks absorber couple to a heat transformer
Experimental Thermal and Fluid Science 46 (2013) 29–36
Contents lists available at SciVerse ScienceDirect
Experimental Thermal and Fluid Science journal homepage: www.elsevier.com/locate/etfs
Experimental study of a graphite disks absorber couple to a heat transformer J. Olarte-Cortés a,⇑, J. Torres-Merino b, J. Siqueiros c a
Posgrado en Ingeniería y Ciencias Aplicadas, Universidad Autónoma del Estado de Morelos, Av. Universidad 1001, Col. Chamilpa, C.P. 62209, Cuernavaca, Morelos, Mexico Facultad de Química, Universidad Nacional Autónoma de México, Circuito interior, Ciudad universitaria, Col. Copilco, Delegación Coyoacán, C.P. 04510, Mexico c Centro de Investigación en Ingeniería y Ciencias Aplicadas, Universidad Autónoma del Estado de Morelos, Av. Universidad 1001, Col. Chamilpa, C.P. 62209, Cuernavaca, Morelos, Mexico b
a r t i c l e
i n f o
Article history: Received 9 March 2012 Received in revised form 25 October 2012 Accepted 10 November 2012 Available online 10 December 2012 Keywords: Absorber Lithium bromide Tar impregnated graphite disks Heat transformer
a b s t r a c t An experimental study was performed on a graphite disk absorber coupled to an absorption heat transformer. The working solution used was H2O/LiBr. Due to corrosion caused to materials such as carbon steel and stainless steel with the solution H2O/LiBr, it was necessary to consider the use of another material such as tar impregnated graphite. This material has interesting properties of corrosion resistance, withstands high temperatures and presents a high thermal conductivity than the graphite without impregnation. However, there does not exist a well established methodology for designing heat exchangers with new geometries for its implementation in heat transformers. For this reason, the heat exchanger was designed with a stainless steel shell and graphite disks arranged internally in a column to carry out the absorption process. In the experimental tests carried out for a thermal load in the range from 625 to 1460 W, heat transfer coefficients in the absorber were obtained in the range from 723 to 1535 W/m2 K. As the number of Reynolds increases from 110 to 144, the heat transfer coefficient increases up to a maximum value of 954 W/m2 K, at Reynolds number at about 144, but when Reynolds number was increased above 147, the heat transfer coefficient decreased. Ó 2012 Elsevier Inc. All rights reserved.
1. Introduction Many industries depend on conventional fuels for the transformation of raw materials into products used by people for their most basic needs. The expansion of industries around the world and the emergence of new ones are the main reason for the consumption of conventional resources such as natural gas, petroleum and its derivates; therefore, the demand increases. In some processes it is possible to reduce the consumption of these conventional sources with the use of thermal devices which can yield high quality energy even when they use a low quality energy supply, it is the case of heat transformers. The main component of an absorption heat transformer is the absorber, in which the absorption process and heat generation are carried out. In the absorber the heat released is proportional to the absorption process. During the last decades, the study of these absorbers has been well documented. Several experimental studies of the thermal behavior have been analyzed and compared with theoretical data in order to model other configurations based on parameters established during the operation of the system under steady-state conditions [1,2]. Several absorber geometries have been studied, such as vertical plate type exchangers [3–5], shell ⇑ Corresponding author. Tel./fax: +52(777)3297984. E-mail address: cuanticofi[email protected] (J. Olarte-Cortés). 0894-1777/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.expthermflusci.2012.11.013
and spiral tube exchangers [6,7], vertical concentric tubes [8–10], shell and horizontal tubes [11], columns of brass disks [12] and horizontal tubes [13–15]. The materials used for construction of flat plate exchangers, vertical concentric tubes and tubes which may be arranged either vertically or horizontally inside a shell, were made of cooper, carbon steel, brass and stainless steel [16]. These materials are selected for their high thermal conductivity and some of them are resistant to the corrosion of certain chemical substances. However, in processes where the solutions are highly corrosive, the thermal conductivity of the material is sacrificed and it is changed for other material with a major resistance to corrosion [17]. Some researchers have experimented with materials resistant to highly corrosive working mixtures, such as cupronickel (copper–nickel alloy), graphite impregnated with phenolic cross-linked resin, high-temperature-treated carbon, and Teflon (PTFE) [18]. Several companies manufacture different types of heat exchangers with impregnated graphite materials using them as their internal structure. The impregnation technique provides the heat exchangers with a high mechanical resistance, resistance to corrosion and to high temperatures, all of which are useful in applications such as evaporators, condensers and reactors. In absorption heat pumps the most commonly used solutions are H2O/LiBr and NH3/H2O, due to their refrigerant–absorbent properties. The former solution is non-toxic, non-volatile, non-pollutant
30
J. Olarte-Cortés et al. / Experimental Thermal and Fluid Science 46 (2013) 29–36
Nomenclature A Cp d h k m Q Re T U X
heat transfer area, m2 specific heat capacity, kJ/kg diameter, m heat transfer coefficient, W/m2 K thermal conductivity, W/m K mass flow rate, kg/s heat rate, W Reynolds number (4 C/l) temperature °C or K overall heat transfer coefficient, W/m2 K concentration (wt% LiBr)
Greek symbols C m/pdi mass flow rate per unit of wetted perimeter, kg/m s
and has a high latent heat. However, its great disadvantage is its corrosiveness. The solution NH3/H2O has high latent heat, is lightly toxic, and its operation pressures are relatively high; however, its main disadvantage is that it requires the use of a rectifier in order to remove the water vapor from the mixture that exits from the generator before it enters to the condenser. The solution H2O/LiBr is used in heat transformers because the temperature levels are higher in comparison with the solution NH3/H2O. There does not seem to be information in the literature concerning the use of heat exchangers built with disks of tar-impregnated graphite. In this paper we show the experimental results of a vertical stainless steel shell absorber and tar-impregnated graphite disks, which were accommodated internally in a column. This heat exchanger was mounted on an absorber test bench and during the experimental tests the parameters such as solution volumetric flow, the fluid temperatures that simulate waste heat, and concentrations of the working solution were modified. The main focus of this study was heat transfer analysis using a non-metallic material with an experimentally unanalyzed geometry in absorbers. 2. Experimental and measurement instruments 2.1. Absorber The studies on absorbers have been largely carried out in vertical plates, vertical tubes (Fig. 1a) and horizontal tube bundles (Fig. 1b), in which the solution descends in falling film over the plate, the tubes or the tube bundles, where the solution distribution is not uniform over the whole surface. In this study the behavior is analyzed using disks with two different slope angles [18], which allow a better distribution of the solution throughout the whole area (Fig. 1c). Fig. 2 shows the vertical absorber built with 18 tar-impregnated graphite disks. The absorber is formed externally by a 316L stainless steel tube joined in its upper and lower ends to two bridles coupled to two headers with screws that give the disks column rigidity and hermetic seal (with 0.17 m in diameter and 0.56 m in length). The upper header is internally designed for the solution to be distributed and to wet the upper and lower surface of the disks. Fig. 3 shows the design of a tar-impregnated graphite disk with 0.1 m in outside diameter. The slope of its surfaces with 7° in the upper and 14° in the lower and its rough surface to improve the distribution of the working mixture. This kind of material has a 50–80 W/m-K thermal conductivity [18]. Neoprenes packing were used as a hermetic seal between the disks and the shell. The con-
Subscripts AB absorber CO condenser EV evaporator GE generator i internal in inlet LM logarithmic mean o external out outlet s solution w water Superscript e equilibrium
centrated working mixture enters through the upper part of the absorber and by gravity goes towards the disks where it gets distributed, while the water vapor enters through the lower part of the absorber and gets mixed with the solution descending from the disks. The working mixture then exits through a lateral connection as a diluted working mixture. The cooling water flows in the annular space formed by the tube and the outside disks. 2.2. Mechanism of heat and mass transfer The overall efficiency of absorption machines depends particularly of the absorber on operating conditions and physical characteristics (heat exchange surface and internal geometry). The studies carried out in the absorbers have been theoretical and experimental. In some theoretical models, heat and mass transfer mechanism is analyzed based on formulations of continuity, diffusion and energy. In other studies, the models analyze the liquid film flowing over the horizontal tubes and inside them, there is a fluid flowing removing heat. For this particular case the boundary layer equations were resolved, which are equations of momentum, energy and mass of the liquid film assuming laminar flow [19]. Design tools for absorbers with H2O/LiBr [20] and NH3–H2O mixtures in film falling on smooth tubes have been developed. Theoretical models assume ideal conditions so that the results show significant differences compared with the experimental results compared in 30% [21]. In experimental works, the analysis of the distribution of the solution in different geometries as banks of tubes, vertical tubes and others geometries has been performed. Effects of absorbed water vapor in lithium bromide falling film in vertical tubes of different geometries, temperature gradients and heat and mass transfer coefficients have been tested experimentally. The mechanism of heat and mass transfer in different kinds of absorbers, occurs by the absorption of water vapor into direct contact with the film surface of the H2O/LiBr solution. The absorbers may be concentric tubes, horizontal tube bundle and vertical flat plates. In the concentric tubes the solution flowing through the inner wall of the inner tube or in the annular section, through the outer wall of the inner tube and the fluid that removes heat flows countercurrent through the annular section or inside the pipe, respectively. For the inner tube, the solution is distributed in falling film on the inner wall of the tube while the vapor fills the internal space of the tube. The fluid flow removes heat by the annular section (Fig. 4a). In the horizontal tube bank, the solution falls along the outer surface of the tube forming a falling film of solution and vapor fills the entire internal space of the heat exchanger shell.
31
J. Olarte-Cortés et al. / Experimental Thermal and Fluid Science 46 (2013) 29–36
(a)
(b)
(c)
Fig. 1. Different geometries for absorbers: (a) vertical tubes, (b) horizontal tube banks and (c) graphite disk.
falling film of solution is spread on a surface, whether the surface is vertical, horizontal or inclined. The water vapor fills the entire space and it is absorbed in direct contact by the solution surfaces generating exothermic reaction, and this heat released goes through the wall of the tube, plate or disk for its transfer towards the fluid that removes heat. 2.3. Instrumentation
Fig. 2. Structure of the graphite disks absorber and its connections.
And inside the bundle tube, a flowing fluid removes heat (Fig. 4b). In the vertical plates, the solution is distributed in falling film on one of the surfaces of the plate and the vapor fills the space, while on the other side of the plate a fluid flows in countercurrent removing heat (Fig. 4c). In the disks used in the present work, the solution is distributed over the inner surfaces in falling film and in the annular section the fluid flows removing the heat generated (Fig. 2). Whatever the geometry, the mechanism is the same, a
Table 1 shows the characteristics of the generator, evaporator, condenser and the solution heat exchanger. The components are connected to each other so that the test bench works as a heat transformer. The connection diagram is shown in Fig. 5. The heating supply consists of a water container installed with two tubular electrical resistors of 2 and 3 kW. These electric resistances provide the required heat to the evaporator and generator and a centrifuge pump of 1/2 hp moves the heating water flow in these components. Other pump of the cooling tower removes the heat in the condenser and the heat generated in the absorber. The volumetric flows of the cooling and heating water at the inlet of each component are measured by rotameters—made of 316L stainless steel and placed at the inlet of the components—with a minimum measurement rate of 745 mL/min and a maximum of 14 LPM. The control of this flow is done with valves. The measurement of the H2O/LiBr solution was similar to that of the cooling and heating water, but in order to regulate the variation of the flow a gear pump of positive displacement with control was used. The working solution used in the experimental test was H2O/ LiBr and its circulation into the generator-absorber circuit was carried out by means of two gear pumps of positive displacement with a maximum volumetric flow control of 5.85 LPM. A third gear pump of 4.55 LPM was used to circulate the condensated water from the condenser to the evaporator. Twenty-four type T (copper-constantan) thermocouples of gauge with a fluorinated ethylene propylene resin coating were installed. The thermocouples were introduced in thermowells connected at the inlet and outlet of each component to measure the temperature during the experimental tests. All of the thermocouples were connected to two 34901A channel multiplexer modules. In order to connect these modules to the computer, they were inserted in a commercial data acquisition switch unit with an RS-232 interphase. To measure the pressure in the generator and the absorber, two manometers were used with a measuring range from 0 to 307.8 kPa, and two previously calibrated pressure transducers with a measuring range from 0 to 204.4 kPa, were connected to the multiplexer module channel for registering the data in the computer.
32
J. Olarte-Cortés et al. / Experimental Thermal and Fluid Science 46 (2013) 29–36
Fig. 3. Section of the tar-impregnated graphite disk.
LiBr solution
LiBr solution
Coolant
Falling film
Falling film Falling film
coolant
LiBr solution
coolant
coolant
Water vapor Water vapor Water vapor
(a)
(b)
(c)
Fig. 4. Direct contact of the water vapor and H2O/LiBr in different geometries.
Table 1 Specifications of the components of the experimental bench. Component
Characteristics
Generator
Heat exchanger of shell and horizontal tubes [22], made of 316L stainless steel, with glass windows (measuring 0.17 m in width and 0.4 m in length) on both sides and with a designed heat capacity of 3 kW It is composed internally of a headstock of four tubes (whose external diameter is 0.0063 m and whose length is 0.50 m) with holes in their upper part evenly distributed in order to wet the columns of tubes with the working mixture. It has four columns of seven stainless steel tubes with a surface of heat transfer of 0.768 m2. Concentric tubes helical heat exchanger [23], constructed of 316L stainless steel. The surface of the heat transfer of each exchanger is 0.090, 0.104 and 0.108 m2 respectively.
Evaporator, condenser and solution heat exchanger
2.4. Operation of the heat transformer The experimental bench for absorbers worked as a single stage heat transformer, as shown in Fig. 6. It was conceived for continuous operation: the heating source supplies low quality heat to the generator and the evaporator. This heat supplied at the generator evaporates part of the working fluid of the diluted solution, increasing its concentration. The produced vapor passes from the generator to the condenser, where it is condensed by exchanging heat with an external fluid. The condensate working fluid is pumped to the evaporator (which is at a higher pressure) where it is converted to vapor using low quality heat from the same heat source as the generator. The vapor formed passes to the absorber. On the other hand, the concentrated working mixture in the generator is pumped to the solution heat exchanger before entering the absorber. In the absorber, the concentrated solution absorbs the working fluid in vapor phase, diluting the solution. The diluted solution is pumped to the generator, passing before by the solution heat exchanger, completing the cycle. The solution in the absorber is heated by the exothermic process of working fluid going into
solution. The cooling tower is the external heat sink whose purpose is to remove the high temperature heat from the absorber and the low temperature heat from the condenser. In the outlet of the absorber a pump is used to control the volumetric flow of the working solution. Quantities of heat QGE and QEV are supplied to the generator and evaporator respectively at the evaporator temperature TEV. A quantity of heat QAB is given out from the absorber at the higher absorber temperature TAB together with a quantity of heat QCO from the condenser at the lower condenser temperature TCO [24].
3. Analysis of experimental data All the analyzed experimental data correspond to steady state working conditions. At the beginning of each experimental test, the volumetric flow, the temperatures and the concentrations of the working mixture were established, although they were modified in a certain number of experiments. For the thermal analysis of the absorber, the number of Reynolds and the overall and indi-
J. Olarte-Cortés et al. / Experimental Thermal and Fluid Science 46 (2013) 29–36
33
Fig. 5. Schematic diagram of the experimental bench.
DT LM ¼ ðT s;in T w;in Þ aðT s;in T s;out Þ bðT w;in T w;out Þ
ð3Þ
Another way of obtaining DTLM is by using the registered temperatures of the absorber [10,11]:
DT LM ¼
ðT s;in T w;out Þ ðT s;out T w;in Þ T T w;out In T s;in s;out T w;in
ð4Þ
Or by using the equilibrium temperatures of the solution [8,13,15] in the following equation
DT LM ¼
ðT es;in T w;out Þ ðT es;out T w;in Þ T e T w;out In T s;in e T s;out
Fig. 6. Schematic diagram of the heat transformer.
vidual heat transfer coefficients were calculated from the data registered in each experimental test. The thermophysical properties of the working solution were taken from correlations proposed in the literature with a wide validity range [18]. The thermal load in the absorber is defined as the heat released in the absorber which is removed by the cooling water. This thermal load was thus calculated as:
Q AB ¼ mw Cpw ðT w;out T w;in Þ
ð1Þ
The overall heat transfer coefficient for the absorber was calculated, using the thermal load in the absorber, the heat transfer area, and Logarithmic Mean Temperature Difference as:
U¼
Q AB ADT LM
ð2Þ
where DTLM can be calculated from the following correlation [14] in which the constants ‘‘a’’ and ‘‘b’’ are, respectively, 0.65 and 0.5:
ð5Þ
w;in
The external heat transfer coefficient ho was obtained by taking into account that the configuration of the external surface of the disks with the shell looks like two concentric tubes. The external heat transfer coefficient ‘ho’ was obtained using the experimental data in the Reynolds and Nusselt dimensionless numbers. Finally, the internal heat transfer coefficient hi was calculated using the following equation:
U ¼ do
1 di hi
1 þ
do k
In ddo i
1 ho
ð6Þ
4. Results and discussion Table 2 shows the thermal load range calculated for each component using the Eq. (1). The values calculated from the thermal load take into account the measured temperatures and mass flow rates of the heating and cooling water for each component. In addition, the working fluid and working solution mass flow rates and temperatures are also shown in the table. The temperatures and mass flow rates ranges correspond to all calculated values of the thermal loads presented in this work. Then the thermal load reported is of 625 W corresponding to 19.6 and 25.3 °C (input and output temperatures respectively) and mass flow rate of 0.028 kg/
34
J. Olarte-Cortés et al. / Experimental Thermal and Fluid Science 46 (2013) 29–36
Table 2 Working ranges of the experimental study. Component
Heat capacity W
Mass flow (kg/s) H2O
H2O-LiBr
H2O
H2O-LiBr Input
Output
Absorber
625
1460
0.008
0.028
0.014
0.018
16.8 26.2
24.1 54.1
67.3 75.0
68.4 86.6
Generator
1600
2700
0.096
0.14
0.018
0.023
75.7 84.6
71.5 79.8
56.2 71.4
63.0 70.1
Condenser
1400
1900
0.048
0.07
15.6 25.6
21.0 31.7
Evaporator
800
1800
0.080
0.085
75.6 84.4
72.1 81.0
Thermal load in the absorber (W)
Min
Temperature (°C)
Max
Min
Max
Min input
Max input
Input
Output
Ts,in=71ºC, Tw,in=17-25ºC, mw,in=0.0289 kg/s
Absorber pressure (kPa)
Thermal load of the absorber (W)
Fig. 7. The thermal load obtained with respect to the pressure in the absorber.
Ts,in=67-71ºC, Tw,in=18-19ºC, mw,in=0.0289 kg/s
Mass flowrate of the solution (kg/s) Fig. 8. The experimental thermal load as a function of the input solution’s mass flow.
s. These data is within the range. These conditions were obtained when the steady state was reached. When there is a direct contact between water vapor and the H2O/LiBr solution surface inside the absorber, the system reaches equilibrium at the solution pressure (water vapor enters with a higher pressure). The temperature difference between the water vapor and the H2O/LiBr solution causes water vapor absorption in the solution. This absorption releases heat (thermal load) and the solution comes out diluted. The heat released will decrease until temperatures are equal. When the equilibrium pressure reached is higher, there will be a smaller temperature difference that will generate a lower absorption of water vapor in the solution and the heat generation will decrease [24]. Fig. 7 shows that the ther-
mal load decreases when the pressure is increased at a concentration of 51 wt%. The same effect is observed at 50 wt%, but it is more pronounced. For steady state operation of each experiment the maximum variations of pressure were ±1.5%. In Fig. 8, the behavior of the thermal load in the absorber is plotted versus the solution mass flow. The increase in the thermal load occurs when the solution mass flow rate is increased from 0.012 to 0.017 kg/s with an inlet concentration measured experimentally of 52 wt%. When the solution mass flow rate lifts and it is above 0.017 kg/s, the thermal load decreases. In a similar way it occurs the trend of the concentration of 51 wt%, but the thermal load was lower than the concentration of 52 wt%. The thermal load values for the concentration of 51 wt%, in the range of the mass flow rate of the solution of 0.019–0.021 kg/s, were not possible to obtain experimentally with the input parameters. The initial increase of the thermal load corresponds to a uniform distribution of the solutions on the disks (uniform thickness on both surfaces). This solution is mixed with the entrance vapor of the working fluid. Further increase of the mass flow rate causes a uniform increase film thickness on surface disk models. However, when the solution mass flow rate is increased above 0.017 kg/s, the thermal load drops because the solution does not have a uniform distribution over the lower surfaces of the disks and then the solution falls evenly over the bottom absorber, decreasing the heat generation. Before placing the disks inside, the shell (absorber), an experimental apparatus was constructed to observe the distribution of the solution on both surfaces of the disks. In this experimental test the H2O/LiBr solution was used and the mass flow rate was changed in the tests. As for a small flow, the distribution of the solution is good on both surfaces of the disks, but a higher mass flow rate causes the solution to slide through the center of the disks to the bottom of the container. The overall heat transfer coefficient U was calculated with the absorber thermal load, the heat transfer area and the Logarithmic Mean Temperature Difference DTLM. Three methods were used to calculate the DTLM: Deng and Ma [14] used a linear correlation to calculate DTLM (Eq. (3)), and it was used in this work. In the second method, inlet and outlet temperatures of the absorber were used to calculate the DTLM (Eq. (4)), [10,11] and in the third method, the DTLM is calculated with pressure and concentration data to determine the equilibrium temperatures of the mixture in the absorption process (Eq. (5)), [8,13,15]. Fig. 9 shows the plot of the overall heat transfer coefficient with respect to the thermal load using the three methods above mentioned. The overall heat transfer coefficient obtained from second method was from 167.71 to 317.41 W/m2 K; the range for the linear correlation was from 127.76 to 205.82 W/m2 K and the range obtained from the third method was from 76.94 to 260.38 W/m2 K. The highest overall heat transfer coefficients are obtained with the second method, while
Yoon et al. [11]
Heat transfer coefficient h i (W/m2 K)
Overall heat transfer coefficient U (W/m2 K)
J. Olarte-Cortés et al. / Experimental Thermal and Fluid Science 46 (2013) 29–36
35
Xin=52 % wt, Tw,in=18-24ºC, mw,in=0.0289 kg/s
Deng et al. [14] Miller et al. [8]
Reynolds number (dimensionless) Thermal load of the absorber (W)
Fig. 11. Variation of the internal heat transfer coefficient with respect to different films’ Reynolds numbers.
Ts,in=67-70ºC, Tw,in=17-19ºC, mw,in=0.0289 kg/s
Heat transfer coefficient hi (W/m 2 K)
Thermal load of the absorber (W)
Fig. 9. Overall heat transfer coefficient as a function of the thermal load.
Xs,in=51 % wt, Tw,in=19ºC, mw,in=0.0289 kg/s
Thermal load of the absorber (W) Difference Xin-Xout in the absorber (% wt) Fig. 10. Thermal load of the absorber with respect to the difference between the concentrations.
the lowest coefficients are obtained using the third method. Three methods followed a similar trend but with differences within the thermal load. The generation of heat inside the absorber takes place with the absorption of the steam proceeding from the evaporator and the H2O/LiBr solution that falls as a film on the disks. With the increased of the absorption of the steam by the working mixture the generation of heat increases. Fig. 10 shows that as the difference between the concentrations in the absorber increases so does the thermal load, but when the difference is in the range from 2 to 2.7 wt%, the thermal load decreases. In the experimental test, where the concentrations measured experimentally were of 50, 51 and 52 wt%, the tendency is very similar. The internal heat transfer coefficients were calculated using the Eqs. (1, 2, 3 and 6). When a film mass flow increases, also the thermal load in the absorber is increased. The increased thermal load is due to a good absorption with adequate mass flow rate under controlled conditions of pressure and temperatures. In order to increase the thermal load there must be an increase in the transference of heat, and this latter depends directly on the heat transfer coefficients. Fig. 11 shows the increase of internal heat transfer coefficient from 750 to 954 W/m2 K when Reynolds number is increased from 109 to 144, but when increasing Reynolds number over 144, the internal heat transfer coefficient decreases, for temperature of solution of 68 °C. Additionally, this behavior is also observed for the input solution temperatures of 72 °C. Fig. 12 shows that as the internal heat transfer coefficient increases, the thermal load of the absorber increases with a linear tendency. For heat transfer coefficient of 750 W/m2 K, the thermal
Fig. 12. Behavior of the internal heat transfer coefficient with respect to the thermal load of the absorber.
load obtained was of 936 W, whereas for heat transfer coefficient of 1397 W/m2 K, the corresponding thermal load was of 1285 W corresponding with a temperature of 67 °C. When the temperatures of the input solution are compared among them, they do not present any significant variation with respect to the thermal load, but for a temperature of 68 °C, the maximum heat coefficient obtained was of 1535 W/m2 K. Several studies have been carried out with different types of heat exchangers, as shown in Table 3. These heat exchangers differed in their geometry, size and construction material. The solution used was H2O/LiBr with additives [9] or any other solution [25] to reduce corrosion. The parameters analyzed by the authors, and listed in Table 3, were the thermal load, the film’s Reynolds number and the heat transfer coefficient. At low Reynolds numbers, with a large contact surface, this corresponds to high heat transfer coefficients and high thermal loads [14]. However, low heat transfer coefficients (within the range of 250–600 W/m2 K) were also obtained, with a small contact surface [9,25]. In the present work, in which the contact surface used was of 0.180 m2 in comparison with other works shown in table 3, the thermal load is less than those reported by Deng and Ma [14]. In the literature, it has not been found experimental or theoretical studies of heat exchangers with similar geometries. Only one author [18] presented a study using three different impregnated graphite tubes with external and internal fins, hence this study presents comparative data of turbulence in the tubes. Therefore, the heat exchangers used in this work in order to have a comparison were selected by using the same solution (H2O/LiBr solution), similar ranges of thermal loads and thus the heat exchangers found with these features are shown in the table 3. Table 3 shows that heat transfer coefficients are similar in order of magnitude.
36
J. Olarte-Cortés et al. / Experimental Thermal and Fluid Science 46 (2013) 29–36
Table 3 Comparison of parameters obtained by different authors. Author
Contact surface (m2)
Description
Type of absorber
Absorber’s heat capacity (kW)
Film’s adimensional Reynolds number
Heat transfer coefficient hi (kW/m2 K)
Bourouis et al. [25]
0.104
Vertical concentric tubes (SS)
0.25–0.70
50–300
0.25–0.60
Deng and Ma [14] Kim et al. [10] Medrano et al. [9]
0.603 0.149 0.104
Horizontal tubes (Cu) Vertical concentric tubes (SS) Vertical concentric tubes (SS)
3.6–5.25 0.4–1–0
12–39 50–150 75–325
1.7–3.15 0.9–1.42 0.32–0.58
Miller and Perez-Blanco [8] Yoon et al. [26] Present work
0.125
LiBr + LiI + LiNO3 + LiCl (falling film) H2O/LiBr (falling film) H2O/LiBr (falling film) H2O/LiBr Li2MoO4 (falling film) H2O/LiBr (falling film)
Vertical tube (glass)
0.64–0.91
70–350
1.25–1.40
0.180
H2O/LiBr (falling film) H2O/LiBr (falling film)
Helical Tar-impregnated graphite disks
0.62–1.46
7–30 110–200
0.44–0.60 0.62–1.53
5. Conclusions Experimental tests were carried out in a tar-impregnated graphite disks absorber installed in a testing bench which worked as a heat transformer. The absorber has a compact design with an area of transfer of 0.180 m2, with only a height of 0.56 m in comparison to other designs. The thermal load obtained in the absorber during the experimental tests ranges from 625 to 1460 W and the internal heat transfer coefficient in the absorber ranges from 723 to 1535 W/ m2 K. The film’s Reynolds number ranges from 109 to 200. The Reynolds number is increased in the solution in the range from 110 to 144 causing an increase in the internal heat transfer coefficient, where the maximum value is of 954 W/m2 K, when Reynolds number is at about 144, but when Reynolds number increase is over 147, the internal heat transfer coefficient decreases. When the internal heat transfer coefficient is increased from 723 to 1535 W/m2 K, the thermal load is also increased from 977 to 1308 W. The analysis carried out in this work obtained parameters from the absorber and their comparison with previous reports show a good approach. The thermal load presents similar values to the obtained by Medrano et al. [9] and Miller and Perez-Blanco [8], but the internal heat transfer coefficient presents values over those obtained by Bourouis et al. [25], Medrano et al. [9] and Yoon et al. [26]. Acknowledgements The authors wish to thank the Consejo Nacional de Ciencia y Tecnología (CONACyT) for the fellowship awarded to J. Olarte Cortés, the Centro de Investigación en Ingeniería y Ciencias Aplicadas (CIICAp) for its technical support and Dr. David Juárez and Dr. Alfredo Hernández for their valuable comments. The graphite disks were designed and constructed by Dr. J. Torres Merino in collaboration with Le Carbone, Lorraine, France. References [1] X. Ma, J. Chen, S. Li, Q. Sha, A. Liang, W. Li, J. Zhang, G. Zheng, Z. Feng, Application of absorption heat transformer to recover waste heat from a synthetic rubber plant, Applied Thermal Engineering 23 (2003) 797–806. [2] A. Huicochea, J. Siqueiros, R.J. Romero, Portable water purification system integrated to a heat transformer, Desalination 165 (2004) 385–391. [3] J.I. Yoon, T.T. Phan, Ch.G. Moon, P. Bansal, Numerical study on heat and mass transfer characteristic of plate absorber, Applied Thermal Engineering 25 (2005) 2219–2235. [4] M. de Vega, J.A. Almendros-Ibañez, G. Ruiz, Performance of a LiBr–water absorption chiller operating with plate heat exchanger, Energy Conversion and Management 47 (2006) 3393–3407.
[5] N. Nishimura, T. Nomura, H. Ivota, N. Uedono, Development of vertical plate type absorber, in: Proc. of the Int. Sorption Heat Pump Conf, Shanghai, China, 2002, pp. 310–313. [6] W. Rivera, Experimental evaluation of a single-stage heat transformer used to increase solar pond’s temperature, Solar Energy 69 (2000) 369–376. [7] D.T. Rose, C.A. Zuritz, H. Perez-Blanco, Thermodynamic analysis and pilot plant design for a solar assisted double-effect refrigeration absorption cycle, International Absorption Heat Pump Conference 31 (1993) 109–116. [8] W.A. Miller, H. Perez-Blanco, Vertical-tube aqueous LiBr falling film absorption using advanced surfaces, International Absorption Heat Pump Conference 31 (1993) 185–202. [9] M. Medrano, M. Bourouis, A. Coronas, Absorption of water vapour in the falling film of wáter-lithium bromide inside a vertical tube at air-cooling thermal conditions, International Journal of Thermal Sciences 41 (2002) 891–898. [10] J. Kim, K. Cho, Analytical and experimental study on the absorption performance in the vertical absorber, in: Proc. of the Int. Sorption Heat Pump Conf., Shanghai, China, 2002, pp. 315–319. [11] J.I. Yoon, E. Kim, K.H. Choi, W.S. Seol, Heat transfer enhancement with a surfactant on horizontal bundle tubes of an absorber, International Journal of Heat and Mass Transfer 45 (2002) 735–741. [12] B.P. Schwarzer, M.S. Rahbar, P. Le Goff, A spiral fin tube: a novel type of falling film heat and mass exchanger, International Absorption Heat Pump Conference 31 (1993) 179–184. [13] I. Greiter, A. Wagner, V. Weiss, G. Alefeld, Experimental investigation of heat and mass transfer in a Horizontal-tube falling film absorber with aqueous solutions, International Absorption Heat Pump Conference 31 (1993) 225–232. [14] S.M. Deng, W.B. Ma, Experimental studies on the characteristics of an absorber using LiBr/H2O solution as working fluid, International Journal of Refrigeration 22 (1999) 293–301. [15] I. Kyung, K.E. Herold, Y.T. Kang, Experimental verification of H2O/LiBr absorber bundle performance with smooth horizontal tubes, International Journal of Refrigeration 30 (2007) 582–590. [16] F.P. Incropera, D.P. De Witt, Fundamentals of heat and mass transfer, third ed., 1990. [17] S. Jiangzhou, R.Z. Wang, Experimental research on characteristics of corrosionresisting nickel alloy tube used in triple-effect LiBr/H2O absorption chiller, Applied Thermal Engineering 21 (2001) 1161–1173. [18] J.M. Torres, Contacteurs gaz–liquide pour pompes á chaleur a absorption multi-étagées, These de doctorado, L’ institute National Polytechnique de Lorraine, Francia, 1997. [19] E.D. Rogdakis, V.D. Papaefthimiou, D.C. Karampinos, A realistic approach to model LiBr–H2O smooth falling film absorption on a vertical tube, Applied Thermal Engineering 23 (2003) 2269–2283. [20] G.A. Florides, S.A. Kalogirou, S.A. Tassou, L.C. Wrobel, Design and construction of a LiBr–water absorption machine, Energy Conversion and Management 44 (2003) 2483–2508. [21] I. Yuekse, W. Schluender, A model of an ammonia–water falling absorber, ASHRAE Transactions 94 (1988) 467–483. [22] M.R. Salimpour, Heat transfer coefficients of shell and coiled tube heat exchangers, Experimental Thermal and Fluid Science 33 (2009) 203–207. [23] T.J. Rennie, V.G.S. Raghavan, Experimental studies of a double-pipe helical heat exchanger, Experimental Thermal and Fluid Science 29 (2005) 919–924. [24] F.A. Holland, J. Siqueiros, S. Santoyo, C.L. Heard, E. Santoyo, Water Purification Using Heat Pumps, E&FN Spon, UK, 1999. [25] M. Bourouis, M. Vallés, M. Medrano, A. Coronas, Absorption in the film of water–(LiBr + LiI + LiNO3 + LiCl) in a vertical tube at air-cooling thermal conditions, International Journal of Thermal Sciences 44 (2005) 491–498. [26] J.I. Yoon, O.K. Kwon, P.K. Bansal, Ch.G. Moon, H.S. Lee, Heat and mass transfer characteristics of a small helical absorber, Applied Thermal Engineering 26 (2006) 186–192.
Contents lists available at SciVerse ScienceDirect
Experimental Thermal and Fluid Science journal homepage: www.elsevier.com/locate/etfs
Experimental study of a graphite disks absorber couple to a heat transformer J. Olarte-Cortés a,⇑, J. Torres-Merino b, J. Siqueiros c a
Posgrado en Ingeniería y Ciencias Aplicadas, Universidad Autónoma del Estado de Morelos, Av. Universidad 1001, Col. Chamilpa, C.P. 62209, Cuernavaca, Morelos, Mexico Facultad de Química, Universidad Nacional Autónoma de México, Circuito interior, Ciudad universitaria, Col. Copilco, Delegación Coyoacán, C.P. 04510, Mexico c Centro de Investigación en Ingeniería y Ciencias Aplicadas, Universidad Autónoma del Estado de Morelos, Av. Universidad 1001, Col. Chamilpa, C.P. 62209, Cuernavaca, Morelos, Mexico b
a r t i c l e
i n f o
Article history: Received 9 March 2012 Received in revised form 25 October 2012 Accepted 10 November 2012 Available online 10 December 2012 Keywords: Absorber Lithium bromide Tar impregnated graphite disks Heat transformer
a b s t r a c t An experimental study was performed on a graphite disk absorber coupled to an absorption heat transformer. The working solution used was H2O/LiBr. Due to corrosion caused to materials such as carbon steel and stainless steel with the solution H2O/LiBr, it was necessary to consider the use of another material such as tar impregnated graphite. This material has interesting properties of corrosion resistance, withstands high temperatures and presents a high thermal conductivity than the graphite without impregnation. However, there does not exist a well established methodology for designing heat exchangers with new geometries for its implementation in heat transformers. For this reason, the heat exchanger was designed with a stainless steel shell and graphite disks arranged internally in a column to carry out the absorption process. In the experimental tests carried out for a thermal load in the range from 625 to 1460 W, heat transfer coefficients in the absorber were obtained in the range from 723 to 1535 W/m2 K. As the number of Reynolds increases from 110 to 144, the heat transfer coefficient increases up to a maximum value of 954 W/m2 K, at Reynolds number at about 144, but when Reynolds number was increased above 147, the heat transfer coefficient decreased. Ó 2012 Elsevier Inc. All rights reserved.
1. Introduction Many industries depend on conventional fuels for the transformation of raw materials into products used by people for their most basic needs. The expansion of industries around the world and the emergence of new ones are the main reason for the consumption of conventional resources such as natural gas, petroleum and its derivates; therefore, the demand increases. In some processes it is possible to reduce the consumption of these conventional sources with the use of thermal devices which can yield high quality energy even when they use a low quality energy supply, it is the case of heat transformers. The main component of an absorption heat transformer is the absorber, in which the absorption process and heat generation are carried out. In the absorber the heat released is proportional to the absorption process. During the last decades, the study of these absorbers has been well documented. Several experimental studies of the thermal behavior have been analyzed and compared with theoretical data in order to model other configurations based on parameters established during the operation of the system under steady-state conditions [1,2]. Several absorber geometries have been studied, such as vertical plate type exchangers [3–5], shell ⇑ Corresponding author. Tel./fax: +52(777)3297984. E-mail address: cuanticofi[email protected] (J. Olarte-Cortés). 0894-1777/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.expthermflusci.2012.11.013
and spiral tube exchangers [6,7], vertical concentric tubes [8–10], shell and horizontal tubes [11], columns of brass disks [12] and horizontal tubes [13–15]. The materials used for construction of flat plate exchangers, vertical concentric tubes and tubes which may be arranged either vertically or horizontally inside a shell, were made of cooper, carbon steel, brass and stainless steel [16]. These materials are selected for their high thermal conductivity and some of them are resistant to the corrosion of certain chemical substances. However, in processes where the solutions are highly corrosive, the thermal conductivity of the material is sacrificed and it is changed for other material with a major resistance to corrosion [17]. Some researchers have experimented with materials resistant to highly corrosive working mixtures, such as cupronickel (copper–nickel alloy), graphite impregnated with phenolic cross-linked resin, high-temperature-treated carbon, and Teflon (PTFE) [18]. Several companies manufacture different types of heat exchangers with impregnated graphite materials using them as their internal structure. The impregnation technique provides the heat exchangers with a high mechanical resistance, resistance to corrosion and to high temperatures, all of which are useful in applications such as evaporators, condensers and reactors. In absorption heat pumps the most commonly used solutions are H2O/LiBr and NH3/H2O, due to their refrigerant–absorbent properties. The former solution is non-toxic, non-volatile, non-pollutant
30
J. Olarte-Cortés et al. / Experimental Thermal and Fluid Science 46 (2013) 29–36
Nomenclature A Cp d h k m Q Re T U X
heat transfer area, m2 specific heat capacity, kJ/kg diameter, m heat transfer coefficient, W/m2 K thermal conductivity, W/m K mass flow rate, kg/s heat rate, W Reynolds number (4 C/l) temperature °C or K overall heat transfer coefficient, W/m2 K concentration (wt% LiBr)
Greek symbols C m/pdi mass flow rate per unit of wetted perimeter, kg/m s
and has a high latent heat. However, its great disadvantage is its corrosiveness. The solution NH3/H2O has high latent heat, is lightly toxic, and its operation pressures are relatively high; however, its main disadvantage is that it requires the use of a rectifier in order to remove the water vapor from the mixture that exits from the generator before it enters to the condenser. The solution H2O/LiBr is used in heat transformers because the temperature levels are higher in comparison with the solution NH3/H2O. There does not seem to be information in the literature concerning the use of heat exchangers built with disks of tar-impregnated graphite. In this paper we show the experimental results of a vertical stainless steel shell absorber and tar-impregnated graphite disks, which were accommodated internally in a column. This heat exchanger was mounted on an absorber test bench and during the experimental tests the parameters such as solution volumetric flow, the fluid temperatures that simulate waste heat, and concentrations of the working solution were modified. The main focus of this study was heat transfer analysis using a non-metallic material with an experimentally unanalyzed geometry in absorbers. 2. Experimental and measurement instruments 2.1. Absorber The studies on absorbers have been largely carried out in vertical plates, vertical tubes (Fig. 1a) and horizontal tube bundles (Fig. 1b), in which the solution descends in falling film over the plate, the tubes or the tube bundles, where the solution distribution is not uniform over the whole surface. In this study the behavior is analyzed using disks with two different slope angles [18], which allow a better distribution of the solution throughout the whole area (Fig. 1c). Fig. 2 shows the vertical absorber built with 18 tar-impregnated graphite disks. The absorber is formed externally by a 316L stainless steel tube joined in its upper and lower ends to two bridles coupled to two headers with screws that give the disks column rigidity and hermetic seal (with 0.17 m in diameter and 0.56 m in length). The upper header is internally designed for the solution to be distributed and to wet the upper and lower surface of the disks. Fig. 3 shows the design of a tar-impregnated graphite disk with 0.1 m in outside diameter. The slope of its surfaces with 7° in the upper and 14° in the lower and its rough surface to improve the distribution of the working mixture. This kind of material has a 50–80 W/m-K thermal conductivity [18]. Neoprenes packing were used as a hermetic seal between the disks and the shell. The con-
Subscripts AB absorber CO condenser EV evaporator GE generator i internal in inlet LM logarithmic mean o external out outlet s solution w water Superscript e equilibrium
centrated working mixture enters through the upper part of the absorber and by gravity goes towards the disks where it gets distributed, while the water vapor enters through the lower part of the absorber and gets mixed with the solution descending from the disks. The working mixture then exits through a lateral connection as a diluted working mixture. The cooling water flows in the annular space formed by the tube and the outside disks. 2.2. Mechanism of heat and mass transfer The overall efficiency of absorption machines depends particularly of the absorber on operating conditions and physical characteristics (heat exchange surface and internal geometry). The studies carried out in the absorbers have been theoretical and experimental. In some theoretical models, heat and mass transfer mechanism is analyzed based on formulations of continuity, diffusion and energy. In other studies, the models analyze the liquid film flowing over the horizontal tubes and inside them, there is a fluid flowing removing heat. For this particular case the boundary layer equations were resolved, which are equations of momentum, energy and mass of the liquid film assuming laminar flow [19]. Design tools for absorbers with H2O/LiBr [20] and NH3–H2O mixtures in film falling on smooth tubes have been developed. Theoretical models assume ideal conditions so that the results show significant differences compared with the experimental results compared in 30% [21]. In experimental works, the analysis of the distribution of the solution in different geometries as banks of tubes, vertical tubes and others geometries has been performed. Effects of absorbed water vapor in lithium bromide falling film in vertical tubes of different geometries, temperature gradients and heat and mass transfer coefficients have been tested experimentally. The mechanism of heat and mass transfer in different kinds of absorbers, occurs by the absorption of water vapor into direct contact with the film surface of the H2O/LiBr solution. The absorbers may be concentric tubes, horizontal tube bundle and vertical flat plates. In the concentric tubes the solution flowing through the inner wall of the inner tube or in the annular section, through the outer wall of the inner tube and the fluid that removes heat flows countercurrent through the annular section or inside the pipe, respectively. For the inner tube, the solution is distributed in falling film on the inner wall of the tube while the vapor fills the internal space of the tube. The fluid flow removes heat by the annular section (Fig. 4a). In the horizontal tube bank, the solution falls along the outer surface of the tube forming a falling film of solution and vapor fills the entire internal space of the heat exchanger shell.
31
J. Olarte-Cortés et al. / Experimental Thermal and Fluid Science 46 (2013) 29–36
(a)
(b)
(c)
Fig. 1. Different geometries for absorbers: (a) vertical tubes, (b) horizontal tube banks and (c) graphite disk.
falling film of solution is spread on a surface, whether the surface is vertical, horizontal or inclined. The water vapor fills the entire space and it is absorbed in direct contact by the solution surfaces generating exothermic reaction, and this heat released goes through the wall of the tube, plate or disk for its transfer towards the fluid that removes heat. 2.3. Instrumentation
Fig. 2. Structure of the graphite disks absorber and its connections.
And inside the bundle tube, a flowing fluid removes heat (Fig. 4b). In the vertical plates, the solution is distributed in falling film on one of the surfaces of the plate and the vapor fills the space, while on the other side of the plate a fluid flows in countercurrent removing heat (Fig. 4c). In the disks used in the present work, the solution is distributed over the inner surfaces in falling film and in the annular section the fluid flows removing the heat generated (Fig. 2). Whatever the geometry, the mechanism is the same, a
Table 1 shows the characteristics of the generator, evaporator, condenser and the solution heat exchanger. The components are connected to each other so that the test bench works as a heat transformer. The connection diagram is shown in Fig. 5. The heating supply consists of a water container installed with two tubular electrical resistors of 2 and 3 kW. These electric resistances provide the required heat to the evaporator and generator and a centrifuge pump of 1/2 hp moves the heating water flow in these components. Other pump of the cooling tower removes the heat in the condenser and the heat generated in the absorber. The volumetric flows of the cooling and heating water at the inlet of each component are measured by rotameters—made of 316L stainless steel and placed at the inlet of the components—with a minimum measurement rate of 745 mL/min and a maximum of 14 LPM. The control of this flow is done with valves. The measurement of the H2O/LiBr solution was similar to that of the cooling and heating water, but in order to regulate the variation of the flow a gear pump of positive displacement with control was used. The working solution used in the experimental test was H2O/ LiBr and its circulation into the generator-absorber circuit was carried out by means of two gear pumps of positive displacement with a maximum volumetric flow control of 5.85 LPM. A third gear pump of 4.55 LPM was used to circulate the condensated water from the condenser to the evaporator. Twenty-four type T (copper-constantan) thermocouples of gauge with a fluorinated ethylene propylene resin coating were installed. The thermocouples were introduced in thermowells connected at the inlet and outlet of each component to measure the temperature during the experimental tests. All of the thermocouples were connected to two 34901A channel multiplexer modules. In order to connect these modules to the computer, they were inserted in a commercial data acquisition switch unit with an RS-232 interphase. To measure the pressure in the generator and the absorber, two manometers were used with a measuring range from 0 to 307.8 kPa, and two previously calibrated pressure transducers with a measuring range from 0 to 204.4 kPa, were connected to the multiplexer module channel for registering the data in the computer.
32
J. Olarte-Cortés et al. / Experimental Thermal and Fluid Science 46 (2013) 29–36
Fig. 3. Section of the tar-impregnated graphite disk.
LiBr solution
LiBr solution
Coolant
Falling film
Falling film Falling film
coolant
LiBr solution
coolant
coolant
Water vapor Water vapor Water vapor
(a)
(b)
(c)
Fig. 4. Direct contact of the water vapor and H2O/LiBr in different geometries.
Table 1 Specifications of the components of the experimental bench. Component
Characteristics
Generator
Heat exchanger of shell and horizontal tubes [22], made of 316L stainless steel, with glass windows (measuring 0.17 m in width and 0.4 m in length) on both sides and with a designed heat capacity of 3 kW It is composed internally of a headstock of four tubes (whose external diameter is 0.0063 m and whose length is 0.50 m) with holes in their upper part evenly distributed in order to wet the columns of tubes with the working mixture. It has four columns of seven stainless steel tubes with a surface of heat transfer of 0.768 m2. Concentric tubes helical heat exchanger [23], constructed of 316L stainless steel. The surface of the heat transfer of each exchanger is 0.090, 0.104 and 0.108 m2 respectively.
Evaporator, condenser and solution heat exchanger
2.4. Operation of the heat transformer The experimental bench for absorbers worked as a single stage heat transformer, as shown in Fig. 6. It was conceived for continuous operation: the heating source supplies low quality heat to the generator and the evaporator. This heat supplied at the generator evaporates part of the working fluid of the diluted solution, increasing its concentration. The produced vapor passes from the generator to the condenser, where it is condensed by exchanging heat with an external fluid. The condensate working fluid is pumped to the evaporator (which is at a higher pressure) where it is converted to vapor using low quality heat from the same heat source as the generator. The vapor formed passes to the absorber. On the other hand, the concentrated working mixture in the generator is pumped to the solution heat exchanger before entering the absorber. In the absorber, the concentrated solution absorbs the working fluid in vapor phase, diluting the solution. The diluted solution is pumped to the generator, passing before by the solution heat exchanger, completing the cycle. The solution in the absorber is heated by the exothermic process of working fluid going into
solution. The cooling tower is the external heat sink whose purpose is to remove the high temperature heat from the absorber and the low temperature heat from the condenser. In the outlet of the absorber a pump is used to control the volumetric flow of the working solution. Quantities of heat QGE and QEV are supplied to the generator and evaporator respectively at the evaporator temperature TEV. A quantity of heat QAB is given out from the absorber at the higher absorber temperature TAB together with a quantity of heat QCO from the condenser at the lower condenser temperature TCO [24].
3. Analysis of experimental data All the analyzed experimental data correspond to steady state working conditions. At the beginning of each experimental test, the volumetric flow, the temperatures and the concentrations of the working mixture were established, although they were modified in a certain number of experiments. For the thermal analysis of the absorber, the number of Reynolds and the overall and indi-
J. Olarte-Cortés et al. / Experimental Thermal and Fluid Science 46 (2013) 29–36
33
Fig. 5. Schematic diagram of the experimental bench.
DT LM ¼ ðT s;in T w;in Þ aðT s;in T s;out Þ bðT w;in T w;out Þ
ð3Þ
Another way of obtaining DTLM is by using the registered temperatures of the absorber [10,11]:
DT LM ¼
ðT s;in T w;out Þ ðT s;out T w;in Þ T T w;out In T s;in s;out T w;in
ð4Þ
Or by using the equilibrium temperatures of the solution [8,13,15] in the following equation
DT LM ¼
ðT es;in T w;out Þ ðT es;out T w;in Þ T e T w;out In T s;in e T s;out
Fig. 6. Schematic diagram of the heat transformer.
vidual heat transfer coefficients were calculated from the data registered in each experimental test. The thermophysical properties of the working solution were taken from correlations proposed in the literature with a wide validity range [18]. The thermal load in the absorber is defined as the heat released in the absorber which is removed by the cooling water. This thermal load was thus calculated as:
Q AB ¼ mw Cpw ðT w;out T w;in Þ
ð1Þ
The overall heat transfer coefficient for the absorber was calculated, using the thermal load in the absorber, the heat transfer area, and Logarithmic Mean Temperature Difference as:
U¼
Q AB ADT LM
ð2Þ
where DTLM can be calculated from the following correlation [14] in which the constants ‘‘a’’ and ‘‘b’’ are, respectively, 0.65 and 0.5:
ð5Þ
w;in
The external heat transfer coefficient ho was obtained by taking into account that the configuration of the external surface of the disks with the shell looks like two concentric tubes. The external heat transfer coefficient ‘ho’ was obtained using the experimental data in the Reynolds and Nusselt dimensionless numbers. Finally, the internal heat transfer coefficient hi was calculated using the following equation:
U ¼ do
1 di hi
1 þ
do k
In ddo i
1 ho
ð6Þ
4. Results and discussion Table 2 shows the thermal load range calculated for each component using the Eq. (1). The values calculated from the thermal load take into account the measured temperatures and mass flow rates of the heating and cooling water for each component. In addition, the working fluid and working solution mass flow rates and temperatures are also shown in the table. The temperatures and mass flow rates ranges correspond to all calculated values of the thermal loads presented in this work. Then the thermal load reported is of 625 W corresponding to 19.6 and 25.3 °C (input and output temperatures respectively) and mass flow rate of 0.028 kg/
34
J. Olarte-Cortés et al. / Experimental Thermal and Fluid Science 46 (2013) 29–36
Table 2 Working ranges of the experimental study. Component
Heat capacity W
Mass flow (kg/s) H2O
H2O-LiBr
H2O
H2O-LiBr Input
Output
Absorber
625
1460
0.008
0.028
0.014
0.018
16.8 26.2
24.1 54.1
67.3 75.0
68.4 86.6
Generator
1600
2700
0.096
0.14
0.018
0.023
75.7 84.6
71.5 79.8
56.2 71.4
63.0 70.1
Condenser
1400
1900
0.048
0.07
15.6 25.6
21.0 31.7
Evaporator
800
1800
0.080
0.085
75.6 84.4
72.1 81.0
Thermal load in the absorber (W)
Min
Temperature (°C)
Max
Min
Max
Min input
Max input
Input
Output
Ts,in=71ºC, Tw,in=17-25ºC, mw,in=0.0289 kg/s
Absorber pressure (kPa)
Thermal load of the absorber (W)
Fig. 7. The thermal load obtained with respect to the pressure in the absorber.
Ts,in=67-71ºC, Tw,in=18-19ºC, mw,in=0.0289 kg/s
Mass flowrate of the solution (kg/s) Fig. 8. The experimental thermal load as a function of the input solution’s mass flow.
s. These data is within the range. These conditions were obtained when the steady state was reached. When there is a direct contact between water vapor and the H2O/LiBr solution surface inside the absorber, the system reaches equilibrium at the solution pressure (water vapor enters with a higher pressure). The temperature difference between the water vapor and the H2O/LiBr solution causes water vapor absorption in the solution. This absorption releases heat (thermal load) and the solution comes out diluted. The heat released will decrease until temperatures are equal. When the equilibrium pressure reached is higher, there will be a smaller temperature difference that will generate a lower absorption of water vapor in the solution and the heat generation will decrease [24]. Fig. 7 shows that the ther-
mal load decreases when the pressure is increased at a concentration of 51 wt%. The same effect is observed at 50 wt%, but it is more pronounced. For steady state operation of each experiment the maximum variations of pressure were ±1.5%. In Fig. 8, the behavior of the thermal load in the absorber is plotted versus the solution mass flow. The increase in the thermal load occurs when the solution mass flow rate is increased from 0.012 to 0.017 kg/s with an inlet concentration measured experimentally of 52 wt%. When the solution mass flow rate lifts and it is above 0.017 kg/s, the thermal load decreases. In a similar way it occurs the trend of the concentration of 51 wt%, but the thermal load was lower than the concentration of 52 wt%. The thermal load values for the concentration of 51 wt%, in the range of the mass flow rate of the solution of 0.019–0.021 kg/s, were not possible to obtain experimentally with the input parameters. The initial increase of the thermal load corresponds to a uniform distribution of the solutions on the disks (uniform thickness on both surfaces). This solution is mixed with the entrance vapor of the working fluid. Further increase of the mass flow rate causes a uniform increase film thickness on surface disk models. However, when the solution mass flow rate is increased above 0.017 kg/s, the thermal load drops because the solution does not have a uniform distribution over the lower surfaces of the disks and then the solution falls evenly over the bottom absorber, decreasing the heat generation. Before placing the disks inside, the shell (absorber), an experimental apparatus was constructed to observe the distribution of the solution on both surfaces of the disks. In this experimental test the H2O/LiBr solution was used and the mass flow rate was changed in the tests. As for a small flow, the distribution of the solution is good on both surfaces of the disks, but a higher mass flow rate causes the solution to slide through the center of the disks to the bottom of the container. The overall heat transfer coefficient U was calculated with the absorber thermal load, the heat transfer area and the Logarithmic Mean Temperature Difference DTLM. Three methods were used to calculate the DTLM: Deng and Ma [14] used a linear correlation to calculate DTLM (Eq. (3)), and it was used in this work. In the second method, inlet and outlet temperatures of the absorber were used to calculate the DTLM (Eq. (4)), [10,11] and in the third method, the DTLM is calculated with pressure and concentration data to determine the equilibrium temperatures of the mixture in the absorption process (Eq. (5)), [8,13,15]. Fig. 9 shows the plot of the overall heat transfer coefficient with respect to the thermal load using the three methods above mentioned. The overall heat transfer coefficient obtained from second method was from 167.71 to 317.41 W/m2 K; the range for the linear correlation was from 127.76 to 205.82 W/m2 K and the range obtained from the third method was from 76.94 to 260.38 W/m2 K. The highest overall heat transfer coefficients are obtained with the second method, while
Yoon et al. [11]
Heat transfer coefficient h i (W/m2 K)
Overall heat transfer coefficient U (W/m2 K)
J. Olarte-Cortés et al. / Experimental Thermal and Fluid Science 46 (2013) 29–36
35
Xin=52 % wt, Tw,in=18-24ºC, mw,in=0.0289 kg/s
Deng et al. [14] Miller et al. [8]
Reynolds number (dimensionless) Thermal load of the absorber (W)
Fig. 11. Variation of the internal heat transfer coefficient with respect to different films’ Reynolds numbers.
Ts,in=67-70ºC, Tw,in=17-19ºC, mw,in=0.0289 kg/s
Heat transfer coefficient hi (W/m 2 K)
Thermal load of the absorber (W)
Fig. 9. Overall heat transfer coefficient as a function of the thermal load.
Xs,in=51 % wt, Tw,in=19ºC, mw,in=0.0289 kg/s
Thermal load of the absorber (W) Difference Xin-Xout in the absorber (% wt) Fig. 10. Thermal load of the absorber with respect to the difference between the concentrations.
the lowest coefficients are obtained using the third method. Three methods followed a similar trend but with differences within the thermal load. The generation of heat inside the absorber takes place with the absorption of the steam proceeding from the evaporator and the H2O/LiBr solution that falls as a film on the disks. With the increased of the absorption of the steam by the working mixture the generation of heat increases. Fig. 10 shows that as the difference between the concentrations in the absorber increases so does the thermal load, but when the difference is in the range from 2 to 2.7 wt%, the thermal load decreases. In the experimental test, where the concentrations measured experimentally were of 50, 51 and 52 wt%, the tendency is very similar. The internal heat transfer coefficients were calculated using the Eqs. (1, 2, 3 and 6). When a film mass flow increases, also the thermal load in the absorber is increased. The increased thermal load is due to a good absorption with adequate mass flow rate under controlled conditions of pressure and temperatures. In order to increase the thermal load there must be an increase in the transference of heat, and this latter depends directly on the heat transfer coefficients. Fig. 11 shows the increase of internal heat transfer coefficient from 750 to 954 W/m2 K when Reynolds number is increased from 109 to 144, but when increasing Reynolds number over 144, the internal heat transfer coefficient decreases, for temperature of solution of 68 °C. Additionally, this behavior is also observed for the input solution temperatures of 72 °C. Fig. 12 shows that as the internal heat transfer coefficient increases, the thermal load of the absorber increases with a linear tendency. For heat transfer coefficient of 750 W/m2 K, the thermal
Fig. 12. Behavior of the internal heat transfer coefficient with respect to the thermal load of the absorber.
load obtained was of 936 W, whereas for heat transfer coefficient of 1397 W/m2 K, the corresponding thermal load was of 1285 W corresponding with a temperature of 67 °C. When the temperatures of the input solution are compared among them, they do not present any significant variation with respect to the thermal load, but for a temperature of 68 °C, the maximum heat coefficient obtained was of 1535 W/m2 K. Several studies have been carried out with different types of heat exchangers, as shown in Table 3. These heat exchangers differed in their geometry, size and construction material. The solution used was H2O/LiBr with additives [9] or any other solution [25] to reduce corrosion. The parameters analyzed by the authors, and listed in Table 3, were the thermal load, the film’s Reynolds number and the heat transfer coefficient. At low Reynolds numbers, with a large contact surface, this corresponds to high heat transfer coefficients and high thermal loads [14]. However, low heat transfer coefficients (within the range of 250–600 W/m2 K) were also obtained, with a small contact surface [9,25]. In the present work, in which the contact surface used was of 0.180 m2 in comparison with other works shown in table 3, the thermal load is less than those reported by Deng and Ma [14]. In the literature, it has not been found experimental or theoretical studies of heat exchangers with similar geometries. Only one author [18] presented a study using three different impregnated graphite tubes with external and internal fins, hence this study presents comparative data of turbulence in the tubes. Therefore, the heat exchangers used in this work in order to have a comparison were selected by using the same solution (H2O/LiBr solution), similar ranges of thermal loads and thus the heat exchangers found with these features are shown in the table 3. Table 3 shows that heat transfer coefficients are similar in order of magnitude.
36
J. Olarte-Cortés et al. / Experimental Thermal and Fluid Science 46 (2013) 29–36
Table 3 Comparison of parameters obtained by different authors. Author
Contact surface (m2)
Description
Type of absorber
Absorber’s heat capacity (kW)
Film’s adimensional Reynolds number
Heat transfer coefficient hi (kW/m2 K)
Bourouis et al. [25]
0.104
Vertical concentric tubes (SS)
0.25–0.70
50–300
0.25–0.60
Deng and Ma [14] Kim et al. [10] Medrano et al. [9]
0.603 0.149 0.104
Horizontal tubes (Cu) Vertical concentric tubes (SS) Vertical concentric tubes (SS)
3.6–5.25 0.4–1–0
12–39 50–150 75–325
1.7–3.15 0.9–1.42 0.32–0.58
Miller and Perez-Blanco [8] Yoon et al. [26] Present work
0.125
LiBr + LiI + LiNO3 + LiCl (falling film) H2O/LiBr (falling film) H2O/LiBr (falling film) H2O/LiBr Li2MoO4 (falling film) H2O/LiBr (falling film)
Vertical tube (glass)
0.64–0.91
70–350
1.25–1.40
0.180
H2O/LiBr (falling film) H2O/LiBr (falling film)
Helical Tar-impregnated graphite disks
0.62–1.46
7–30 110–200
0.44–0.60 0.62–1.53
5. Conclusions Experimental tests were carried out in a tar-impregnated graphite disks absorber installed in a testing bench which worked as a heat transformer. The absorber has a compact design with an area of transfer of 0.180 m2, with only a height of 0.56 m in comparison to other designs. The thermal load obtained in the absorber during the experimental tests ranges from 625 to 1460 W and the internal heat transfer coefficient in the absorber ranges from 723 to 1535 W/ m2 K. The film’s Reynolds number ranges from 109 to 200. The Reynolds number is increased in the solution in the range from 110 to 144 causing an increase in the internal heat transfer coefficient, where the maximum value is of 954 W/m2 K, when Reynolds number is at about 144, but when Reynolds number increase is over 147, the internal heat transfer coefficient decreases. When the internal heat transfer coefficient is increased from 723 to 1535 W/m2 K, the thermal load is also increased from 977 to 1308 W. The analysis carried out in this work obtained parameters from the absorber and their comparison with previous reports show a good approach. The thermal load presents similar values to the obtained by Medrano et al. [9] and Miller and Perez-Blanco [8], but the internal heat transfer coefficient presents values over those obtained by Bourouis et al. [25], Medrano et al. [9] and Yoon et al. [26]. Acknowledgements The authors wish to thank the Consejo Nacional de Ciencia y Tecnología (CONACyT) for the fellowship awarded to J. Olarte Cortés, the Centro de Investigación en Ingeniería y Ciencias Aplicadas (CIICAp) for its technical support and Dr. David Juárez and Dr. Alfredo Hernández for their valuable comments. The graphite disks were designed and constructed by Dr. J. Torres Merino in collaboration with Le Carbone, Lorraine, France. References [1] X. Ma, J. Chen, S. Li, Q. Sha, A. Liang, W. Li, J. Zhang, G. Zheng, Z. Feng, Application of absorption heat transformer to recover waste heat from a synthetic rubber plant, Applied Thermal Engineering 23 (2003) 797–806. [2] A. Huicochea, J. Siqueiros, R.J. Romero, Portable water purification system integrated to a heat transformer, Desalination 165 (2004) 385–391. [3] J.I. Yoon, T.T. Phan, Ch.G. Moon, P. Bansal, Numerical study on heat and mass transfer characteristic of plate absorber, Applied Thermal Engineering 25 (2005) 2219–2235. [4] M. de Vega, J.A. Almendros-Ibañez, G. Ruiz, Performance of a LiBr–water absorption chiller operating with plate heat exchanger, Energy Conversion and Management 47 (2006) 3393–3407.
[5] N. Nishimura, T. Nomura, H. Ivota, N. Uedono, Development of vertical plate type absorber, in: Proc. of the Int. Sorption Heat Pump Conf, Shanghai, China, 2002, pp. 310–313. [6] W. Rivera, Experimental evaluation of a single-stage heat transformer used to increase solar pond’s temperature, Solar Energy 69 (2000) 369–376. [7] D.T. Rose, C.A. Zuritz, H. Perez-Blanco, Thermodynamic analysis and pilot plant design for a solar assisted double-effect refrigeration absorption cycle, International Absorption Heat Pump Conference 31 (1993) 109–116. [8] W.A. Miller, H. Perez-Blanco, Vertical-tube aqueous LiBr falling film absorption using advanced surfaces, International Absorption Heat Pump Conference 31 (1993) 185–202. [9] M. Medrano, M. Bourouis, A. Coronas, Absorption of water vapour in the falling film of wáter-lithium bromide inside a vertical tube at air-cooling thermal conditions, International Journal of Thermal Sciences 41 (2002) 891–898. [10] J. Kim, K. Cho, Analytical and experimental study on the absorption performance in the vertical absorber, in: Proc. of the Int. Sorption Heat Pump Conf., Shanghai, China, 2002, pp. 315–319. [11] J.I. Yoon, E. Kim, K.H. Choi, W.S. Seol, Heat transfer enhancement with a surfactant on horizontal bundle tubes of an absorber, International Journal of Heat and Mass Transfer 45 (2002) 735–741. [12] B.P. Schwarzer, M.S. Rahbar, P. Le Goff, A spiral fin tube: a novel type of falling film heat and mass exchanger, International Absorption Heat Pump Conference 31 (1993) 179–184. [13] I. Greiter, A. Wagner, V. Weiss, G. Alefeld, Experimental investigation of heat and mass transfer in a Horizontal-tube falling film absorber with aqueous solutions, International Absorption Heat Pump Conference 31 (1993) 225–232. [14] S.M. Deng, W.B. Ma, Experimental studies on the characteristics of an absorber using LiBr/H2O solution as working fluid, International Journal of Refrigeration 22 (1999) 293–301. [15] I. Kyung, K.E. Herold, Y.T. Kang, Experimental verification of H2O/LiBr absorber bundle performance with smooth horizontal tubes, International Journal of Refrigeration 30 (2007) 582–590. [16] F.P. Incropera, D.P. De Witt, Fundamentals of heat and mass transfer, third ed., 1990. [17] S. Jiangzhou, R.Z. Wang, Experimental research on characteristics of corrosionresisting nickel alloy tube used in triple-effect LiBr/H2O absorption chiller, Applied Thermal Engineering 21 (2001) 1161–1173. [18] J.M. Torres, Contacteurs gaz–liquide pour pompes á chaleur a absorption multi-étagées, These de doctorado, L’ institute National Polytechnique de Lorraine, Francia, 1997. [19] E.D. Rogdakis, V.D. Papaefthimiou, D.C. Karampinos, A realistic approach to model LiBr–H2O smooth falling film absorption on a vertical tube, Applied Thermal Engineering 23 (2003) 2269–2283. [20] G.A. Florides, S.A. Kalogirou, S.A. Tassou, L.C. Wrobel, Design and construction of a LiBr–water absorption machine, Energy Conversion and Management 44 (2003) 2483–2508. [21] I. Yuekse, W. Schluender, A model of an ammonia–water falling absorber, ASHRAE Transactions 94 (1988) 467–483. [22] M.R. Salimpour, Heat transfer coefficients of shell and coiled tube heat exchangers, Experimental Thermal and Fluid Science 33 (2009) 203–207. [23] T.J. Rennie, V.G.S. Raghavan, Experimental studies of a double-pipe helical heat exchanger, Experimental Thermal and Fluid Science 29 (2005) 919–924. [24] F.A. Holland, J. Siqueiros, S. Santoyo, C.L. Heard, E. Santoyo, Water Purification Using Heat Pumps, E&FN Spon, UK, 1999. [25] M. Bourouis, M. Vallés, M. Medrano, A. Coronas, Absorption in the film of water–(LiBr + LiI + LiNO3 + LiCl) in a vertical tube at air-cooling thermal conditions, International Journal of Thermal Sciences 44 (2005) 491–498. [26] J.I. Yoon, O.K. Kwon, P.K. Bansal, Ch.G. Moon, H.S. Lee, Heat and mass transfer characteristics of a small helical absorber, Applied Thermal Engineering 26 (2006) 186–192.