Evaluation of a small capacity, hot water driven, air-cooled [H.sub.2]O-LiBr absorption machine.

A prototype of an air-cooled absorption chiller of about 2 kW for air conditioning using [H.sub.2]O-LiBr has been developed. The unit has been conceived as a laboratory experimental test device with removable components to facilitate modifications of the initial design. Several tests have been carried out under different conditions. The experimental results have been compared with the theoretical ones based on global mass and energy balances over the different components of the system. Detailed simulation models for each heat exchanger have been developed and implemented in the numerical codes to calculate the overall heat transfer coefficients and subcooling values for the whole system simulation. The conclusions reported will lead to future design revisions and improvements to achieve better performance and reliability.

INTRODUCTION

In the last decades, a significant increase in electricity consumption has been produced due to the growth in cooling demand. In order to save energy in cooling systems, it is necessary to develop new technologies to take advantage of alternative energies, e.g., solar energy or waste heat.

In the case of solar cooling installations, the main obstacle that impedes extended use of absorption systems is the large initial investment necessary for both solar collectors and the absorption machine. For low-capacity installations (less than 15 kW), the price of the chiller is the most limiting factor. Therefore, innovative designs of both solar collectors and absorption chillers (Ziegler 2002) are necessary to reduce prices. Moreover, greater simplicity of the absorption cooling installation would reduce the cost of the investment significantly. Therefore, air cooling for the absorber and condenser can be an important issue to avoid using a cooling tower. For solar air-conditioning applications, the refrigerant-absorbent [H.sub.2]O-LiBr is clearly preferred with respect to N[H.sub.3]-[H.sub.2]O due to its higher performance, but the use of LiBr as absorbent implies the risk of crystallization, more important in air-cooled systems.

Different prototypes of air-cooled [H.sub.2]O-LiBr machines have been developed in the past decade for air-conditioning in buildings (Ohuchi et al. 1994; Tongu et al. 1993; Enjoji 1998; Ishino and Kawasaki 1998; Kawakami et al. 1998). However, those machines were gas fired. Therefore, the type of cycle used in those cases was double effect in order to get the maximum advantage of the input energy. LiBr is sometimes used as absorbent together with LiI to overcome the problem of crystallization (Tongu et al. 1993; Ishino and Kawasaki 1998). The main problem for all those machines concerns the high electrical consumption of the fans to create adequate cooling effect on the absorber and the condenser because reduced and compact designs were the main premises. Izquierdo et al. (2001) presented a suitable air-cooled absorber used for absorption systems in public transport that took advantage of the engine's waste heat.

In this paper, a prototype of a low-power absorption cooling machine with an air-cooled condenser and an absorber driven by hot water at the generator is described in detail. Its input energy is hot water below 100[degrees]C. Therefore, a single-effect cycle is suitable for this situation (Figure 1). Although the final performance of the system is slightly reduced due to the air cooling, in both the absorber and condenser the main premise of design has been to keep the electrical fan consumption in these elements within moderate limits, maintaining a reasonable performance. The prototype has a mechanical solution pump, which leads to different behavior of the machine compared to the heat-powered pumps used in the Yazaki commercial water-fired chillers (these chillers experiment with a drastic decrease of the COP due to the decrease of the circulated mass flow at low driving temperatures [Yazaki 1996]). The machine studied has a horizontal-tube, a falling film generator, and an evaporator. The air-cooled absorber and condenser consist of vertical finned tube batteries, whereas the falling film is placed inside the tubes. Different temperature, pressure, and mass flow sensors are located in the most significant points in the machine. This unit has been conceived as a laboratory experimental test device with removable components to facilitate modifications of the initial design.

This paper is divided into two main sections. The first section is devoted to the research approach and methodology. The numerical model of the whole refrigeration cycle and the detailed models used for the heat exchangers are presented. Details of the experimental device developed for validation purposes are also presented. The second section focuses on the experimental data obtained and their comparison with the numerical model.

[FIGURE 1 OMITTED]

RESEARCH APPROACH AND METHODOLOGY

In this section, the model employed for the simulation of the whole absorption cooling system and the experimental apparatus used for its validation are described. The simulation of the whole absorption cycle is based in two levels of modeling: (1) the cycle modeling level, where the main overall values of the cycle are calculated, and (2) the heat exchanger detailed modeling level, where the overall heat transfer coefficients and the subcooling values at the outlet of the heat exchangers, needed by the above-mentioned cycle modeling level, are calculated using the input data of the experimentation.

Numerical Model: Cycle Simulation

In order to study the type of air-cooled absorption system described in the previous section, a mathematical model to simulate the whole system (see Figure 1) has been developed. For each element, overall mass and energy balances are imposed. At the design mode, the nominal cooling capacity (heat exchanged at the evaporator) and the heat transfer efficiencies of the other heat exchangers are data; the overall heat transfer coefficient multiplied by the area of each heat exchanger, the heat exchanged in each element, and the main data of the absorption cycle (enthalpies, temperatures, pressures, LiBr mass fractions, and mass flows) are obtained. At the rating mode, both the area and the overall heat transfer coefficients of the heat exchangers are data. Thus, the effectiveness and heat exchanged in each element are obtained in addition to the main data of the absorption cycle. Negligible heat losses and pressure drops in the connections between elements are assumed. For each i-connection, these conditions, together with mass conservation, produce = [dot.m.sub.i] = [dot.m.sub.i'], [c.sub.i'] = [c.sub.i], [h.sub.i] = [h.sub.i'], and [p.sub.i] = [p.sub.i']. Under steady-state conditions, and neglecting both kinetic and potential energy, the equations considered in the analysis for the different elements are mass, LiBr mass, and energy balance,

[summation][dot.m.sub.out] - [summation][dot.m.sub.in] = 0, (1)

[summation]([dot.m]c)[.sub.out] - [summation]([dot.m]c)[.sub.in] = 0, and (2)

[summation][dot.m][h.sub.out] - [summation][dot.m][h.sub.in] = [dot.Q] + [dot.W], (3)

and, finally, equilibrium relations of LiBr aqueous solution and [H.sub.2]O, enthalpy expressions for the LiBr aqueous solution, [H.sub.2]O, and air:

T = [T.sub.sat,sol](p,c) (4)

T = [T.sub.sat,wat](p) (5)

h [approximately equal to] [h.sub.sol](T,c) (6)

h [approximately equal to] [h.sub.wat](T,[x.sub.g]) (7)

h [approximately equal to] [h.sub.air](T,w) (8)

It is assumed that the enthalpy and entropy of compressed liquid (for LiBr solution and pure [H.sub.2]O) and [H.sub.2]O vapor only depend on the temperature. The equilibrium relations for the [H.sub.2]O-LiBr solution and enthalpy expressions were obtained from ASHRAE (1997) and the entropy expressions from Kaita (2001). For the case of pure [H.sub.2]O, the data were obtained from Furukawa (1991) except for the entropy expressions, which were obtained from NIST (1998). All the UAs are defined according to UA = [dot.Q]/LMTD, where LMTD is the logarithmic mean temperature difference. The UA and subcooling values are calculated in an accurate way by means of the detailed models presented in the next subsection.

Other additional hypotheses are assumed in order to close the nonlinear system of equations generated (see Figure 1):

* The subcooling solution at the absorber outlet (SUB) is fixed as a parameter of the cycle (point 1).

* Equilibrium of the LiBr solution is assumed in the generator outlet (point 4).

* The superheated vapor in the generator outlet is at the same temperature as the outlet solution (points 7 and 4).

* The outlet of [H.sub.2]O at the condenser is saturated liquid (point 8).

* The outlet of [H.sub.2]O from the vessel to the refrigerant pump and from the evaporator to the vessel is saturated liquid (points 10 and 12, respectively).

* The outlet of [H.sub.2]O at the evaporator outlet to the absorber and at the vessel outlet to the equalizer line is saturated vapor (points 13 and 14, respectively).

* The pressure losses are only considered in the expansion valves (points 8'-9 and 5'-6).

* Adiabatic expansion in the valves are conditioned (points 8'-9 and 5'-6).

* Adiabatic pumps are modeled (points 1'-2 and 10'-11).

The corresponding equations for the absorber are given below as examples. In the primary stream, equations of mass, LiBr, energy, and momentum conservation are, respectively,

[dot.m.sub.1] - [dot.m.sub.6'] - [dot.m.sub.13'] - [dot.m.sub.14'] = 0, (9)

[dot.m.sub.1][c.sub.1] - [dot.m.sub.6'][c.sub.6'] = 0, (10)

[dot.m.sub.1][h.sub.1] - [dot.m.sub.6'][h.sub.6'] - [dot.m.sub.13'][h.sub.13'] - [dot.m.sub.14'][h.sub.14'] = [dot.Q.sub.ab], and (11)

[p.sub.1] = [p.sub.6'] = [p.sub.13'] = [p.sub.14']. (12)

For the secondary stream, the equations are [dot.m.sub.15] = [dot.m.sub.16]; [dot.m.sub.16][h.sub.16] - [dot.m.sub.15][h.sub.15] = [dot.Q.sub.ab]; and [p.sub.15] = [p.sub.16]. Equation 6 is applied to relate the temperature, LiBr concentration, and enthalpy at points 1 and 6'; Equation 7 to relate the temperature and enthalpy at points 13' and 14'; and Equation 8 is applied at points 15 and 16. Finally, the equilibrium relation involving the absorber outlet is applied:

[T.sub.6'] = [T.sub.sat,sol]([p.sub.6'],[c.sub.6']) - SUB (13)

where SUB is the subcooling at the absorber outlet. A similar procedure is performed for the rest of the elements of the absorption cycle. As a consequence, a nonlinear system of equations can be schematically written as follows:

[[phi].sub.1] = [f.sub.1]([[phi].sub.1],[[phi].sub.2],...,[[phi].sub.n])

[[phi].sub.2] = [f.sub.2]([[phi].sub.1],[[phi].sub.2],...,[[phi].sub.n])

...

[[phi].sub.n] = [f.sub.n]([[phi].sub.1],[[phi].sub.2],...,[[phi].sub.n]) (14)

The equation system is solved using an iterative procedure (Gauss-Seidel method), where in each function the most updated values of the variables ([[phi].sub.i]) are used according to a previously determined order. Another method of resolution that has been tested (Powell 1970) is based on the combination of a Newton-Raphson method together with a method of minimization of the residual of the equations (the set of equations is equaled to zero). This method was only more effective in some occasions, and its behavior was quite unpredictable, making its use difficult for the designer. Consequently, it was discarded.

Numerical Model: Detailed Simulation of the Heat Exchangers

For the design of the heat exchangers of the air-cooled absorption chiller prototype, detailed numerical codes have been developed. These codes are also used for prediction purposes. They are capable of simulating the heat and mass transfer processes implied under certain hypotheses or using the appropriate empirical information if necessary. The solution heat exchanger has been calculated using standard methods for heat exchangers.

The thermophysical properties of [H.sub.2]O-LiBr were obtained from different sources--density, viscosity, and thermal conductivity from DiGuilio et al. (1990a, 1990b), heat capacity from ASHRAE (1997), mass diffusivity from Reid et al. (1977), and surface tension from Kim et al. (1994). In the case of pure [H.sub.2]O, the source for the thermophysical properties was Furukawa (1991). For the air, the source of the enthalpy expressions was ASHRAE (1997).

Air-Cooled Heat Exchangers (Absorber, Condenser), Air Side. For the simulation of the air side of the air-cooled elements (absorber and condenser), a code developed for simulating the thermal and fluid dynamic behavior of fin-and-tube heat exchangers (Oliet et al. 2000, 2002) was used. This code adopts a strategy of resolution based on discretization of the heat exchanger into macro control volumes around the tubes (see Figure 2). Over these macro volumes, the equations of conservation of mass, momentum, and energy are applied for the air and the energy equation for the solid elements (tubes and fins). The inner flow is a specific subroutine, in this case falling film absorption or pure fluid condensation, which provides the necessary boundary conditions for the calculation of the solid elements. In order to keep computer time consumption within reasonable limits, the mathematical formulation requires the knowledge of some empirical information, specifically the local heat transfer coefficients, the local friction factors, and the mass transfer coefficients for the absorption process. These fundamental or elementary empirical coefficients do not depend on a specific heat exchanger but on the kind of heat transfer surfaces and local flow structure. This information can be obtained from the literature (e.g., Kim et al. [1997]).

Serpentine Heat Exchangers (Generator, Evaporator), Water Side. In the case of the generator and the evaporator, the type of heat exchanger consists of a serpentine of horizontal tubes wetted by LiBr aqueous solution in the case of the generator and pure [H.sub.2]O in the case of the evaporator. The resolution strategy has a similar methodology as for the case of air-cooled heat exchangers (Castro et al. 2002). The heat exchanger is divided into control volumes along the tubes (see Figure 3). In these control volumes, the equations of conservation of mass, momentum, and energy are applied for the inner flow (water stream in this case) and the energy equation for the solid elements (tubes). The calculation of the outer flow is a specific subroutine, in this case falling film desorption or pure fluid falling film evaporation. This subroutine provides the necessary boundary conditions for the calculation of the solid elements. The model needs empirical information of heat transfer and pressure drop coefficients of forced convection inside tubes (Wong 1977). This inner flow is solved in a step-by-step procedure.

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

In the solid elements, the energy equation is solved neglecting the heat transfer outside the wetted area (elbows and manifolds).

Resolution of Absorption/Desorption Processes (Absorber, Generator). In the outer part of the generator, the same subroutine is applied as for the solution side of the air-cooled absorber. In this case, the formulation is adapted to the change in the gravity direction with respect to the tube surface. For this calculation, the following hypotheses are assumed:

* Steady-state flow.

* Newtonian fluid.

* Liquid film in laminar incompressible flow.

* The radius of the tube is much greater than the falling film thickness. Therefore, curvature effects are neglected.

* Constant physical properties at the conditions of the tube inlet.

* Diffusion in flow direction is neglected.

* Convection terms are negligible in the direction orthogonal to the flow.

* No resistance to mass transfer at the interface.

* No resistance to heat transfer at the vapor phase.

* No interfacial shear stress between liquid and vapor phases.

* Thermodynamic equilibrium at the interface.

* Pressure gradients negligible.

* For falling film flowing down inside vertical tubes, initial velocity is considered fully developed laminar regime.

* For falling film flowing down outside horizontal tubes, initial velocity is considered uniform and parallel to tube surface with the same value as the one reached for free drop between tubes.

* Dufour or Soret effects have not been considered.

* Calculation of the fraction of wetted area without taking into account any particular velocity profile of the falling rivulets. The mass flow per unit length is re-adapted to the new area.

According to the above-mentioned hypotheses, the conservation equations of mass, momentum, energy, and LiBr are as follows:

[[partial derivative]u/[partial derivative]x] + [[partial derivative]v/[partial derivative]y] = 0 (15)

u[[partial derivative]u/[partial derivative]x] + v[[partial derivative]u/[partial derivative]y] = gsin[theta] + v[[[partial derivative].sup.2]u/[partial derivative][y.sup.2]] (16)

u[[partial derivative]T/[partial derivative]x] + v[[partial derivative]T/[partial derivative]y] = [alpha][[[partial derivative].sup.2]T/[partial derivative][y.sup.2]] (17)

u[[partial derivative]c/[partial derivative]x] + v[[partial derivative]c/[partial derivative]y] = D[[[partial derivative].sup.2]c/[partial derivative][y.sup.2]] (18)

The boundary conditions are detailed in Figure 4.

This model reasonably predicts absorption and desorption phenomena over smooth surfaces (Grossman 1983; Andberg 1986; Yang and Wood 1992; Wang et al. 1995) if a complete wetted area is achieved and no waves are formed. In order to consider noncomplete wetted area effects, the Mikielewicz and Moszynski (1976) model has been implemented. This model assumes a circular profile of the rivulets and proposes a formula for calculating the fraction of wetted area as a function of the contact angle and mass flow per unit length, with the criteria of minimum energy. This wetted area is then introduced as a factor that reduces the heat and mass transfer with respect to the case of a completely wetted area, without taking into account any particular profile. Consequently, the mass flow rate per unit length is corrected to the new wetted area.

[FIGURE 4 OMITTED]

X = A'/A = f([GAMMA], [zeta]) (19)

The criterion considered for calculating the minimum flow rate to achieve a completely wetted area is the one given by Hobler (1964), also reported by Mickielewicz and Moszynski (1976). For the case of the generator, a correction coefficient for horizontal tubes has been applied (Tang et al. 1991).

The influence of falling drops between tubes is considered in the desorption process at the generator. The models suggested by Kirby and Perez-Blanco (1994) of forming and falling droplets have been applied.

Falling Film Condensation. The correlation proposed by Graham (1969) of dropwise condensation has been used for falling film condensation. Additional effects, such as subcooling or interfacial shear stress, have been neglected due to the conditions of the condenser: low temperature differences (low degree of subcooling) and low vapor velocity inside the tubes.

Falling Film Evaporation. Finally, for the calculation of evaporation processes of pure refrigerant in the outer part of the horizontal-tube falling film evaporator, a semi-empirical model has been implemented (Chyu and Bergles 1987). In this model, four regions of heat transfer are considered, each with its own expression of heat transfer coefficient, according to the falling film flow and the temperature field: (1) stagnation flow region, (2) impingement flow region, (3) thermal developing flow region, and (4) fully developed flow region. The same expression of wetted area (Mickielewicz and Moszynski 1976) with the correction for horizontal tubes (Tang et al. 1991) has been used.

[FIGURE 5 OMITTED]

Experimental Device

Figure 5 schematically shows the prototype with its different circuits. The device consists of two shells, one situated at the top and the other at the bottom, and two air-cooled heat exchangers in the middle, the absorber and condenser. These elements are in series. Thus, the inlet air conditions at the condenser correspond to the outlet air conditions at the absorber (point 17 = point 16). The two shells are divided into two parts, according to the two pressure levels of the absorption cycle. To deal with a wide range of working conditions, this prototype has two extra pumps with respect to the standard single-effect cycle. The first one is for circulating [H.sub.2]O in the evaporator to drive different mass flows of refrigerant (see points 10'-11 in Figures 1 and 2). The second extra pump is placed to circulate the solution from the generator to the absorber (points 23-24). This pump would not be necessary in the case of the absence of measuring elements, i.e., mass flowmeters that produce additional pressure drops in the line. Therefore, the solution expansion valve is not necessary (5' = 6). There is also an extra line for taking samples to a mass spectrometer to measure the presence of noncondensable substances and other undesirable species during the performance of the unit.

Two extra lines are be used in de-crystallization mode. A valve (v1) is opened and activates the bypass line between the generator and absorber in case the normal line between them through the solution heat exchanger is obstructed due to crystallization. A second valve (v2) is also opened, and the connection line between the refrigerant circuit and the strong solution is activated to dilute the crystals.

All the heat exchange surfaces are smooth. The tubes that compose the absorber and condenser are made of copper and their dimensions are 16 mm outer diameter, 1300 mm long (1200 mm finned in the middle part), with a wall thickness of 0.4 mm. Their arrangement is staggered, with a tube distance in the airflow direction (y) of 33 mm and a tube spacing in the direction normal to the airflow (z) of 38 mm. The fins are made of aluminum and are 0.12 mm thick, with longitudes in the y and z directions of 136 mm and 341 mm, respectively, with herringbone wave geometry. Both heat exchangers consist of 36 tubes (four tubes in the y direction and nine tubes in the z direction). The only difference is the fin pitch-2 mm for the absorber and 3 mm for the condenser.

The generator and the evaporator consist of batteries of horizontal smooth tubes made of stainless steel of 12.7 mm outer diameter, 0.89 mm wall thickness, and 300 mm useful length. Their arrangement is inline, and the tube distances in the horizontal and vertical directions are 25.4 mm and 15.7 mm, respectively. The generator consists of 96 tubes (16 tubes in the vertical direction and 6 tubes in the horizontal direction) and the evaporator of 48 tubes (12 tubes in the vertical direction and 4 tubes in the horizontal direction).

In order to achieve high flexibility, different o-ring joints and fittings have been used. Special care was taken to purge incondensable substances (once before every one-day test). Their concentration is measured by means of a quadrupole mass spectrometer. Experimental results in a previous work (Castro et al. 2002) indicated discrepancies with respect to the values of COP and cooling capacity obtained from the theoretical model. Different reasons were reported in that work (theoretical UA of the heat and mass exchange components calculated by means of more detailed models that did not take into account the wetted area in the falling film heat exchangers). However, an additional reason for the discrepancies was observed: the inefficient performance of the drop separator at the generator. The presence of LiBr in the refrigerant circuit was detected, with the consequent reduction of the evaporator performance. This aspect has been completely solved in the present setup.

Figure 6 shows the absorption chiller. Under working conditions, the machine is placed in a wind tunnel where the airflow is measured by means of a hot wire anemometer. A matrix of measurement points has been taken (7 in the horizontal direction x 11 in the vertical direction) and distributed uniformly. Each measurement has been averaged in time.

The tests are performed under steady-state conditions. Therefore, the input conditions of the machine are maintained constant. When all the data measured achieve a stable value, the machine is maintained at these conditions for a long period of time to ensure independence from the initial values. The average heat imbalance of the whole system for all tests considering the primary circuits has been less than 4% and considering the secondary circuits has been estimated at approximately 12%.

[FIGURE 6 OMITTED]

EXPERIMENTAL DATA AND COMPARISON WITH NUMERICAL RESULTS

The evaluation of the performance of the machine has been carried out comparing the experimental results with numerical ones calculated by means of the simulation of the whole cycle. The measured COPs of the cycle (ratio between the heat exchanged at the evaporator and the heat exchanged at the generator) and the cooling capacity (the heat exchanged at the evaporator) are compared with the calculated ones, taking into account the overall heat transfer coefficients multiplied by the heat exchanger areas (UA) obtained from more detailed models. In all simulated cases, the influence of humidity is not significant (approximately 1% of the heat dissipated by the air-cooled heat exchangers). To ensure a similar quality of the points measured, before starting the measurements, the machine is purged until the air concentration measured at the absorber with the mass spectrometer is lower than 2% in volume.

Figure 7 illustrates results of temperatures at different points in the absorption cycle (see Figure 1) compared with theoretical ones calculated with the model described using estimated UAs from the more detailed models. The external working conditions are the following: input hot water temperature [T.sub.19] = 85[degrees]C, outlet chilled water temperature [T.sub.22] = 9[degrees]C, and environmental temperature [T.sub.15] = 35[degrees]C. In general there is an acceptable agreement between the numerical results and the experimental temperature data. The calculated condensation pressure (7605 Pa) is higher than in the measurements (7510 Pa) and the evaporation pressure is lower in the calculation (948 Pa) than in the experiment (1024 Pa). Concerning the LiBr concentrations, the measured concentrations were higher than the calculated ones, both for weak (56.57% LiBr measured vs. 56.31% LiBr calculated) and strong (57.31% LiBr measured vs. 57.07% LiBr calculated) solutions. These discrepancies in the main results of the cycle can be explained by the differences observed between numerical values and experimental data for the UAs (see Table 1).

[FIGURE 7 OMITTED]

The absorption chiller unit was tested with varying external conditions: input hot water temperature, input chilled water temperature, and environment temperature (air-cooling temperature). The solution mass flow was maintained at a value of about 0.048 kg/s. The mass flows of the secondary streams were: 0.12 kg/s for the hot water stream, 0.93 kg/s for the refrigerating airstream, and 0.07 kg/s for the chilled-water stream. Table 2 shows the most relevant experimental data: condensation pressure, evaporation pressure, low LiBr concentration level, and high LiBr concentration level. The concentrations measured were quite far from crystallization, even at 95[degrees]C of input hot water temperature. From these results, it can be concluded that these values have enough margin to reduce the size of the absorber.

For 9[degrees]C and 12[degrees]C (approximately) of output chilled water temperature, Figures 8 and 9 show the evolution of the COP of the machine with respect to the input hot water temperature. In each plot, two families of curves are represented, one at an environmental temperature of 30[degrees]C and the other one at 35[degrees]C. Each family of curves consists of three plots: (1) experimental COP obtained with the data measured in the primary circuits, (2) experimental COP obtained with the data measured in the secondary circuits, (3) theoretical COP obtained using calculated UAs.

There is a good agreement between the experimental values of COP calculated from the primary fluid data and the calculated ones from the mathematical models (medium discrepancy less than 3%), except in test 15 (see Table 2), where the value is underpredicted (about 20%). However, if the experimental data are considered from the secondary circuits, the COP values are underpredicted (about 15% of medium discrepancy), especially near the minimum driving temperature (the maximum discrepancy is found in test 15). The evolution in the COP indicates an important decrease in the performance of the machine between environmental temperatures of 30[degrees]C and 35[degrees]C. The maximum value of the COP is approximately 0.65. For all the tested cases, the measured efficiency of the solution heat exchanger was approximately 0.7.

[FIGURE 8 OMITTED]

[FIGURE 9 OMITTED]

Figures 10 and 11 show the evolution of the machine's capacity. As in the COP plots, one of the figures plots the capacity for output chiller water temperature of 9[degrees]C and the other for 12[degrees]C (approximately). There are also two families of curves in each plot, one at an environmental temperature of 30[degrees]C and the other one at 35[degrees]C. Each family of curves consists of three similar plots.

In general, the calculated capacity is lower than the one measured from the primary circuits, with a medium difference of 7.7%. This is due to the general underestimation of about 15%-20% of the UA related to the generator. If the experimental data from the secondary circuits are considered, the calculated values are higher, about 9.2% of medium difference. The figures show an almost linear dependency between the capacity of the machine and the input hot water temperature. As was observed in the COP plots, there is an important decrease in the capacity of the machine when the environmental temperature increases from 30[degrees]C to 35[degrees]C (between 1000 W and 1500 W). On the contrary, there is an increase of the capacity at 12[degrees]C of output chilled water (about 500 W). Although this temperature could not be valid for fan-coil applications, these working conditions can be suitable for chilled ceiling or wall panels.

[FIGURE 10 OMITTED]

[FIGURE 11 OMITTED]

CONCLUSIONS

A laboratory prototype of an air-cooled absorption chiller of [H.sub.2]O-LiBr was designed, modeled, and tested. From the numerical and experimental results obtained, the following conclusions are highlighted:

1. Reasonable agreement has been obtained between the predicted values and the experimental ones.

2. The maximum COP of the machine was approximately 0.65, with a measured efficiency of the solution heat exchanger of approximately 0.7.

3. A significant decrease of the COP and the capacity of the machine was observed between environmental temperatures of 30[degrees]C and 35[degrees]C, and an increase of the capacity of the machine was observed between output chilled water temperatures from 9[degrees]C to 12[degrees]C.

4. There is an almost linear dependency of the capacity of the machine with the hot water temperature.

5. For all cases, the machine is far from crystallization (maximum concentration measured, 60% LiBr). The only problems of crystallization were derived from inadequate shutdowns of the machine.

6. Due to the cycle configuration (single effect), the temperature lifts between the evaporator and condenser are limited if the maximum hot water temperature is restricted to a maximum value of 95[degrees]C. Consequently, at environmental temperatures higher than 35[degrees]C, it would be necessary to use hot water temperatures higher than 95[degrees]C. In this way, evaporation temperatures of approximately 5[degrees]C-6[degrees]C can be reached. However, at higher evaporator temperatures (8[degrees]C-12[degrees]C), other applications different from fan-coils can be applied (chilled ceiling or wall panels). A single-effect cycle can be used in those cases.

7. The electrical consumption of the fan has been estimated at approximately 250 W. However, it is still possible to reduce it with more optimized designs: increased face area of the air-cooled heat exchangers, absorber-condenser in parallel configuration, other tube-fin arrangements, fan choice, etc.

ACKNOWLEDGMENTS

This research was supported financially by the European Commission (ref. JOE3-CT98-7003) and the Spanish Government, project ref. TIC2003-07970.

NOMENCLATURE 

A = heat exchange area

A' = wetted heat exchange area

c = LiBr mass fraction

D = mass diffusivity

g = gravity

h = enthalpy

H = phase-change heat

i = position of control volume in x direction

j = position of control volume in y direction

k = position of control volume in z direction

l = falling film length

L = length of a side of the air-liquid heat exchanger

[dot.m] = mass flow

n = normal direction to free surface

N = number of control volumes in one direction

p = pressure

[dot.Q] = heat

T = temperature

u = velocity in x direction

U = overall heat transfer coefficient

v = velocity in y direction

w = humidity

[dot.W] = mechanical work

x = coordinate parallel to gravity

X = fraction of wetted area

[x.sub.g] = vapor mass fraction

y = coordinate parallel to airflow

z = coordinate

Symbols

[alpha] = thermal diffusivity 

[phi] = generic variable

[lambda] = thermal conductivity

[nu] = kinematic viscosity 

[rho] = density

[theta] = angle of the normal direction of the wall with respect to gravity

[zeta] = static contact angle

Subscripts and Superscripts

a = absorption

ab = absorber

air = air

cn = condenser

ev = evaporator

gn = generator

i = internal

in = input

out = output

p = pump

sat = saturation

SHX = solution heat exchanger

sol = LiBr aqueous solution

w = wall

wat = water

REFERENCES

Andberg, J.W. 1986. Absorption of vapours into liquid films flowing over cooled horizontal tubes. PhD thesis, University of Texas, Austin.

ASHRAE. 1997. 1997 ASHRAE Handbook--Fundamentals. Atlanta: American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.

Castro, J., L. Leal, P. Pozo, C.D. Perez-Segarra, and C. Oliet. 2002. Development and performance of an air-cooled water-LiBr absorption cooling machine. Proceedings of the Forum International Sur les Energies Renouvelables I, pp. 59-65.

Chyu, M.C., and A.E. Bergles. 1987. An analytical and experimental study of falling-film evaporation on a horizontal tube. Journal of Heat Transfer 109:983-90.

DiGuilio, R.M., R.J. Lee, S.M. Jeter, and A.S. Teja. 1990a. Properties of lithium bromide-water solutions at high temperatures and concentrations--I: Thermal conductivity. ASHRAE Transactions 96(1):702-08.

DiGuilio, R.M., R.J. Lee, S.M. Jeter, and A.S. Teja. 1990b. Properties of lithium bromide-water solutions at high temperatures and concentrations--II: Density and viscosity. ASHRAE Transactions 96(1):709-14.

Enjoji, K. 1998. Development of a direct gas-fired absorption duct-type air-conditioner with an air-cooled absorber and condenser. Proceedings of the International Gas Research Conference, San Diego, pp. 708-18.

Furukawa, M. 1991. Practical expressions for thermodynamic and transport properties of commonly used fluids. Journal of Thermophysics 5(4):524-31.

Graham, C. 1969. The limiting heat transfer mechanisms of dropwise condensation. PhD thesis, Massachusetts Institute of technology, Cambridge, MA.

Grossman, G. 1983. Simultaneous heat and mass transfer in film absorption under laminar flow. International Journal of Heat and Mass Transfer 26(3):357-71.

Hobler, T. 1964. Minimal surface wetting (in Polish). Chemia Stosow 2B:145-59.

Ishino, H., and S. Kawasaki. 1998. Development of air-cooled type absorption chiller/heater using a new working fluid (LiBr+LiI/[H.sub.2]O). Proceedings of the International Gas Research Conference, San Diego, pp. 687-98.

Izquierdo, M., E. Martin, A. Lecuona, and P.A. Rodriguez. 2001. Absorbedor de gotas iguales de flujos paralelos, Patent no. 2161119, Oficina Espanola de Patentes y Marcas.

Kaita, Y. 2001. Thermodynamic properties of lithium bromide-water solutions at high temperatures. International Journal of Refrigeration 24:374-90.

Kawakami, R., N. Nishiyama, M. Aimi, and S. Tongu. 1998. Development of small-sized air-cooled absorption air-conditioner. Proceedings of the International Gas Research Conference, San Diego, pp. 699-707.

Kim, K.J., N.S. Berman, and B.D. Wood. 1994. Surface tension of aqueous lithium bromide + 2-ethyl-1-hexanol. Journal of Chemical Engineering Data 39:122-24.

Kim, N.H., J.H. Yun, and R.L. Webb. 1997. Heat transfer and friction correlations for wavy plate fin-and-tube heat exchangers. Journal of Heat Transfer 119:560-67.

Kirby, M.J., and H. Perez-Blanco. 1994. A design model for horizontal tube water/lithium bromide absorbers. Proceedings of Heat Pump and Refrigeration Systems Design 32:1-10, Analysis and Applications. American Society of Mechanical Engineers.

Mikielewicz, J., and J.R. Moszynski. 1976. Minimum thickness of a liquid film flowing vertically down a solid surface. International Journal of Heat and Mass Transfer 19:771-76.

NIST. 1998. Thermodynamic Properties of Refrigerants and Refrigerant Mixtures Database. Gaithersburg, MD: National Institute of Standards and Technology.

Ohuchi, T., M. Aizawa, R. Kawakami, A. Nishiguchi, T. Hatada, and Y. Kunugi. 1994. A study on a hot-water-driven air-cooled absorption refrigerating machine. Proceedings of the 1994 International Refrigeration Conference at Purdue, Purdue University, West Lafayette, Indiana, pp. 275-80.

Oliet, C., C.D. Perez-Segarra, O. Garcia-Valladares, and A. Oliva. 2000. Advanced numerical simulation of compact heat exchangers. Application to automotive, refrigeration and air conditioning industries. Proceedings of the Third European Congress on Computational Methods in Applied Sciences and Engineering (ECCOMAS), in CD format.

Oliet, C., C.D. Perez-Segarra, S. Danov, and A. Oliva. 2002. Numerical simulation of dehumidifying fin-and-tube heat exchangers. Model strategies and experimental comparisons. Proceedings of the 2002 International Refrigeration Engineering Conference at Purdue, Purdue University, West Lafayette, Indiana, Paper R5-5, in CD format.

Powell, M. 1970. Numerical methods for nonlinear algebraic equations. Vol. A, Hybrid Method for Nonlinear Equations. London: Gordon and Breach Science.

Reid, R., J. Prausnitz, and T. Sherwood. 1977. The Properties of Gases and Liquids. New York: McGraw-Hill.

Tang, J., B. Yu-chi, and Z. Lu. 1991. Minimum wetting rate of film flow on solid surface. Proceedings of the XVIIIth International Congress of Refrigeration II, Montreal, pp. 519-23.

Tongu, S., Y. Makino, K. Ohnishi, and S. Nakatsugawa. 1993. Practical operating of small-sized air-cooled double-effect absorption chiller-heater by using lithium bromide and aqueous. AES-Vol. 31, International Absorption Heat Pump Conference, Purdue University, West Lafayette, Indiana, pp. 125-32.

Wang, C.Q., Z. Lu, D.Q. Li, B. Yu-Chi, and Y.G. Sun. 1995. Heat and mass transfer in falling film generator of lithium bromide absorption refrigerating machine. Proceedings of the 19th International Congress of Refrigeration, The Hague, The Netherlands IIIa:209-14.

Wong, H. 1977. Heat Transfer for Engineers. London: Longman.

Yang, R., and B. Wood. 1992. A numerical modeling of an absorption process on a liquid falling film. Solar Energy 48(3):195-98.

Yazaki. 1996. Manual de Instalacion de Grupos Refrigerantes por Absorcion de Agua Caliente, WFC 10. Yazaki Ltd.

Ziegler, F. 2002. State of the art in sorption heat pumping and cooling technologies. International Journal of Refrigeration 25:450-59.

Jesus Castro, PhD

Assensi Oliva, PhD

Carlos David Perez-Segarra, PhD

Jordi Cadafalch, PhD

Received March 22, 2005; accepted September 5, 2006

Jesus Castro, Assensi Oliva, Carlos David Perez-Segarra, and Jordi Cadafalch are with the Centre Tecnologic de Transferencia de Calor (CTTC), Universitat Politecnica de Catalunya (UPC).

Table 1. UA Comparison for the Case Studied

UA (W/K)  Absorber  Generator  Condenser  Evaporator

Exp.      840.0     422.5       990.4     422.5
Num.      448.3     389.7      1109.3     269.3

Table 2. Main Experimental Data of the Cycle at Different Working
Conditions

      [T.sub.19]    [T.sub.15]    [T.sub.21]    [p.sub.8,cn]
Test  ([degrees]C)  ([degrees]C)  ([degrees]C)  (Pa)

 1    75.0          30.2           8.5          5743
 2    80.0          30.5           8.6          6082
 3    85.0          30.1           8.5          6243
 4    90.0          30.0           8.6          6508
 5    95.0          30.0           8.5          7165
 6    80.0          34.9           8.5          7326
 7    85.0          35.2           8.5          7510
 8    90.0          35.1           8.6          7699
 9    95.0          35.2           8.5          8611
10    75.0          30.5          11.7          6261
11    80.0          30.2          11.7          6731
12    85.0          30.4          11.7          7321
13    90.0          30.3          11.7          7782
14    95.0          30.5          11.6          8254
15    75.0          35.1          11.7          7493
16    80.0          35.2          11.7          7820
17    85.0          35.0          11.7          8031
18    90.0          35.2          11.7          8360
19    95.0          35.2          11.6          8762

      [p.sub.13,ev]  [c.sub.3]   [c.sub.6]
Test  (Pa)           Frac. LiBr  Frac. LiBr

 1     921           0.5480      0.5557
 2     907           0.5566      0.5675
 3     858           0.5661      0.5775
 4     847           0.5707      0.5842
 5     845           0.5799      0.5972
 6    1046           0.5561      0.5603
 7    1024           0.5657      0.5731
 8     946           0.5768      0.5863
 9     973           0.5820      0.5941
10    1221           0.5300      0.5399
11    1190           0.5408      0.5538
12    1206           0.5475      0.5632
13    1151           0.5588      0.5769
14    1128           0.5667      0.5862
15    1241           0.5379      0.5407
16    1210           0.5488      0.5545
17    1178           0.5591      0.5682
18    1167           0.5621      0.5732
19    1136           0.5742      0.5883