International Journal of Scientific & Engineering Research, Volume 4, Issue 4, April-2013 1306

ISSN 2229-5518

Computer-Aided Design and Simulation of

Working Fluid Pairs for Absorption Refrigerators

Eman Abdel-Hakim Tora*

AbstractThe performance of real absorption refrigeration systems is influenced by their working fluids. Despite there being many pair s that can act as working fluids, the conventional pairs (ammoniawater and waterlithium bromide) are still the most used. This paper reports an investigation to find an alternative pair via a three-step approach. The first is computer-aided molecular design using ICAS- ProCAMD software to preliminarily find a pair with analogous thermophysical properties. The second step is plotting Dühring c urves to display the thermodynamic behaviour of the proposed pair, the required data for which are obtained from Aspen plus . Upon obtaining satisfactory results from the second step, the third step is simulation of an absorption refrigerator operating under differe nt conditions. The results show that n-butaneethanol can act as a suitable pair, where butane is the refrigerant and ethanol the absorbent. It has an acceptable pressure range (high pressure of approximately 4 atm, low pressure of around 1 atm) and generator temperature (91° C). In addition, there is no crystallisation probability at 0°C. However, the coefficient of p erformance is up to 0.33, which is low compared to both of the conventional pairs.

Index Termsalternative working fluid pair, computer aided mixture design, simulation, absorption refrigerator.

1 INTRODUCTION

—————————— ——————————

bsorption refrigerators (ARs) can be driven by low-quality heat sources such as process waste heat and solar energy. Replacing the commercially dominant vapour compression refrigerators with ARs can reduce electricity consumption. Fur- thermore, ARs are environmentally friendly due to their use of green natural materials such as water and ammonia as the refrig-
erant [1].
An AR has a working fluid that is a pair/solution comprising
TABLE 1

PROPERTIES OF THE MOST-USED AR WORKING FLUIDS (NH3–H2O AND H2O–LIBR) [2]
Property Ammonia–Water Water–Lithium
Bromide
Refrigerant
High latent heat Moderate Superior
a refrigerant and an absorbent. Back in the 1970s, water and sul-
phuric acid were used in the early AR systems. An AR machine dating back to 1859 used NH3–H2O as the working pair; after about a century another working fluid, the H2O–LiBr pair, was
Moderate vapour pressure
Low freezing tem- perature
Extremely high Markedly low
Excellent Limited application
introduced for the first time [2],[3]. Owing to the superior
tradeoffs of these two pairs, the other developed alternative work-
Low viscosity Good
Absorbent
ing fluids have almost no recognised commercial foothold. These
tradeoffs are listed in Table 1. Newly developed working fluids
Low vapour pres- sure
Poor Superb
for ARs succeeded in providing some of these properties, but
failed significantly to meet other requirements. Many substances can act as working fluids for ARs [4],[5],[6]. Thus, this paper tackles the search for a new working fluid for absorption refriger-
Low viscosity Acceptable
Mixture
ation with properties analogous to those of the NH3–H2O and
H2O–LiBr pairs. Computer-aided molecular design (CAMD) is
Solid phase for-
mation
No Yes
used successfully to find new compounds such as solvents for
many chemical processes [7], and absorbents for the refrigerant
Low toxicity Poor Fine
Affinity between
(water) in certain absorption refrigeration configurations [8].
Since the ProCAMD software package can be used to design molecules and mixtures, this software is used to find alternative working fluids with thermophysical and thermodynamic proper-
refrigerant and ab- sorbent
Good
ties that correspond to or surpass the most commercially success- ful working fluid tradeoffs.

————————————————

* Eman Abdel-Hakim Tora currently works as a researcher at the Depart- ment of Chemical Engineering & Pilot Plant in National Research Centre (NRC), Egypt. E-mail: emantora@gmail.com; eman.tora@yahoo.com

This paper considers a three-step approach through the use of
CAMD and process simulation software to find an alternative working fluid pair.

2 PROPERTIES OF THE SUBSTITUTE PAIR

NH3–H2O is a successful working pair that is used in indus- trial applications to provide cooling over a wide temperature

IJSER © 2013

http://www.ijser.org

International Journal of Scientific & Engineering Research, Volume 4, Issue 4, April-2013 1307

ISSN 2229-5518

range since the normal boiling point is −33.34°C and the freezing point is approximately −74°C. Similarly, H2O–LiBr can work using a low-quality heat source at around 70°C as an average. Finding a substitute refrigerant that replaces NH3 and still works well with water as the absorbent can be described as follows.
Regarding the pure refrigerant, the following properties are the design target:
- Low boiling point at atmospheric pressure
- High latent heat of vaporisation
- Low heat of mixing with the absorbent
- Low viscosity
- Not toxic
- The normal boiling point and the vapour pressure of the al- ternative refrigerant should not be close to those of the ab- sorbent (water) to allow easy separation and regeneration by heating. Moreover, ideally, one should have a mixture with separation efficiency that does not require a rectification col- umn, which in turn can reduce the capital cost and the oper- ating cost of the AR – thus enabling its commercial usage.
Regarding the mixture, the properties to be considered are as follows:
- The boiling point of the solution is in a range that can be provided by low-quality heat sources – for this study, it is taken as a temperature less than 150°C
- High miscibility between the refrigerant and the absorbent
- No azeotrope formation
Furthermore, there are a set of properties that both the pure refrigerant and the mixture should have:
- Completely miscible (it is a necessity to have a completely miscible solution to have a homogeneous solution in the va- pour and liquid phases under all conditions)
- Low heat of mixing (the absorption process is an exothermic
one with negative heat of mixing, thus the mixture should be a real one with negative deviation from Raoult’s law)
- No solid formation under operation conditions
- Not toxic or at least safe in the allowed range
After the desired properties and the constraints have been determined and the problem formulated, ProCAMD is run to generate the designed compounds. A list of the results is given in Table 3.
An Aspen plus simulation has been conducted to check the possibility of separation and the need for a rectifier, which is ad- dressed in detail in the next section.
Therefore, CAMD is used to generate possible mole- cules/mixtures; process simulation software (Aspen plus) is em- ployed to validate the cycle performance when the suggested pair is used. New substitutes are determined, and their comparative potential is assessed in terms of coefficient of performance (COP), high and low cycle pressure and heat source required.

2.1 Alternative Pair Design by ProCAMD

CAMD starts with a certain given user-preselected set of func- tional groups, then combines them, estimates their properties, filters and eventually selects the candidates with the desired properties. A working fluid to be designed consists of two com-
pounds, one acting as a refrigerant and the other as an absorbent, with the provision that the pair is completely soluble. Here, the design of the working fluid is achieved through three steps: work- ing fluid pair design is conducted using ICAS-ProCAMD through the design of a refrigerant based on physical properties, the ab- sorbent is designed according to solubility, and then the thermo- dynamics of the solution is checked.

2.1.1 Design of refrigerant based on physical properties

Preselecting functional groups allows one to consider alco-
hol, ketone, aldehyde, ester, ether, amine, amide, acid, and dou- ble- and triple-bond acyclic compounds. In contrast, chlorine-, bromine-, iodine-, sulphur-, phenol- and silicon-based com- pounds are dismissed.
The physical properties affecting the process are predeter- mined. These are the boiling point, vapour pressure, solubility, latent heat of vaporisation and freezing/melting point.
Constraints are then set for these properties. These con- straints result in compounds with the desired characteristics. The values of ammonia–water properties are taken as the average target values. Table 2 lists these vital properties for this case and their constraints, which are upper and lower acceptable values. An information screen from a ProCAMD run is shown in Fig. 1.
TABLE 2
CONSIDERED PHYSICAL PROPERTIES AND CON- STRAINTS OF THE WORKING FLUIDS

Physical properties

Unit

Constraints

Normal boiling

point (NBP)

K

0 ≤ NBP ≤ 275

Freezing point

(FP)

K

0 ≤ FB ≤ 200

Latent heat of va- porisation ()

kJ mol−1

5 ≤ ≤ 130

(NH3) = 23.27

(H2O) = 40.65

Molecular weight

(M)

−1

g mol

M ≤ 265

(M = 236.8 for R-

110 hexachloro- ethane )

Vapour pressure

(VP)

atm

1  VP  11

Minimum pressure is 1 atm to avoid

vacuum; maximum allowed pressure is

11 atm to be in an acceptable range, as that is the NH3 av- erage high pressure

Solubility (S)

S: gm R per

100 gm A

R:refrigerant,

‘solute’

A: absorbent,

‘solvent’

Completely soluble

By limiting compound design to a molecular weight of less than 265 g mol−1, normal boiling point of around 0°C and

IJSER © 2013

http://www.ijser.org

International Journal of Scientific & Engineering Research, Volume 4, Issue 4, April-2013 1308

ISSN 2229-5518


maximum freezing point of −73°C, ProCAMD considered 10 310 compounds, yet selected only seven compounds matching the desired properties. These are listed in Table 3 in ascending order based on normal boiling point.
Fig. 1. ProCAMD run information and results.
TABLE 3
DESIGNED REFRIGERANTS AS SUBSTITUTES FOR NH3
IN NH3–H2O WORKING FLUID

Name

Formula

kJ mol−1

NMP

K

NBP

K

NBP

°C

MWt

g mol−1

Propane

2 CH3 1 CH2

19.71

63.15

203.08

−69.92

44.09

Dimethyl ether

1 CH3 1 CH3O

21.86

125.01

234.03

−38.97

46.07

Isobutane

3 CH3 1 CH

21.95

57.18

242.22

−30.78

58.12

Not in the data-

base

2 CH3 1 CH2O

22.54

112.44

250.31

−22.69

60.09

n-Butane

2 CH3 2 CH2

24.36

104.62

263.12

−9.88

58.12

Acetaldehyde |

Methacrolein

1 CH3 1 CHO

23.85

159.74

269.68

−3.32

44.05

Not in the data-

base

3 CH3 1 CH–O

24.88

173.24

274.1

1.1

74.12

The ProCAMD results show that n-butane and acetaldehyde have latent heat of vaporisation almost equal to that of ammonia; however, their normal boiling points are higher. Therefore, they are not superior to ammonia as refrigerants, yet are very similar. Dimethyl ether has almost the same latent heat of vaporisation and normal boiling point as ammonia.
2.1.2 Design of absorbent according to solubility
Now one turns to the selection of an absorbent that is able to ab- sorb the pre-specified refrigerants. After determining potential substitute refrigerants, absorbents with high affinity for these refrigerants are to be determined. To do so, the ICAS Data Base Manager, particularly ‘Solvent Properties Page’, is used to speci- fy solvent ‘absorbents’. The results are given in Table 4.

3. PERFORMANCE OF ARS APPLYING THE ALTERNATIVE PAIR ( N-BUTANE – ETHANOL)

- Dühring curve
- Aspen plus simulation
Now the whole mixture is investigated in terms of the highest and lowest pressures that can be employed corresponding to feasible condenser cooling temperatures. Table 4 illustrates that ethyl ether–propane, ethyl ether–n-butane and ethanol–n- butane are very soluble, which means that they are highly likely
potential candidates. Therefore, their thermodynamic behaviour is determined to check if they are suitable for AR cycle perfor- mance. It is a good result to avoid using water as the absorbent as in the conventional pair, thus avoiding water evaporator malfunc- tion. The results given will be limited to the pair that gives the most satisfactory performance. In this study, ethanol–n-butane is investigated in details as follow.
TABLE 4
PRELIMINARY PRESELECTED PAIRS BASED ON THER- MOPHYSICAL properties

R e f r i g e r a n t

Propane

Dimethylether

Isobtane

n-Butane

Acetaldhyde

Absorbent

Miscibility

Water

S

S

SS

S

S

Ethanol

S

S

S

VS

S

Ethyl ether

VS

S

S

VS

S

Acetone

SS

S

S: soluble; VS: very solu-

ble; SS: slightly soluble.

3.1 Dühring Curve

For n-butane as refrigerant ethanol as the absorbent, Fig. 2
shows Dühring curves, and, based on them, the appropriate oper- ating conditions to provide cooling at 0°C are as follows: conden- ser at 42.6°C, desorber at 90°C and absorber temperature of

25°C.
Fig. 2. Dühring curves of n-butane–ethanol pair.
This pair provides cooling at a level comparable to that of water–lithium bromide, yet it surpasses the latter based on the following advantages:
1. No vacuum is needed as in the case of water–lithium bro- mide
2. The low pressure is 1 atm and the high pressure is 4 atm
(moderate pressure)
3. Condenser cooling can be carried out using ambient air, as with water
4. No crystallisation risk

3.2 Simulation Using Aspen plus

Based on the Dühring curve results, simulation of an AR cycle has been performed under various conditions. Figure 3 shows snapshots of the simulation under three different operating condi- tions.

IJSER © 2013

http://www.ijser.org

International Journal of Scientific & Engineering Research, Volume 4, Issue 4, April-2013 1309

ISSN 2229-5518


The operating conditions of the cycles when using the pro- posed pair are moderate and acceptable in terms of the re- quired heat demand temperature and pressures. However, for the proposed pair to give a cooling range analogous to that of ammonia–water, the COP of the system is low, as listed in Table 5.
Regarding the cooling range of H2O–LiBr, the proposed pair is superior to H2O–LiBr from the viewpoint of no crystallisa- tion risk, cooling at lower temperature level, no vacuum and both the low and high pressures being acceptable. However, H2O–LiBr still has a higher COP than the proposed pair.
TABLE 5
OPERATING CONDITIONS AND PERFORMANCE OF N-BUTANE–ETHANOL PAIR COMPARED TO LITHIUM BROMIDE–WATER PAIR

Tevp (°C)

Tabs

°C

PL

atm

Tg

°C

PH

atm

Tcond

°C

COP

0

16

1

91.5

4

42.6

0.22

10

31

1.5

91.5

4

42.6

0.25

20

43

2

91.5

4

42.6

0.3

5. CONCLUDING REMARKS

The following conclusions can be drawn based on the
Dühring curve and simulation results:
- The proposed working pair (n butane–ethanol) can act as a working pair for absorption refrigerator
- When employing the proposed pair, an absorption refrigera-
tion system has the requirements of mild pressure and a low- temperature heat source
- No crystallisation risk
- The proposed pair can provide cooling at temperatures lower than that be provided by H2O–LiBr, but with a lower COP than if using NH3–H2O
- Employing the new pair involves tradeoffs between the pref-
erable operating conditions and COP value.

ACKNOWLEDGMENT

The author would like to express thanks to CAPEC and Prof. R. Gani, Department of Chemical Engineering, Technical University of Denmark, for providing the ICAS software.
Fig. 3. Aspen plus simulation of n-butane–ethanol pair at differ- ent cooling temperatures.

4. RESULTS

The results as given in Table 5 are promising and confirm that the proposed n-butane–ethanol pair is comparable to water–lithium bromide in terms of the heat source temperature (91.5°C), evapo- rator temperature (0–20°C) and absorber temperature (16–43°C). However, the COP is 0.22–0.3 when 0.2 wt. fraction n-butane is the strong solution leaving the absorbent.

References

[1] P. Srikhirin, S. Aphornratana and S. Hungpaibulpatana:
Renew. Sustain. Energy Rev., 2001, 5, 343– 372.
[2] K. Herold, R. Radermacher and S. A. Klein: ‘Absorption chillers and heat pumps’; 1996, Boca Raton, Florida, CRC Press.
[3] L. Radermacher: Mech. Eng., August, 1989, 68–73.

IJSER © 2013

http://www.ijser.org

International Journal of Scientific & Engineering Research, Volume 4, Issue 4, April-2013 1310

ISSN 2229-5518

[4] W. B. Gosney: ‘Principle of refrigeration’; 1982, Cam- bridge, Cambridge University Press.
[5] J. Sun, L. Fu and S. Zhang: Renew. Sustain. Energy Rev.,
2012, 16, 1899–1906.
[6] L. W. Wang, R. Z. Wang and R. G. Oliveira: ‘A review on adsorption working pairs for refrigeration’, Report, Energy Division, National Laboratory, Oak Ridge, Ten- nessee, 2009.
[7] R. A. Macriss, J. M Gutraj and T. S. Zawacki: ‘Absorp- tion fluids data survey: final report of worldwide data’, Report ORNL/Sub/84-47989/3, Institute of Gas Tech- nology, Chicago, Illinois, 1988.
[8] ICAS documentation: CAPEC, Department of Chemical
Engineering, Technical University of Denmark.

IJSER © 2013

http://www.ijser.org