Design & Construction of LiBr - H2O Machine
Design & Construction of LiBr - H2O Machine
Design & Construction of LiBr - H2O Machine
www.elsevier.com/locate/enconman
a,*
Department of Mechanical Engineering, Higher Technical Institute, P.O. Box 20423, Nicosia 2152, Cyprus
b
Department of Mechanical Engineering, Brunel University, Uxbridge, Middlesex, UB8 3PH, UK
Received 22 July 2002; accepted 11 December 2002
Abstract
The objective of this paper is to present a method to evaluate the characteristics and performance of a
single stage lithium bromide (LiBr)water absorption machine. The necessary heat and mass transfer
equations and appropriate equations describing the properties of the working uids are specied. These
equations are employed in a computer program, and a sensitivity analysis is performed. The dierence
between the absorber LiBr inlet and outlet percentage ratio, the coecient of performance of the unit in
relation to the generator temperature, the eciency of the unit in relation to the solution heat exchanger
area and the solution strength eectiveness in relation to the absorber solution outlet temperature are
examined. Information on designing the heat exchangers of the LiBrwater absorption unit are also presented. Single pass, vertical tube heat exchangers have been used for the absorber and for the evaporator.
The solution heat exchanger was designed as a single pass annular heat exchanger. The condenser and the
generator were designed using horizontal tube heat exchangers. The calculated theoretical values are
compared to experimental results derived for a small unit with a nominal capacity of 1 kW. Finally, a cost
analysis for a domestic size absorber cooler is presented.
2003 Elsevier Science Ltd. All rights reserved.
Keywords: Absorption cooling; Lithium bromide; Sensitivity analysis; Heat exchangers
1. Introduction
In hot climates, the heating and cooling demand of domestic dwellings can be reduced substantially with various measures such as good insulation, double glazing, use of thermal mass and
ventilation. However, due to the high summer temperatures, the cooling demand cannot be reduced to the level of thermal comfort with passive and low energy cooling techniques, and
*
0196-8904/03/$ - see front matter 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0196-8904(03)00006-2
2484
Nomenclature
A
CIN
COUT
CEQ
Di , Do
f
F
Fi , Fo
g
hi , ho
hfg
hm
hs
k
K1
K2
kl
M
NuD
P
Pr
Prs
Re
ReD
U
Vm
Tv
Tw
X
Greek
C
d
DTln
DT0
DTL
q
ql
qv
l
ll
m
2485
therefore, an active cooling system is required. It is preferable that such a system is not powered
by electricity, the production of which depends entirely on fuel.
Solar energy, which is available in such climates, could be used to power an active cooling
system based on the absorption cycle. Lithium bromide (LiBr)water absorption units are the
most suitable for solar applications, since low cost solar collectors may be used to power the
generator of the machine. Such absorption units though, are not yet readily available in small
residential sizes. After a search in the world market, only one manufacturer was found commercially producing LiBrwater absorption refrigerators (Yazaki of Japan). Therefore, the possibility of producing absorption air conditioning systems in small sizes for residential buildings
and the economics of using such a refrigerator, assisted by solar energy, needs to be investigated.
The objective of this paper is to present a method to evaluate the characteristics and performance of a single stage LiBrwater absorption machine. The necessary heat and mass transfer
equations and appropriate equations describing the properties of the working uids are specied.
These equations are employed in a computer program, and a sensitivity analysis is performed.
Information on designing the heat exchangers of the LiBrwater absorption unit is also presented.
Single pass, vertical tube heat exchangers have been used for the absorber and for the evaporator.
The solution heat exchanger was designed as a single pass annular heat exchanger. The condenser
and the generator were designed using horizontal tube heat exchangers. The calculated theoretical
values are compared to experimental results derived for a small unit with a nominal capacity of
1 kW. Finally, a cost analysis for a domestic size absorber cooler is presented.
2. Absorption cooling
Absorption machines are thermally activated, and for this reason, high input (shaft) power is
not required. In this way, where power is expensive or unavailable, or where there is waste, gas,
geothermal or solar heat available, absorption machines provide reliable and quiet cooling [1]. A
number of refrigerantabsorbent pairs are used, for which the most common ones are LiBrwater
and ammoniawater. These two pairs oer good thermodynamic performance, and they are environmentally benign.
Absorption refrigeration system uid pairs should meet a number of important requirements,
which are [1]:
1. Absence of solid phaseThe refrigerantabsorbent pair should not form a solid phase over the
range of composition and temperature to which it will be subjected. The formation of solids
may stop the ow and cause problems to the equipment.
2. Volatility ratioThe refrigerant should be more volatile than the absorbent so that it can be
separated easily by heating.
2486
3. AnityThe absorbent should have a strong anity with the refrigerant under the conditions
in which absorption takes place. This anity allows less absorbent to be circulated for the same
refrigerating eect, and therefore, sensible heat losses are less. Also, a smaller liquid heat exchanger is required to transfer heat from the absorbent to the pressurised refrigerantabsorbent
solution. A disadvantage is that extra heat is required in the generator to separate the refrigerant from the absorbent.
4. PressureModerate operating pressures should be used in order to avoid the use of heavywalled equipment and reduce the electrical power required to pump the uids from the low
pressure side to the high pressure side. Also, very low pressure (vacuum) will require the use
of large volume equipment and special means of reducing pressure drop in the refrigerant vapour ow.
5. StabilityHigh chemical stability is required to avoid the undesirable formation of gases, solids
or corrosive substances.
6. CorrosionThe uids should be non-corrosive. If the uids are corrosive, corrosion inhibitors
should be used, which may inuence the thermodynamic performance of the equipment.
7. SafetyIdeally, the uids must be non-toxic and non-inammable.
8. Latent heatTo keep the circulation rate of the refrigerant and absorbent at a minimum, the
latent heat of the refrigerant should be as high as possible.
The ammoniawater pair is not suitable for use with solar collectors because of the high
temperature needed in the generator (125170 C). This temperature can only be obtained with
medium concentration ratio parabolic collectors, which have increased maintenance requirements
due to the need for tracking the sun.
The generator temperatures needed for the LiBrwater pair are lower (75120 C). These
temperatures can be achieved with high performance at plate collectors, compound parabolic
collectors and evacuated tube collectors that are of lower cost and easier to install and operate
than medium concentration ratio parabolic collectors.
2487
CONDENSER
7
3
4
GENERATOR
SOLUTION
HEAT
EXCHANGER
EXPANSION
VALVE
10,11
5
6
ABSORBER
EVAPORATOR
2
PUMP
Qc
Qg
CONDENSER
GENERATOR
7
P
R REFRIGERANT
E FLOW
S RESTRICTOR
S
U
9
R
E
SOLUTION HEAT
EXCHANGER
2
PUMP
1
10
EVAPORATOR
Qe
11
SOLUTION
FLOW
RESTRICTOR
ABSORBER
Qa
TEMPERATURE
With reference to the numbering system shown in Fig. 1, at point (1), the solution is rich in
refrigerant and a pump (2) forces the liquid through a heat exchanger to the generator (3). The
temperature of the solution in the heat exchanger is increased.
2488
7
4
3
5
9,10
1,2
CRYSTALLISATION LINE
TEMPERATURE
In the generator, thermal energy is added and the refrigerant boils o the solution. The refrigerant vapour (7) ows to the condenser, where heat is rejected as the refrigerant condenses.
The condensed liquid (8) ows through a ow restrictor to the evaporator (9). In the evaporator,
the heat from the load evaporates the refrigerant, which ows back to the absorber (10). A small
portion of the refrigerant leaves the evaporator as liquid spillover (11). At the generator exit (4),
the uid consists of the absorbentrefrigerant solution, which is cooled in the heat exchanger.
From points (6)(1), the solution absorbs refrigerant vapour from the evaporator and rejects heat
through a heat exchanger.
The above procedure can also be presented in a Duhring chart (Fig. 3). This chart is a pressure
temperature graph where the diagonal lines represent constant LiBr mass fraction, with the pure
water line at the left and crystallisation line at the right.
4. Design of a single eect lithium bromidewater absorption cycle system
To perform estimations of equipment sizing and performance evaluation of a single-eect LiBr
water absorption cooler, basic assumptions and input values must be considered. With reference
to Figs. 13, the basic assumptions are:
1.
2.
3.
4.
5.
6.
7.
The method of design is demonstrated below. The design parameters considered are listed in
Table 1.
2489
Table 1
Design parameters for the single eect LiBrwater absorption cooler
Parameter
Capacity
Evaporator temperature
Generator solution exit temperature
Weak solution mass fraction
Strong solution mass fraction
Solution heat exchanger exit temperature
Generator (desorber) vapour exit temperature
Liquid carryover from evaporator
Symbol
Q_ e
T10
T4
X1
X4
T3
T7
m_ 11
Example-value
10 kW
6 C
90 C
55% LiBr
60% LiBr
65 C
85 C
0:025m_ 10
Since the evaporator output power Q_ e is 10.0 kW and m_ 11 2:5%m_ 10 the mass ow rates can be
calculated. The results are shown in Table 2.
4.2. Absorber analysis
Since the values of m_ 10 and m_ 11 are known, mass balances around the absorber give:
m_ 1 m_ 10 m_ 11 m_ 6
x1 m_ 1 x6 m_ 6
and
2490
Table 2
Data for single eect LiBrwater cooling system
Point #
h (kJ/kg)
m_ (kg/s)
p (kPa)
T (C)
X (%LiBr)
1
2
3
4
5
6
7
8
9
10
11
83
83
145.4
212.2
144.2
144.2
2628
185.3
185.3
2511.8
23.5
0.053
0.053
0.053
0.0486
0.0486
0.0486
0.0044
0.0044
0.0044
0.0043
0.00011
0.934
9.66
9.66
9.66
9.66
0.934
9.66
9.66
0.934
0.934
0.934
34.9
34.9
65
90
54.8
44.5
85
44.3
6
6
6
55
55
55
60
60
60
0
0
0
0
0
Remarks
Sub-cooled liquid
Superheated steam
Saturated liquid water
Saturated vapour
Saturated liquid water
The mass fractions x1 and x6 in Eq. (4) are inputs, and therefore, m_ 1 and m_ 6 can be calculated. The
results are shown in Table 2.
The heat transfer rate in the absorber can be determined from the enthalpy values at each of the
connected state points. At point (1), the enthalpy is determined from the input mass fraction
(55%) and the assumption that the state is saturated liquid at the same pressure as the evaporator
(0.934 kPa). This value is h1 83 kJ/kg. The enthalpy value at point 6 is determined from the
throttling model, which gives h6 h5 .
The enthalpy at point 5 is not known but can be determined from the energy balance on the
solution heat exchanger, assuming an adiabatic shell as follows:
m_ 2 h2 m_ 4 h4 m_ 3 h3 m_ 5 h5
The temperature at point 3 is an input value (65 C), and since the mass fraction for points 1 to 3
is the same, the enthalpy at this point is determined as 145.4 kJ/kg. Actually, the state at point 3
may be subcooled liquid. However, at the conditions of interest, the pressure has an insignicant
eect on the enthalpy of the subcooled liquid, and the saturated value at the same temperature
and mass fraction can be an adequate approximation. The enthalpy at point 2 is determined from
an isentropic pump model. The minimum work input w can, therefore, be obtained from:
w m_ 1 v1 p2 p1
In Eq. (6), it is assumed that the specic volume (v, m3 /kg) of the liquid solution does not change
appreciably from point (1)(2). The specic volume of the liquid solution can be obtained from a
curve t [3] (see Appendix A).
For the present study, since all variables are known (Table 2), the pump power can be calculated as:
w 0:29 W
Eq. (5) can now be solved for the unknown enthalpy value at point 5, giving h5 144:2 kJ/kg. The
temperature at point 5 can also be determined from the enthalpy value and is 54.8 C.
2491
Q_ e
Q_ g
10
Table 3
Energy ows at the various components of the system
Description
Capacity (evaporator output power)
Pump minimum work input
Absorber heat, rejected to the environment
Heat input to the generator
Condenser heat, rejected to the environment
Coecient of performance
Symbol
Q_ e
w
Q_ a
Q_ g
Q_ c
COP
kW
10.0
0.29
13.42
14.2
10.78
0.704
45
40
0.9
35
0.8
0.7
30
0.6
25
0.5
20
15
0.4
0.3
10
0.2
0.1
0
45
47.5
50
52.5
55
Absorber inlet LiBr percentage
2492
0
57.5
correlations for the uid properties shown in Appendix A. Such a sensitivity analysis is presented
below.
4.6. Sensitivity analysis
4.6.1. Eect of absorber inlet LiBr percentage ratio
To check this eect, the following conditions were assumed:
1.
2.
3.
4.
5.
6.
7.
8.
Since the absorber exit LiBr percentage ratio is kept constant at 60%, the greater the dierence
between the absorber LiBr inlet and outlet percentage ratios is, the smaller will be the mass circulating in the absorber. Additionally, as seen in Fig. 4, the COP increases with decreasing pump
mass ow. On the other hand, to keep the cycle running at the specied stage, the temperature at
the exit of the absorber has to be maintained at a lower level. However, this presents diculties
with the cooling water temperature of the absorber heat exchanger. Normally, this temperature
may be between 20 and 25 C, which means that the lowest temperature at the exit of the absorber
would be around 30 C.
0.9
Generator pressure (kPa)
28
0.8
24
0.7
20
Generator pressure
0.6
16
0.5
12
COP
0.4
0.3
0.2
0.1
0
65
70
75
80
85
90
95
100
105
110
0
115
2493
As seen in Fig. 5, when the generator temperature is increased, the generator pressure is
also increased, and this has the eect of lowering the COP of the unit, considering that the pressures
and temperatures at other points of the unit are kept constant. The pump mass ow changes
slightly from 0.04 kg/s at a generator exit temperature of 65 C to 0.037 kg/s at 115 C.
4.6.3. Eect of heat exchanger temperatures
To check this eect, the following conditions were assumed:
1.
2.
3.
4.
5.
6.
7.
The solution heat exchanger increases the eciency of the unit. The greater the heat exchanger
area is, the greater its eect will be, since useful energy can be extracted from the generator solution fed to the absorber and be transferred to the solution returning to the generator, where it
2494
Te mperat ure C
70
60
50
40
30
Generator exit solution temp
Absorber inlet solution temp
Absorber oulet solution temp
Generator inlet solution temp
20
10
0
0.7 2
0.74
0.76
0.78
0.8
0.82
0.84
COP
will be heated to the evaporating point. As is observed in Fig. 6, the COP of the unit under the
specied conditions is increased from 0.72 to a maximum of 0.84.
4.6.4. Eect of solution strength
To check the solution strength eectiveness, a constant dierence of 6% between the absorber
inlet LiBr percentage ratio and absorber outlet ratio was considered. Also, the following conditions were assumed:
1. Evaporator capacity 10 kW.
2. Evaporator temperature 6 C.
3. Generator solution exit temperature, T 4 75 C.
0.8
60
0.7
50
0.6
40
0.5
0.4
30
absorber inlet LiBr%
Generator pressure
Absorber exit temp
COP
Pump mass flow
20
10
0.3
0.2
0.1
0
52
54
56
58
60
62
Absorber exit LiBr percentage ratio
64
0.9
70
2495
5. Crystallization
LiBr is a salt, and in its solid state, it has a crystalline structure. When LiBr is dissolved in
water, there is a specic minimum solution temperature for any given salt concentration. The salt
begins to leave the solution and crystallize below this minimum temperature.
In an absorption machine, if the solution concentration is too high or the solution temperature
is reduced too low, crystallization may occur. This is most likely to occur in the solution heat
exchanger, interrupting the machine operation. 1 In such a case, the concentrated solution temperature needs to be raised above its saturation point so that the salt crystals will return to the
solution, freeing the machine.
The most frequent cause of crystallization is air leakage into the machine, which results in
increased pressure in the evaporator. This, in turn, results in higher evaporator temperatures and,
consequently, lower capacities. At high load conditions, the control system increases the heat
input to the generator, resulting in increased solution concentrations to the point where crystallization may occur. Non-absorbable gases, like hydrogen, produced during corrosion, can also
be present, which reduce the performance of both the condenser and the absorber [4]. A direct
method for keeping the required pressure is to evacuate the vapour space periodically with a
vacuum pump.
Excessively cold condenser water, coupled with a high load condition, is another cause for
crystallization. In modern designs, the cooling tower water temperature is allowed to oat with
variations of load and outdoor air temperature. In this way, by using cooling water temperatures
that are below design, the unit performance is improved. However, in the solution heat exchanger
under high load conditions, the entering temperature of the dilute solution may be low enough to
reduce the temperature of the highly concentrated solution returning from the generator to the
crystallization point.
A third reason is electric power failure. During normal shutdown, the machine undergoes a
dilution cycle, which lowers the concentration of the solution throughout the machine. In such a
case, the machine may cool to ambient temperature without crystallization occurring in the
Absorption Refrigeration. Trane Air Conditioning Clinic. The Trane Company, La Crosse, Wisconsin 54601.
2496
solutions. Crystallization is most likely to occur when the machine is stopped while operating at
full load, when highly concentrated solutions are present in the solution heat exchanger.
6. Heat exchangers sizing
In single pass heat exchangers, the temperature dierence DT between the hot and the cold
uids is not constant but varies with distance along the heat exchanger. In the heat transfer
analysis, it is convenient to establish a mean temperature dierence (DTm ) between the hot and
cold uids such that the total heat transfer rate Q_ between the uids can be determined from the
following simple expression:
Q_ AU DTm
11
DTm F DTln F
DT0 DTL
ln DT0 =DTL
12
Also, for Eq. (11), the overall heat transfer coecient U based on the outside surface of the tube
is dened as [5]:
U
1
Do =Di 1=hi Do =Di Fi 1=2kDo lnDo =Di Fo 1=ho
13
The above equations can be used in a computer program for designing the unit heat exchangers.
For the design of the heat exchangers, the cooling water inlet and outlet temperatures must be
assumed. The cooling water inlet temperature depends exclusively on the available source of
water, which may be a cooling tower or a well.
6.1. Condenser heat exchanger design
The overall heat transfer coecient is given by Eq. (13). For this equation, the value of the
fouling factors (Fi ; Fo ) at the inside and outside surfaces of the tube can be taken as 0.09 m2 K/kW
[6]. For a copper heat exchanger, k can be calculated from curve tting [5] (Appendix A). The heat
transfer coecients, hi , ho , for the inside and outside ow need to be calculated.
The PetukhovPopov equation in Ref. [7] for turbulent ow inside a smooth tube gives:
NuD
f =8ReD Pr
K1 K2 f =81=2 Pr2=3 1
14
2497
0:25
15
The physical properties in Eq. (15) should be evaluated at the mean wall surface and vapour
saturation temperatures.
The above equation is recommended in the case of condensation on a single horizontal tube,
although a comparison of the average heat transfer coecient for vertical surfaces with that found
by experiments has shown that the measured heat transfer coecient is about 20% higher than the
values suggested by theory [5].
By substituting the above values in Eq. (13), a resulting overall heat transfer coecient (U ) may
be determined.
The logarithmic mean temperature dierence may be evaluated from Eq. (12), and Eq. (11) is
used to evaluate the necessary heat transfer area of the pipe.
6.2. Generator heat exchanger design
The generator provides sensible heat and the latent heat of vaporisation. The sensible heat
raises the inlet stream temperature to the saturation temperature. This amount of heat is normally
about 13% of the total heat required [4]. The heat of vaporisation consists of the heat of vaporisation of pure water and the latent heat of mixing of the liquid solution. Typically, the heat of
mixing is about 11% of the heat of vaporisation for water/LiBr.
The above analysis indicates that the heat to be provided by the generator can be based on the
heat of vaporisation of pure water, increased by about 23% in a typical design.
Although considerable research work has been done in the past on pool boiling of liquids, data
on water/LiBr solutions are not extensive [9]. Experimental results indicate that boiling is not
signicantly aected by the tube diameter but is aected signicantly by the solution concentration. As the solution concentration increases, the heat transfer coecient decreases. Also, the
heat transfer coecient increases as the heat ux increases. Average heat transfer coecients were
found to vary between 1600 and 7500 W/m2 K [9]. For the above work, stainless steel tubes were
used. The tube material also aects the heat transfer coecient as shown by an empirical relation
developed by Rohsenow in Ref. [7] for nucleate boiling. Since no formula is available for calculating the exact heat transfer coecient of the generator heat exchanger, the exact heat transfer
coecient needs to be determined experimentally.
6.3. Solution heat exchanger design
For the solution heat exchanger design, Eq. (13) is used to calculate the overall heat transfer
coecient. As before, the value of the fouling factors (FI ; Fo ) at the inside and outside surfaces of
the tube are taken as 0.09 m2 K/kW [6] and k for copper at a mean temperature may be evaluated.
The heat transfer coecients, hi , ho , for the inside and outside ow are determined as detailed
below.
The LiBr solution properties at a mean temperature are calculated (Appendix A) and Re is
determined from
2498
Re 4m_ =pDl
For laminar ow, the Nusselt number (Nu) is [6]:
Nu hi Di =K 3:66
16
ReDH
l
Al
For laminar ow, as before Nu ho D=K 3:66, and therefore, ho may be evaluated.
By substituting the above values in Eq. (13), the resulting overall heat transfer coecient U
based on the outside surface of the tube is subsequently evaluated.
The logarithmic mean temperature is then calculated, and nally, from Eq. (11), the needed
area of the heat exchanger is determined.
6.4. Tube absorber
For this design, the solution lm can ow downward either on horizontal or on vertical tubes.
The construction of the horizontal tube absorber may be more costly because of the large
length of welds, which may present problems with maintaining a vacuum. For this reason, an
alternative design with vertical tubes, housed in a cylindrical shell, may be employed, as shown in
Fig. 8. The theoretical analysis of the vertical tube absorber is presented below.
6.5. Absorber heat exchanger design
In the case of this study, the water vapour produced in the evaporator (Fig. 6) is absorbed in
the ow of the LiBrwater solution and is not condensing directly on the heat exchanger tubes.
The design of the heat exchanger, therefore, requires values for the heat and mass transfer coecients.
A number of researchers have studied the absorption of water vapour in falling lms of LiBr
water solutions. Morioka et al. [10] conducted experiments on steam absorption for lms owing
down a vertical pipe. They obtained experimental results, which show that for lm Reynolds
numbers in the range of 40400, the heat transfer coecients of the lm are between 1500 and
3000 W/m2 K. The average absorption of mass ux (kg/m2 s) is compared with the numerical
results derived from a laminar ow theoretical model proposed by the authors. The agreement of
the results is good for lm Reynolds numbers up to 100, but the experimental values are far higher
above lm Reynolds numbers of 200.
Grossman [11] described a theoretical analysis of the combined heat and mass transfer process
in the absorption of gas or vapour into a laminar liquid lm. Simultaneous equations are described that give the temperature and concentration variations at the liquidgas interface and at
the wall. A constant temperature and an adiabatic wall case were considered. The Nusselt and
2499
Solution outlet
Sherwood numbers were found to depend on the Peclet and Lewis numbers as well as on the
equilibrium characteristics of the working uids.
Conlisk [12] developed a design procedure for predicting the absorption capacity of a given
tube based on the governing geometrical and physical parameters. The theoretical approach
developed can predict the amount of mass absorbed in a given length of tube.
Patnaik et al. [13] presented a model based on the solution of dierential equations to calculate
the axial solution concentration and temperature distributions along a vertical tube absorber. The
absorption of water vapour into the falling lm of the solution of LiBr was modelled, employing
equations extracted from the literature, incorporating information on wavy-laminar ows. The
usefulness of the model was demonstrated by generating absorber performance charts.
A practical model for absorption of vapours into a laminar lm of water and LiBr falling along
a constant temperature vertical plate was described in Andberg and Vliet [14]. The model developed considers non-isothermal absorption, and the equations showed good agreement with
experimental results. The objective of the study was to develop an eective model of the absorption process in order to facilitate the study of absorber design from a theoretical viewpoint.
For this reason, some simplications were made, like assuming that the bank of tubes of an actual
absorption unit are replaced with a constant temperature vertical at plate. Also, because of the
complexity of the problem involving the solution of momentum, energy and diusion equations
with their boundary conditions, a simplied method was developed for determining the various
quantities involved. Because of the simplicity and the good agreement with experimental results,
this method was chosen for determining the number of absorber tubes required.
2500
The independent variables, which aect the absorption of vapours, are solution mass ow rate,
solution inlet concentration, absorber pressure and wall temperature.
The length of plate (L) is correlated to the solution mass ow rate by the expression:
L amb
17
where a 132ln100 AP =86:0, b 1:33, m mass ow rate per unit width of plate (kg/m s)
and Ap absorption percentage.
The absorption percentage (Ap ), is dened as:
Ap
CIN COUT
100
CIN CEQ
18
Determination of the equilibrium concentration, CEQ , requires the solution of the following set of
expressions:
A 2:00755 0:16976X 3:133362 0:001X 2 1:97668 0:00001X 3
B 321:128 19:322X 0:374382X 2 2:0637 0:001X 3
C 6:21147
D 2886:373
E 337269:46
T 0 2E=D D2 4EC LOGP =6894:80:5 459:72
TW 5=9AT 0 B 32
The above set of expressions requires an iterative type of solution to nd CEQ , given TW and P .
In the case of this study, TW and P are known, and therefore, CEQ may be evaluated. Eq. (18) may
then be solved to calculate AP .
Assuming a certain length for every tube of the absorber, Eq. (17) gives the mass ow m, and
since the tube external area per meter length is known, the required number of tubes can be
calculated.
The next step is to check the area of tubes needed to cool the solution to the required level.
Patnaik et al. [13] suggest that Wilkes correlation, valid for a constant heat ux wall with progressively decreasing dierence from isothermal wall outside the entrance region, can be used for
the falling lm. It is assumed that the ow is fully developed in a wavy, laminar regime and that
the bulk solution temperature prole is linear with respect to the transverse coordinate. Wilkes
correlation is:
ks
0:029Res 0:53 Prs0:344
19
hs
d
The lm thickness d (m) is given by:
1=3
3lC
20
d
q2 g
and the solution Reynolds number (Res ) for the vertical tube is:
Res 4C=l
21
2501
Water outlet
Entering
water
Entering
water
Vapour
outlet
Falling film
Water inlet
In the case of this study, the mean properties of the solution are evaluated. Assuming the number
of pipes to be used and substituting the above values into Eqs. (19)(21), a solution convective
heat transfer coecient hs results.
Subsequently, the cooling water properties at the mean temperature are evaluated and substituted into Eq. (14). By replacing Nu hi Di =K, hs is determined. By substituting the above
values into Eq. (13), the resulting overall heat transfer coecient (U ) based on the outside surface
of the tube is also determined. DTln also needs to be evaluated, and from Eq. (11), the resulting
length of each pipe is determined. This length needs to be equal to the initial length assumed. The
above procedure may be repeated until the pipe length matches.
6.6. Evaporator heat exchanger design
To facilitate construction, the evaporator heat exchanger may be constructed in a similar way
to the absorber heat exchanger. Fig. 9 shows a schematic representation of the vertical tube
evaporator. Water passing through the evaporator tubes supplies the required heat to vaporise the
falling lm of water around every tube.
A search of the literature has shown that the preferred construction method is to allow the liquid
to enter inside a tube. The uid inside the tube is heated by the run of uid at the outer surface of the
tube, so that progressive vaporisation occurs. The heat transfer coecient increases with distance
from the entrance, since heat is added continuously to the uid. It is also not yet possible to predict
all of the characteristics of this process quantitatively because of the great number of variables upon
which the process depends and the complexity of the various two phase ow patterns that occur as
the quality of the vapourliquid mixture increases during vaporisation [7]. Therefore, in the case of
this study, the mean heat transfer coecient needs to be determined experimentally.
2502
Table 4
LiBrwater absorption refrigeration system calculations based on a generator temperature of 75 C and a solution heat
exchanger exit temperature of 55 C.
Point
H (kJ/kg)
m_ (kg/s)
P (kPa)
T (C)
%LiBr (X)
1
2
3
4
5
6
7
8
9
10
11
83
83
124.7
183.2
137.8
137.8
2612.2
131.0
131.0
2511.8
23.45
0.00517
0.00517
0.00517
0.00474
0.00474
0.00474
0.000431
0.000431
0.000431
0.000421
0.000011
0.93
4.82
4.82
4.82
4.82
0.93
4.82
4.82
0.93
0.93
0.93
34.9
34.9
55
75
51.5
44.5
70
31.5
6
6
6
55
55
55
60
60
60
0
0
0
0
0
Description
Symbol
Q_ e
Q_ a
Q_ g
Q_ c
COP
Remarks
Subcooled liquid
Superheated steam
Saturated liquid
Saturated vapour
Saturated liquid
kW
1.0
1.28
1.35
1.07
0.74
Table 5
Overall heat transfer coecients of the various heat exchangers of the 1 kW unit
Heat exchanger
Generator
Condenser
Evaporator
Absorber
Solution heat exchanger
Cooling/heating water
temperature (C)
Entering
Leaving
Theoretically estimated
Actual value
92
27
13
30
88
28.5
11
31
16007500
2980
625
130
2300
3265
195
400
130
7. Experimental results
To investigate the accuracy of the theoretical procedure described above for the design of the
various heat exchangers, a 1 kW model was designed and constructed [15]. The designed conditions chosen are indicated in Table 4.
A summary of the experimental results is presented in Table 5.
8. General observations
Because of corrosion problems, the life expectancy of LiBr absorption machines is approximately
20 years. The presence of oxygen can greatly accelerate the degradation, and thus, leaks must be
avoided. If the vacuum must be broken for any reason, it is important to ll the vapour space with
nitrogen or other inert gas to a pressure above atmospheric to avoid introduction of oxygen. Metals
2503
Table 6
Comparative cost of a 1 kW and a 10 kW LiBrwater absorption unit
Item
Generator and condenser unit
Evaporator
Absorber
Piping
Pumps
Automatic controls, sensors and valves
LiBr salt
Total cost
50
150
150
30
370
600
200
400
1000
1000
100
500
700
600
1550
4300
like copper and iron in the presence of aqueous LiBr, which is an electrolyte, produce ions that leave
the solid surface and combine with oxygen at a distance from the surface. This leaves the solid
surface of the base metal available for more attack. By controlling the solution pH level to be
slightly basic instead of acidic, oxides are formed directly on the solid surface, forming a relatively
stable oxide coating (passivation), protecting the surface. The pH level can be controlled by adding
small amounts of hydrobromic acid (HBr). Since both elements constituting HBr are already
present within the solution, the solution properties are not signicantly altered. Corrosion inhibitors can also reduce corrosion rates. Such inhibitors are lithium chromate (Li2 CrO4 ) lithium molybdate (Li2 MoO4 ) or lithium nitrate (LiNO3 ) in amounts on the order of 1% by solution weight [4].
9. Cost analysis
The unit manufactured for this project is a prototype unit, and the total cost of its construction
is about C 3000. 2 A large part of this amount was spent for purchasing ow meters (C 500),
auxiliary equipment like vacuum pumps, glass tubes etc. (C 500) and experimentation with
dierent types of valves and materials.
The actual cost of a working unit of 1 kW capacity can be as low as C 1550. A breakdown of
this cost is shown in Table 6. Based on the 1 kW unit, the cost of a 10 kW unit, which can cover
the needs of a typical insulated house, can be estimated as C 4300 (Table 6).
It must also be stated that the above unit cannot be produced in a basic workshop. Automatic
or semi-automatic welding machines and personnel specialised on welding components for vacuum are necessary. Also special equipment for weld testing and soundness is necessary. The cost
of such infrastructure is not included in the unit costing. On the other hand, the produced unit is
only a prototype, and a lot of research can be done in order to improve the eciency of the heat
exchangers by using a surfactant additive, called octyl alcohol. Such an additive can increase the
heat transfer performance in the absorber. Also, other types of heat exchanging surfaces, like
rough or nned surfaces may increase the eciency and reduce the size of the unit.
It must also be noted that in order to complete the requirements of a 10 kW unit, a cooling
tower or water from a well is necessary to provide adequate cooling water for the condenser and
2
2504
absorber heat exchangers. The cost of such a tower is about C 400. Also the installation of the
absorption unit and cooling tower will demand about C 100 extra cost for pumps and piping,
compared to the installation cost of a conventional 10 kW electric chiller.
10. Conclusions
A method to evaluate the characteristics and performance of a single stage LiBrwater absorption machine was presented, and the necessary heat and mass transfer equations and appropriate equations describing the properties of the working uids were specied. These equations
were used in a computer program, and a sensitivity analysis was performed. This analysis shows
that the greater the dierence between the absorber LiBr inlet and outlet percentage ratios is, the
smaller will be the mass circulating in the absorber. On the other hand, to keep the cycle running
at a specied stage, the temperature at the exit of the absorber has to be maintained at a lower
level when the absorber exit LiBr percentage ratio is lower.
Considering that the pressures and temperatures at other points of the unit are kept constant,
the COP of the unit is lowered when the generator temperature is increased, leading to an increase
of the generator pressure.
The solution heat exchanger increases the eciency of the unit. The greater the heat exchanger
area, the greater its eect is.
Finally, when checking the solution strength eectiveness for a constant dierence of 6% between the absorber inlet LiBr percentage ratio and absorber outlet ratio, it was found that a smaller
percentage ratio in LiBr solutions would have slightly better results. The main draw back is the
absorber solution outlet temperature (T 1), which will have to be kept at a low temperature for
smaller percentage ratios in LiBr solutions. A reasonable temperature at the exit of the absorber
would be around 30 C, which would result in an absorber outlet LiBr percentage above 58%.
Information on designing the heat exchangers of the LiBrwater absorption unit are also
presented. The calculated theoretical values are compared to experimental results derived for a
small unit with a nominal capacity of 1 kW.
The cost for a nominal 10 kW unit that can cover the needs of a typical insulated house was
estimated to be C 4300. The total cost of an absorption unit together with all necessary secondary devices and installation cost is estimated as C 4800. The price, therefore, of a LiBrwater
absorption refrigeration unit is high compared to a similar capacity electric chiller that is only C
1500. It should be noted, however, that the absorption units oer possibilities of use of renewable
energy sources and waste heat, whereas the electric chiller uses electricity that is produced from
fossil fuels and have harmful eects on the environment.
Appendix A. Material properties
A.1. LiBr solution enthalpy [4]
Range 0% < X < 40% LiBr
CT temperature (C)
h enthalpy (kJ/kg)
2505
X %LiBr
T CT 9=5 32
CA0 33:1054264, CA1 0:13000636, CA2 0:00097096, CB0 1:0090734,
CB1 0:01377507, CB2 0:000085131
h CA0 CA1 X CA2 X 2 T CB0 CB1 X CB2 X 2 2:326
A.2. LiBr solution and refrigerant pressure and temperatures [1]
Range 45% < X < 70% LiBr
Tsol solution temperature (C), range 5 < Tsol < 175 C
Tref refrigerant saturated temperature (C), range 15 < Tref < 110 C
P saturation pressure (kPa)
A0 2:00755, A1 0:16976, A2 0:003133362, A3 0:0000197668
B0 124:937, B1 7:71649, B2 0:152286, B3 0:0007959
RA A0 X 0 A1 X 1 A2 X 2 A3 X 3
RB B0 X 0 B1 X 1 B2 X 2 B3 X 3
C 7:05, D 1596:49, E 104095:5
Log P C D=Tref 273 E=Tref 2732
Tref 2E=D D2 4EC log P 0:5 273
Tsol RB Tref RA
A.3. Enthalpy of LiBr solution [1]
Range 40% < X < 70% LiBr
T solution temperature (C)
h enthalpy (kJ/kg)
Solution temperature range 15 C < T < 165 C
A0 2024:33, A1 163:309, A2 4:88161, A3 0:06302948, A4 0:0002913704
B0 18:2829, B1 1:1691757, B2 0:03248041, B3 0:0004034184, B4 0:0000018520569
C0 0:037008214, C1 0:0028877666, C2 0:000081313015
C3 0:00000099116628, C4 0:0000000044441207
RA A0 X 0 A1 X 1 A2 X 2 A3 X 3 A4 X 4
RB B0 X 0 B1 X 1 B2 X 2 B3 X 3 B4 X 4
RC C0 X 0 C1 X 1 C2 X 2 C3 X 3 C4 X 4
h RA T RB RC T 2
A.4. Density of LiBr solution [3]
Range 20% < X < 60%
T solution temperature (C), 0 < T < 200 C,
qx LiBr solution density (kg/m3 )
X0 X =100
qx 1145:36 470:84X0 1374:79X02 0:333393 0:571749X0 273 T
2506
2507
A.10. Latent heat of condensation of water vapour (curve t Ref. [1]R2 0:9999
hfg latent heat of condensation (kJ/kg)
T saturation temperature (C)
hfg 0:00132635T 2 2:29983657T 2500:43063
A.11. Density of saturated water (curve t Ref. [1]R2 0:9986)
qg density of saturated water (kg/m3 )
T saturation temperature (C)
qg 1=0:00001147965T 4 0:00297197798T 3 0:28077931731T 2 11:83083758T
202:9035477661
A.12. Enthalpy of superheated steam (curve ts from data presented in Ref. [2]R2 0:992)
Hsh enthalpy of superheated steam (kJ/kg),
Tref refrigerant (water) saturated temperature [see Section 2] (C)
Tactual temperature of vapour (C)
P pressure of vapour (kPa)
T degrees of superheat (C)
T Tactual Tref
HSH1 32:508 lnP 2513:2
HSH2 0:00001P 2 0:1193P 2689
Hsh HSH2 HSH1 =100T HSH1
A.13. Water properties (curve ts from data presented in Ref. [2])
T temperature of water (C)
K thermal conductivity (W/m K) [R2 0:9999]
l absolute viscosity (N s/m2 ) (kg/m s) [R2 0:9999]
m kinematic viscosity (m2 /s) [R2 0:9999]
q density (kg/m3 ) [R2 0:9999]
cp specic heat (J/kg K) [R2 0:9994]
K 6:5104167D)10T 4 0:00000018923611T 3 2:671875E)05T 2
0:0027103175T 0:5520119
l 0:0000010:000031538716146T 4 8:913055428199999D)03T 3
0:9795876934T 2 55:4567974T 1791:74424
m 0:00013:17708333331008 T 4 0:0000089652777778T 3 0:00098270833333T 2
0:055322222222T 1:7876666667
q 0:000015451T 3 0:0059003T 2 0:019075T 1002:3052
cp 0:000003216145833T 4 0:000798668982T 3 0:0780295139T 2 3:0481614T
4217:7377
2508