- Email: [email protected]
Evaluation of the performance of solid sorption refrigeration systems using carbon dioxide as refrigerant
Applied Thermal Engineering 26 (2006) 1807–1811 www.elsevier.com/locate/apthermeng
Evaluation of the performance of solid sorption refrigeration systems using carbon dioxide as refrigerant Y. Zhong *, R.E. Critoph, R. Thorpe Department of Engineering, University of Warwick, Coventry CV4 7AL, United Kingdom Received 10 January 2005; accepted 2 February 2006 Available online 3 April 2006
Abstract The refrigerant–adsorbent pairs at present mostly used in research on solid-sorption refrigeration cycles are ammonia–carbon, methanol–carbon, water–silica gel and water–zeolite. Porosity tests have been carried out on an alternative, carbon dioxide, as refrigerant with several types of carbon, zeolite and silica gel as adsorbent. The results of fitting to the Dubinin equation and modelling of cycles based on these pairs are presented. Due to its low latent heat, the performance of adsorption refrigeration systems using CO2 as refrigerant is poor. However, carbon dioxide may be useful where the toxicity and incompatibility of ammonia and copper is valued. Ó 2006 Elsevier Ltd. All rights reserved. Keywords: Adsorption; Refrigeration; Carbon dioxide
1. Introduction There has been wide range of possible adsorbent/refrigerant pairs for solid-sorption cooling and heating. Historically, Michael Faraday first demonstrated an adsorption refrigeration system in 1848 utilising ammonia and silver chloride as working pair [1]. In 1929, Hulse and Miller described an adsorption system using in the United States for the air conditioning system of railway carriages [2], in which Silica gel–sulphur dioxide was chosen as the adsorbent/adsorbate pair. In recent times, the adsorption pairs mostly used include ammonia–carbon [3,4], ammonia-salts [5,6], methanol–carbon [7,8], water–silica gel [9,10] and water–zeolite [11,12]. The sub-atmospheric refrigerants (water and methanol) have excellent thermodynamic properties but their specific cooling power (kW of cooling per kg of adsorbent) can be limited by low mass transfer rates resulting from pressure drops, internal diffusion etc. Ammonia, being at much higher pressures (typically 5– 20 bar) does not suffer these mass transfer limitations but does have the disadvantages of its toxicity and incompati*
Corresponding author. Tel.: +44 2476530273. E-mail address: [email protected] (Y. Zhong).
1359-4311/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.applthermaleng.2006.02.006
bility with copper. Its thermodynamic properties, whilst not as good as water, are sufficient to obtain useful COP’s, particularly when used in regenerative cycles. With the advent of such highly regenerative cycles, such as multiple-beds [13], thermal waves [14] and the convective thermal wave [15], it may be worth considering high pressure refrigerants other than ammonia which have advantages such as compatibility with copper, or more convenient working pressure and so on. The new refrigerant tested is carbon dioxide. As we know, CO2 is one of the few natural refrigerants, which is neither flammable nor toxic. It is inexpensive, widely available and does not affect the global environment as do many other refrigerants. CO2 has a GWP = 1 (global warming potential, the GWP of HFC is 1000–3000), but the net global warming impact when used as a technical gas is zero, since the gas is a waste product from industrial production. Therefore, CO2 is an excellent alternative among the natural refrigerants, especially in applications where the toxicity and flammability of ammonia and hydrocarbons may be a problem. The promising applications of CO2 refrigeration systems include automotive air conditioning, heat pumps, residential/commercial air conditioning and various refrigeration areas [16,17]. However,
1808
Y. Zhong et al. / Applied Thermal Engineering 26 (2006) 1807–1811
there has been very little work on adsorption refrigeration systems using CO2 as the refrigerant. This paper presents an experimental study on the porosity characteristics of CO2 with various types of activated carbon, zeolite and silica gel. The performance of adsorption systems using CO2 is also predicted using a numerical model.
ance, is shown schematically in Fig. 1. The adsorbent sample is contained in a small basket, which is suspended inside the sealed steel chamber. The temperature of the chamber is controlled by a heater (Polystat CC302). Two thermocouples within the chamber are used to measure the gas and the wall temperature to ensure an equilibrium state. When a new temperature is selected, readings are only taken when the temperature varieties measured by the two thermocouples are within 0.5 °C for a period of at least 30 min and the weight measurement is similarly stable. The system pressure was measured directly with a calibrated Druck PDCR 920 transducer. Adsorbate concentration was calculated from the weight of the sample, corrected for gas buoyancy. The CO2 was introduced as a solid which was allowed to evaporate, this being more convenient and controllable than introducing it as a gas. In this series of tests, the experimental sample (contained in a stainless steel basket) was exposed to pure carbon dioxide. The temperature within the steel chamber ranged from 30 °C to 150 °C and the pressure from 10 bar to 50 bar.
2. Porosity test rig
3. Adsorbents selected
Two different sets of experimental apparatus have been used to measure the data. Both have adequate levels of accuracy (concentration ±0.001 kg adsorbate/kg adsorbent), but the newer apparatus is more highly automated. The former method used a sealed sample vessel with temperature and pressure instrumentation. CO2 was introduced from a gas bottle and the vessel and its instrumentation weighed to determine the exact mass of CO2. The vessel was then heated electrically, equilibrium pressure and temperature being measured over the required range. The concentration at those temperature and pressure combinations was then calculated from knowledge of the gas properties and the mass of CO2 in the system. After another weighing to check that no CO2 had escaped, the process was repeated with a range of different masses of CO2 within the vessel. This (older) test equipment was used for all samples except the 208C carbon. The newer method, which utilises a Rubotherm high temperature, medium pressure magnetic suspension bal-
Three different types of adsorbent were tested. The first is activated carbon, which included Sutcliffe Speakman [now Chemviron] 208C, NORIT A5798, NORIT 2030 and NORIT RB1. 208C is coconut-shell-based gas adsorption carbon which is low-cost and generally used in our laboratories. The other two kinds of adsorbent are silica gel, which included MD263 and HPV silica, and zeolite, which included zeolite 3A, 4A, 13X and NA-Y. All these adsorbents selected in our porosity experiments are widely used in recent research on adsorption refrigeration systems [18].
Solid CO 2
Pressure transducer
Thermocouple
Vacuum pump
Sample
Steel chamber
Thermocouple
Oil bath
Fig. 1. Schematic flow diagram of porosity test rig.
4. Test results The following pairs were tested for their porosity: the refrigerant is carbon dioxide and the adsorbents are those mentioned in part three. The concentration of an adsorbent sample was measured between room temperature and about 200 °C and corresponding pressure from 5 bar to 60 bar.
Table 1 Porosity test results Adsorbent
Refrigerant
Limiting concentration (x0)
K
n
SEE
208C carbon NORIT A5798 carbon NORIT 2030 carbon NORIT RB1 carbon MD263 silica HPV silica Zeolite 13X Zeolite 3A Zeolite 4A Zeolite NA-Y
Carbon Carbon Carbon Carbon Carbon Carbon Carbon Carbon Carbon Carbon
0.3242 0.5049 0.5929 0.5184 0.6591 0.6257 0.2741 0.2139 0.2754 0.5553
2.5135 3.5318 2.2717 2.9064 4.281 3.5887 1.9792 0.8503 0.6865 0.7826
1.1602 1.7473 0.7331 1.3221 0.7887 0.3752 4.3149 1.8918 0.4754 0.2054
0.0087 0.0037 0.0109 0.0025 0.0016 0.0025 0.0074 0.0048 0.0043 0.009
dioxide dioxide dioxide dioxide dioxide dioxide dioxide dioxide dioxide dioxide
Y. Zhong et al. / Applied Thermal Engineering 26 (2006) 1807–1811
1809
All the experimental data were fitted to a modified Dubinin–Astakhov equation: n T 1 x ¼ x0 exp K ð1Þ T sat where x0, K and n are constants for the pair; Tsat, saturation temperature (K); T, adsorbent temperature (K). A program written in MATLAB is used to find the values of x0, K and n, which minimize the sum of the squares of the differences in concentration predicted by the equation and the measured. The results and the standard estimate of error (see) are given in Table 1. The data given in Table 1 implies that the carbon dioxide is adsorbed well by most of the adsorbents. Figs. 2–4 show the isosteres on a Clapeyron diagram. The points labelled with a ‘+’ in these diagrams are the experimental data. Fig. 4. Clapeyron diagram for zeolite 3A and carbon dioxide.
To determine whether the performance of adsorption systems using CO2 as refrigerant is satisfactory, complete cycle analyses were carried out. 5. Cycle modelling method The method used to calculate the heating required and cooling produced follows Meunier [19] and Critoph and Turner [20]. The concentration is obtained by the use of Eq. (1) and the effective specific heat (dQ/dT) along an isostere is given by: dQ ¼ cpc ðT Þ þ xcpa dT
Fig. 2. Clapeyron diagram for 208C carbon and carbon dioxide.
ð2Þ
cpc is the adsorbent specific heat. In the absence of complete information, the value obtained by Turner [21] for 208C as a function of temperature has been used for all four carbons: cpc ¼ 174 þ 2:245T ðJ=kgÞ
ð3Þ
The value 921 J/kg [22] has been used for all types of silica and the value 836 J/kg [23] has been used for all types of zeolite. Here, cpa is the adsorbate specific heat. It is commonly assumed similar to the specific heat of the liquid phase at the same temperature [24], but this assumption breaks down at higher than critical temperatures. It could also be argued that cv is more appropriate than cp along an isostere [25]. In this work a constant value of 2986 J/kg is taken for carbon dioxide. During desorption the effective specific heat is given by dQ ox ¼ cpc ðT Þ þ xcpa H ð4Þ dT oT p where H is the heat of desorption, given by H ¼ Rðp; T ÞA Fig. 3. Clapeyron diagram for MD263 silica gel and carbon dioxide.
T T sat
ð5Þ
1810
Y. Zhong et al. / Applied Thermal Engineering 26 (2006) 1807–1811
R is the gas constant and A is the slope of the saturated adsorbate line on the Clapeyron diagram; Tsat is the saturation temperature corresponding to the system pressure. All measurements were made below the critical pressure and so the slope and Tsat have real physical significance. The effective specific heat is calculated for every 1 K change in temperature during both the heating and cooling phases. The total heating or cooling load of a bed is simply the integral of the above equations, actually taken as the sum of the value of effective specific heat calculated every degree. The cooling and the heat rejected in the condenser are evaluated by considering the mass of refrigerant desorbed and then adsorbed per unit mass of adsorbent during every cycle. The net concentration change is xconc xdil. The useful cooling obtained from it is: Qc ¼ ðxconc xdil Þðhgasev hliquidcon Þ
ð6Þ
where: xconc, maximum concentration; xdil, minimum concentration; hgasev, specific enthalpy of gas leaving the evaporator (kJ/kg); hliquidcon, specific enthalpy of the condensed liquid (kJ/kg). The condensing temperature used was 30 °C, a little below the critical temperature. The cycle COP is calculated for both a basic (one-bed) cycle and two-bed cycle in which some of the heat of adsorption can be pre-heat the desorbing-bed. The assumption is made that the two-beds can transfer heat until they are within 10 K of each other. This process is illustrated in Figs. 5–7.
Fig. 6. Cycle heat load for MD263 and carbon dioxide with 5 °C evaporating and 200 °C generating temperatures.
6. Cycle COPs In Figs. 5–7, the area under the upper curves gives the heat input between any two temperatures and the area under the lower (negative) curves gives the heat rejected. The discontinuities are the changes from isosteric heating to isobaric or vice versa. The quantity of heat that can be Fig. 7. Cycle heat load for zeolite 3 A and carbon dioxide with 5 °C evaporating and 200 °C generating temperatures.
transferred from one-bed to another, assuming an approach temperature of 10 K is also be shown. These figures illustrate that the cycle COPs of these adsorption systems are very low, since the large concentration swings are negated by the lower latent heat of CO2. To illustrate and confirm this supposition, the latent heat of CO2 was artificially set to be the same as ammonia (which is about 6 times that of CO2). The result is shown in Fig. 8, which indicate a much higher COP although of course this result is not for a real refrigerant. 7. Conclusion
Fig. 5. Cycle heat load for 208C and carbon dioxide with 5 °C evaporating and 200 °C generating temperatures.
An alternative refrigerant (CO2) and several adsorbents have been studied experimentally in this paper. Measurements of the porosity characteristics show that carbon
Y. Zhong et al. / Applied Thermal Engineering 26 (2006) 1807–1811
Fig. 8. Cycle heat load for 208C and carbon dioxide with 5 °C evaporating and 200 °C generating temperatures (the latent heat of CO2 has been artificially changed).
dioxide is adsorbed well by a wide range of adsorbents. The performance of adsorption systems using CO2 as the refrigerant has been simulated by cycle modelling. The results show that the performance of single and two-bed cycles are poor due to low latent heat. If there are applications where COP is not of primary importance, but toxicity and the ability to use copper is valued, then CO2 might still be useful. Acknowledgement Chemviron for samples of carbon. References [1] R. Thevenot, A history of refrigeration throughout the world, IIR, Paris, 1979. [2] G.E. Hulse, Freight Car Refrigeration by an adsorption system employing silica gel, Refrigerating Engineer 17 (2) (1924). [3] R.E. Critoph, An ammonia carbon solar refrigerator for vaccine cooling, Renewable Energy 5 (1) (1994) 502–508. [4] Z. Tamainot-Telto, R.E. Critoph, Adsorption refrigeration using monolithic carbon–ammonia pair, International Journal of Refrigeration 20 (2) (1997) 146–155. [5] B. Spinner, Les transformateurs thermochimiques a ammoniac, Proc. Symp. le Froid a Sorption Solide, Paris, (1992). [6] U. Rockenfeller, L.D. Kirol, P. Sarkisian et al., Advanced heat pump staging for complex compound chemisorption systems, Proc. Symp. le Froid a Sorption Solide, Paris, (1992).
1811
[7] Z.F. Li, K. Sumathy, A solar powered ice maker with the solid adsorption pair of activated carbon and methanol, International Journal of Energy Research 23 (1999) 517–527. [8] R. Exell, S.C. Bhattacharya, Y.R. Upadhyaya, Research and development of solar powered desiccant refrigeration for cold storage applications, Asian Institute of Technology, Bangkok, AIT Research Report, 1993. [9] B.B. Saha, S. Koyama, J.B. Lee, et al., Performance evaluation of a low-temperature waste heat driven multi-bed adsorption chiller, International Journal of Multiphase Flow 29 (8) (2003) 1249–1263. [10] A. Akahira, K.C.A. Alam, Y. Hamamoto, et al., Mass recovery adsorption refrigeration cycle-improving cooling capacity, International Journal of Refrigeration 27 (3) (2004) 225–234. [11] P.H. Grenier, J.J. Guilleminot, F. Meunier, et al., Solar powered solid adsorption cold store, ASME Journal of Solar Energy Engineering 110 (1998) 192–197. [12] M. Tather, A. Erdem-Senatalar, The effects of thermal gradients in a solar adsorption heat pump utilizing the zeolite–water pair, Applied Thermal Engineering 19 (1999) 1157–1172. [13] F. Meunier, Second law analysis of solid adsorption heat pump operating on reversible cascade cycle application to the zeolite–water pair, Heat Recovery System 15 (2) (1985) 133–141. [14] S.V. Shelton, W.J. Wepfer, D.J. Miles, Ramp wave analysis of the solid/vapour heat pump, Journal of Energy Research and Technology (112) (1990) 69–77. [15] R.E. Critoph, Forced convection adsorption cycles, Applied Thermal Engineering (18) (1988) 799–807. [16] P. Neksa, CO2 heat pump systems, International Journal of Refrigeration 25 (2002) 421–427. [17] M.H. Kim, J. Pettersen, C.W. Bullard, Fundamental process and system design issues in CO2 vapor compression systems, Progress in Energy and Combustion Science 30 (2004) 119–174. [18] N.C. Srivastava, I.W. Eames, A review of adsorbents and adsorbates in solid–vapour adsorption heat pump systems, Applied Thermal Engineering 18 (1998) 707–714. [19] F. Meunier, Second law analysis of solid adsorption heat pump operating on reversible cascade cycle: application to the zeolite–water pair, Heat Recovery System 5 (2) (1985) 133–141. [20] R.E. Critoph, H.L. Turner, Performance of ammonia-activated carbon and ammonia zeolite heat pump adsorption cycle, Proc. Pompes a Chaleur Chimiques De Hautes Performances Perpignan, Lavoisier, Paris (September) 1989. [21] L. Turner, Improvement of activated charcoal–ammonia adsorption heat pumping/refrigeration cycles. Investigation of porosity and heat / mass transfer characteristics, Ph.D. Thesis, University of Warwick, UK, (1992). [22] K.C. Ng, H.T. Chua, C.Y. Chung, et al., Experimental investigation of the silica gel–water adsorption isotherm characteristics, Applied Thermal Engineering 21 (16) (2001) 1631–1642. [23] N.B. Amar, L.M. Sun, F. Meunier, Numerical analysis of adsorption temperature wave regenerative heat pump, Applied Thermal Engineering 16 (5) (1996) 405–418. [24] R.E. Critoph, Adsorption refrigeration and heat pumps, in: Trimothy D. Burchell (Ed.), Carbon Materials for Advanced Technologies, Pergamon Press, 1999. [25] R.E. Critoph, Evaluation of alternative refrigerant–adsorbent pairs for refrigeration cycles, Applied Thermal Engineering 16 (11) (1996) 891–900.
Evaluation of the performance of solid sorption refrigeration systems using carbon dioxide as refrigerant Y. Zhong *, R.E. Critoph, R. Thorpe Department of Engineering, University of Warwick, Coventry CV4 7AL, United Kingdom Received 10 January 2005; accepted 2 February 2006 Available online 3 April 2006
Abstract The refrigerant–adsorbent pairs at present mostly used in research on solid-sorption refrigeration cycles are ammonia–carbon, methanol–carbon, water–silica gel and water–zeolite. Porosity tests have been carried out on an alternative, carbon dioxide, as refrigerant with several types of carbon, zeolite and silica gel as adsorbent. The results of fitting to the Dubinin equation and modelling of cycles based on these pairs are presented. Due to its low latent heat, the performance of adsorption refrigeration systems using CO2 as refrigerant is poor. However, carbon dioxide may be useful where the toxicity and incompatibility of ammonia and copper is valued. Ó 2006 Elsevier Ltd. All rights reserved. Keywords: Adsorption; Refrigeration; Carbon dioxide
1. Introduction There has been wide range of possible adsorbent/refrigerant pairs for solid-sorption cooling and heating. Historically, Michael Faraday first demonstrated an adsorption refrigeration system in 1848 utilising ammonia and silver chloride as working pair [1]. In 1929, Hulse and Miller described an adsorption system using in the United States for the air conditioning system of railway carriages [2], in which Silica gel–sulphur dioxide was chosen as the adsorbent/adsorbate pair. In recent times, the adsorption pairs mostly used include ammonia–carbon [3,4], ammonia-salts [5,6], methanol–carbon [7,8], water–silica gel [9,10] and water–zeolite [11,12]. The sub-atmospheric refrigerants (water and methanol) have excellent thermodynamic properties but their specific cooling power (kW of cooling per kg of adsorbent) can be limited by low mass transfer rates resulting from pressure drops, internal diffusion etc. Ammonia, being at much higher pressures (typically 5– 20 bar) does not suffer these mass transfer limitations but does have the disadvantages of its toxicity and incompati*
Corresponding author. Tel.: +44 2476530273. E-mail address: [email protected] (Y. Zhong).
1359-4311/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.applthermaleng.2006.02.006
bility with copper. Its thermodynamic properties, whilst not as good as water, are sufficient to obtain useful COP’s, particularly when used in regenerative cycles. With the advent of such highly regenerative cycles, such as multiple-beds [13], thermal waves [14] and the convective thermal wave [15], it may be worth considering high pressure refrigerants other than ammonia which have advantages such as compatibility with copper, or more convenient working pressure and so on. The new refrigerant tested is carbon dioxide. As we know, CO2 is one of the few natural refrigerants, which is neither flammable nor toxic. It is inexpensive, widely available and does not affect the global environment as do many other refrigerants. CO2 has a GWP = 1 (global warming potential, the GWP of HFC is 1000–3000), but the net global warming impact when used as a technical gas is zero, since the gas is a waste product from industrial production. Therefore, CO2 is an excellent alternative among the natural refrigerants, especially in applications where the toxicity and flammability of ammonia and hydrocarbons may be a problem. The promising applications of CO2 refrigeration systems include automotive air conditioning, heat pumps, residential/commercial air conditioning and various refrigeration areas [16,17]. However,
1808
Y. Zhong et al. / Applied Thermal Engineering 26 (2006) 1807–1811
there has been very little work on adsorption refrigeration systems using CO2 as the refrigerant. This paper presents an experimental study on the porosity characteristics of CO2 with various types of activated carbon, zeolite and silica gel. The performance of adsorption systems using CO2 is also predicted using a numerical model.
ance, is shown schematically in Fig. 1. The adsorbent sample is contained in a small basket, which is suspended inside the sealed steel chamber. The temperature of the chamber is controlled by a heater (Polystat CC302). Two thermocouples within the chamber are used to measure the gas and the wall temperature to ensure an equilibrium state. When a new temperature is selected, readings are only taken when the temperature varieties measured by the two thermocouples are within 0.5 °C for a period of at least 30 min and the weight measurement is similarly stable. The system pressure was measured directly with a calibrated Druck PDCR 920 transducer. Adsorbate concentration was calculated from the weight of the sample, corrected for gas buoyancy. The CO2 was introduced as a solid which was allowed to evaporate, this being more convenient and controllable than introducing it as a gas. In this series of tests, the experimental sample (contained in a stainless steel basket) was exposed to pure carbon dioxide. The temperature within the steel chamber ranged from 30 °C to 150 °C and the pressure from 10 bar to 50 bar.
2. Porosity test rig
3. Adsorbents selected
Two different sets of experimental apparatus have been used to measure the data. Both have adequate levels of accuracy (concentration ±0.001 kg adsorbate/kg adsorbent), but the newer apparatus is more highly automated. The former method used a sealed sample vessel with temperature and pressure instrumentation. CO2 was introduced from a gas bottle and the vessel and its instrumentation weighed to determine the exact mass of CO2. The vessel was then heated electrically, equilibrium pressure and temperature being measured over the required range. The concentration at those temperature and pressure combinations was then calculated from knowledge of the gas properties and the mass of CO2 in the system. After another weighing to check that no CO2 had escaped, the process was repeated with a range of different masses of CO2 within the vessel. This (older) test equipment was used for all samples except the 208C carbon. The newer method, which utilises a Rubotherm high temperature, medium pressure magnetic suspension bal-
Three different types of adsorbent were tested. The first is activated carbon, which included Sutcliffe Speakman [now Chemviron] 208C, NORIT A5798, NORIT 2030 and NORIT RB1. 208C is coconut-shell-based gas adsorption carbon which is low-cost and generally used in our laboratories. The other two kinds of adsorbent are silica gel, which included MD263 and HPV silica, and zeolite, which included zeolite 3A, 4A, 13X and NA-Y. All these adsorbents selected in our porosity experiments are widely used in recent research on adsorption refrigeration systems [18].
Solid CO 2
Pressure transducer
Thermocouple
Vacuum pump
Sample
Steel chamber
Thermocouple
Oil bath
Fig. 1. Schematic flow diagram of porosity test rig.
4. Test results The following pairs were tested for their porosity: the refrigerant is carbon dioxide and the adsorbents are those mentioned in part three. The concentration of an adsorbent sample was measured between room temperature and about 200 °C and corresponding pressure from 5 bar to 60 bar.
Table 1 Porosity test results Adsorbent
Refrigerant
Limiting concentration (x0)
K
n
SEE
208C carbon NORIT A5798 carbon NORIT 2030 carbon NORIT RB1 carbon MD263 silica HPV silica Zeolite 13X Zeolite 3A Zeolite 4A Zeolite NA-Y
Carbon Carbon Carbon Carbon Carbon Carbon Carbon Carbon Carbon Carbon
0.3242 0.5049 0.5929 0.5184 0.6591 0.6257 0.2741 0.2139 0.2754 0.5553
2.5135 3.5318 2.2717 2.9064 4.281 3.5887 1.9792 0.8503 0.6865 0.7826
1.1602 1.7473 0.7331 1.3221 0.7887 0.3752 4.3149 1.8918 0.4754 0.2054
0.0087 0.0037 0.0109 0.0025 0.0016 0.0025 0.0074 0.0048 0.0043 0.009
dioxide dioxide dioxide dioxide dioxide dioxide dioxide dioxide dioxide dioxide
Y. Zhong et al. / Applied Thermal Engineering 26 (2006) 1807–1811
1809
All the experimental data were fitted to a modified Dubinin–Astakhov equation: n T 1 x ¼ x0 exp K ð1Þ T sat where x0, K and n are constants for the pair; Tsat, saturation temperature (K); T, adsorbent temperature (K). A program written in MATLAB is used to find the values of x0, K and n, which minimize the sum of the squares of the differences in concentration predicted by the equation and the measured. The results and the standard estimate of error (see) are given in Table 1. The data given in Table 1 implies that the carbon dioxide is adsorbed well by most of the adsorbents. Figs. 2–4 show the isosteres on a Clapeyron diagram. The points labelled with a ‘+’ in these diagrams are the experimental data. Fig. 4. Clapeyron diagram for zeolite 3A and carbon dioxide.
To determine whether the performance of adsorption systems using CO2 as refrigerant is satisfactory, complete cycle analyses were carried out. 5. Cycle modelling method The method used to calculate the heating required and cooling produced follows Meunier [19] and Critoph and Turner [20]. The concentration is obtained by the use of Eq. (1) and the effective specific heat (dQ/dT) along an isostere is given by: dQ ¼ cpc ðT Þ þ xcpa dT
Fig. 2. Clapeyron diagram for 208C carbon and carbon dioxide.
ð2Þ
cpc is the adsorbent specific heat. In the absence of complete information, the value obtained by Turner [21] for 208C as a function of temperature has been used for all four carbons: cpc ¼ 174 þ 2:245T ðJ=kgÞ
ð3Þ
The value 921 J/kg [22] has been used for all types of silica and the value 836 J/kg [23] has been used for all types of zeolite. Here, cpa is the adsorbate specific heat. It is commonly assumed similar to the specific heat of the liquid phase at the same temperature [24], but this assumption breaks down at higher than critical temperatures. It could also be argued that cv is more appropriate than cp along an isostere [25]. In this work a constant value of 2986 J/kg is taken for carbon dioxide. During desorption the effective specific heat is given by dQ ox ¼ cpc ðT Þ þ xcpa H ð4Þ dT oT p where H is the heat of desorption, given by H ¼ Rðp; T ÞA Fig. 3. Clapeyron diagram for MD263 silica gel and carbon dioxide.
T T sat
ð5Þ
1810
Y. Zhong et al. / Applied Thermal Engineering 26 (2006) 1807–1811
R is the gas constant and A is the slope of the saturated adsorbate line on the Clapeyron diagram; Tsat is the saturation temperature corresponding to the system pressure. All measurements were made below the critical pressure and so the slope and Tsat have real physical significance. The effective specific heat is calculated for every 1 K change in temperature during both the heating and cooling phases. The total heating or cooling load of a bed is simply the integral of the above equations, actually taken as the sum of the value of effective specific heat calculated every degree. The cooling and the heat rejected in the condenser are evaluated by considering the mass of refrigerant desorbed and then adsorbed per unit mass of adsorbent during every cycle. The net concentration change is xconc xdil. The useful cooling obtained from it is: Qc ¼ ðxconc xdil Þðhgasev hliquidcon Þ
ð6Þ
where: xconc, maximum concentration; xdil, minimum concentration; hgasev, specific enthalpy of gas leaving the evaporator (kJ/kg); hliquidcon, specific enthalpy of the condensed liquid (kJ/kg). The condensing temperature used was 30 °C, a little below the critical temperature. The cycle COP is calculated for both a basic (one-bed) cycle and two-bed cycle in which some of the heat of adsorption can be pre-heat the desorbing-bed. The assumption is made that the two-beds can transfer heat until they are within 10 K of each other. This process is illustrated in Figs. 5–7.
Fig. 6. Cycle heat load for MD263 and carbon dioxide with 5 °C evaporating and 200 °C generating temperatures.
6. Cycle COPs In Figs. 5–7, the area under the upper curves gives the heat input between any two temperatures and the area under the lower (negative) curves gives the heat rejected. The discontinuities are the changes from isosteric heating to isobaric or vice versa. The quantity of heat that can be Fig. 7. Cycle heat load for zeolite 3 A and carbon dioxide with 5 °C evaporating and 200 °C generating temperatures.
transferred from one-bed to another, assuming an approach temperature of 10 K is also be shown. These figures illustrate that the cycle COPs of these adsorption systems are very low, since the large concentration swings are negated by the lower latent heat of CO2. To illustrate and confirm this supposition, the latent heat of CO2 was artificially set to be the same as ammonia (which is about 6 times that of CO2). The result is shown in Fig. 8, which indicate a much higher COP although of course this result is not for a real refrigerant. 7. Conclusion
Fig. 5. Cycle heat load for 208C and carbon dioxide with 5 °C evaporating and 200 °C generating temperatures.
An alternative refrigerant (CO2) and several adsorbents have been studied experimentally in this paper. Measurements of the porosity characteristics show that carbon
Y. Zhong et al. / Applied Thermal Engineering 26 (2006) 1807–1811
Fig. 8. Cycle heat load for 208C and carbon dioxide with 5 °C evaporating and 200 °C generating temperatures (the latent heat of CO2 has been artificially changed).
dioxide is adsorbed well by a wide range of adsorbents. The performance of adsorption systems using CO2 as the refrigerant has been simulated by cycle modelling. The results show that the performance of single and two-bed cycles are poor due to low latent heat. If there are applications where COP is not of primary importance, but toxicity and the ability to use copper is valued, then CO2 might still be useful. Acknowledgement Chemviron for samples of carbon. References [1] R. Thevenot, A history of refrigeration throughout the world, IIR, Paris, 1979. [2] G.E. Hulse, Freight Car Refrigeration by an adsorption system employing silica gel, Refrigerating Engineer 17 (2) (1924). [3] R.E. Critoph, An ammonia carbon solar refrigerator for vaccine cooling, Renewable Energy 5 (1) (1994) 502–508. [4] Z. Tamainot-Telto, R.E. Critoph, Adsorption refrigeration using monolithic carbon–ammonia pair, International Journal of Refrigeration 20 (2) (1997) 146–155. [5] B. Spinner, Les transformateurs thermochimiques a ammoniac, Proc. Symp. le Froid a Sorption Solide, Paris, (1992). [6] U. Rockenfeller, L.D. Kirol, P. Sarkisian et al., Advanced heat pump staging for complex compound chemisorption systems, Proc. Symp. le Froid a Sorption Solide, Paris, (1992).
1811
[7] Z.F. Li, K. Sumathy, A solar powered ice maker with the solid adsorption pair of activated carbon and methanol, International Journal of Energy Research 23 (1999) 517–527. [8] R. Exell, S.C. Bhattacharya, Y.R. Upadhyaya, Research and development of solar powered desiccant refrigeration for cold storage applications, Asian Institute of Technology, Bangkok, AIT Research Report, 1993. [9] B.B. Saha, S. Koyama, J.B. Lee, et al., Performance evaluation of a low-temperature waste heat driven multi-bed adsorption chiller, International Journal of Multiphase Flow 29 (8) (2003) 1249–1263. [10] A. Akahira, K.C.A. Alam, Y. Hamamoto, et al., Mass recovery adsorption refrigeration cycle-improving cooling capacity, International Journal of Refrigeration 27 (3) (2004) 225–234. [11] P.H. Grenier, J.J. Guilleminot, F. Meunier, et al., Solar powered solid adsorption cold store, ASME Journal of Solar Energy Engineering 110 (1998) 192–197. [12] M. Tather, A. Erdem-Senatalar, The effects of thermal gradients in a solar adsorption heat pump utilizing the zeolite–water pair, Applied Thermal Engineering 19 (1999) 1157–1172. [13] F. Meunier, Second law analysis of solid adsorption heat pump operating on reversible cascade cycle application to the zeolite–water pair, Heat Recovery System 15 (2) (1985) 133–141. [14] S.V. Shelton, W.J. Wepfer, D.J. Miles, Ramp wave analysis of the solid/vapour heat pump, Journal of Energy Research and Technology (112) (1990) 69–77. [15] R.E. Critoph, Forced convection adsorption cycles, Applied Thermal Engineering (18) (1988) 799–807. [16] P. Neksa, CO2 heat pump systems, International Journal of Refrigeration 25 (2002) 421–427. [17] M.H. Kim, J. Pettersen, C.W. Bullard, Fundamental process and system design issues in CO2 vapor compression systems, Progress in Energy and Combustion Science 30 (2004) 119–174. [18] N.C. Srivastava, I.W. Eames, A review of adsorbents and adsorbates in solid–vapour adsorption heat pump systems, Applied Thermal Engineering 18 (1998) 707–714. [19] F. Meunier, Second law analysis of solid adsorption heat pump operating on reversible cascade cycle: application to the zeolite–water pair, Heat Recovery System 5 (2) (1985) 133–141. [20] R.E. Critoph, H.L. Turner, Performance of ammonia-activated carbon and ammonia zeolite heat pump adsorption cycle, Proc. Pompes a Chaleur Chimiques De Hautes Performances Perpignan, Lavoisier, Paris (September) 1989. [21] L. Turner, Improvement of activated charcoal–ammonia adsorption heat pumping/refrigeration cycles. Investigation of porosity and heat / mass transfer characteristics, Ph.D. Thesis, University of Warwick, UK, (1992). [22] K.C. Ng, H.T. Chua, C.Y. Chung, et al., Experimental investigation of the silica gel–water adsorption isotherm characteristics, Applied Thermal Engineering 21 (16) (2001) 1631–1642. [23] N.B. Amar, L.M. Sun, F. Meunier, Numerical analysis of adsorption temperature wave regenerative heat pump, Applied Thermal Engineering 16 (5) (1996) 405–418. [24] R.E. Critoph, Adsorption refrigeration and heat pumps, in: Trimothy D. Burchell (Ed.), Carbon Materials for Advanced Technologies, Pergamon Press, 1999. [25] R.E. Critoph, Evaluation of alternative refrigerant–adsorbent pairs for refrigeration cycles, Applied Thermal Engineering 16 (11) (1996) 891–900.