- Email: [email protected]
Contributions of system components and operating conditions to the performance of desiccant cooling systems
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 4 ( 2 0 1 1 ) 9 2 2 e9 2 7
available at www.sciencedirect.com
w w w . i i fi i r . o r g
journal homepage: www.elsevier.com/locate/ijrefrig
Contributions of system components and operating conditions to the performance of desiccant cooling systems Jae Dong Chung a,*, Dae-Young Lee b a b
Department of Mechanical Engineering, Sejong University, Seoul 143-747, Republic of Korea Korea Institute of Science and Technology, Seoul 136-791, Republic of Korea
article info
abstract
Article history:
This study systematically analyzes the effect of various kinds of design parameters on the
Received 7 December 2010
performance of a desiccant cooling system under two different system configurations. The
Received in revised form
considered parameters include system components such as the sensible heat exchanger,
22 February 2011
regenerative evaporative cooler and desiccant wheel, as well as operating conditions of
Accepted 8 March 2011
outdoor conditions, regenerative temperature and rate of outdoor influx. Numerical
Available online 15 March 2011
simulation has been conducted for these 11 design parameters with 3 levels. The orthogonal array L27(313) is adopted for the analysis of variance. In the range of the parameters
Keywords:
considered, the regenerative temperature is found to be the most dominant parameter of
Cooling system
contribution ratio of 31.9% and 23.9% for each system configuration. In the case of confined
Desiccant wheel
interest of the applications such as a district cooling system or a solar system using
Regenerative temperature
medium-temperature collectors, the cooling performance of the regenerative evaporative
Evaporative cooler
cooler is the most crucial for the system performance. ª 2011 Elsevier Ltd and IIR. All rights reserved.
Impact des composants des syste`mes et des conditions de fonctionnement sur la performance des syste`mes de refroidissement a` de´shydratant Mots cle´s : Syste`me de refroidissement ; Roue de´shydratante ; Tempe´rature de re´ge´ne´ration ; Refroidisseur a` e´vaporation
1.
Introduction
Indoor air quality (IAQ) has received increasing attention in recent years (Yu et al., 2009). Besides temperature control to control IAQ, the optimum level of indoor humidity should be maintained to ensure a comfortable and healthy environment. The desiccant cooling system is one of the promising solutions
for these purposes. Desiccant cooling systems also have advantages in environmental-conscious operations and separate controls for sensible and latent cooling loads which lead to comfortable indoor air quality. In addition, the desiccant cooling system is a heat-driven cycle and so it has the promise of being able to use low density energy such as natural gas, waste heat and solar energy. The US-Department of Energy estimated
* Corresponding author. Tel.: þ82 2 3408 3776; fax: þ82 2 3408 4333. E-mail address: [email protected] (J.D. Chung). 0140-7007/$ e see front matter ª 2011 Elsevier Ltd and IIR. All rights reserved. doi:10.1016/j.ijrefrig.2011.03.003
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 4 ( 2 0 1 1 ) 9 2 2 e9 2 7
Nomenclature a ANOVA b BP c cp COP D f F0 h Le N _ m Q ST t T
channel height (m) analysis of variance channel width (m) rate of outdoor influx channel wall thickness (m) specific heat of dry air water (J kg1 K1) coefficient of performance dew point temperature ( C) mass fraction of desiccant in the wheel ratio of mean squares of factor i to error enthalpy (J kg1) Lewis number number of experiment mass flow rate (kg h1) cooling capacity per unit air flow rate (J kg1) sum of square of experiment time (s) temperature ( C)
that desiccant cooling systems could reduce annual energy consumption by 117.2 million MWh and carbon dioxide emissions by 6 million tons by 2010 (Pesaran et al., 1992). Various aspects of desiccant cooling systems have been intensively investigated by many researchers. The reported works are related to feasibility studies (Mavroudaki et al., 2002; Halliday et al., 2002), performance predictions (Dai et al., 2001b; Mazzei et al., 2002), wheel optimization (Maclaine-Cross, 1988; Collier and Cohen, 1991; Zheng et al., 1995; Dai et al., 2001a; Chung et al., 2009, 2010) and the development of new materials (Cui et al., 2005; Jia et al., 2006; Chung and Lee, 2009). The wheel is considered the most crucial component of the system, but comparatively scant attention has been given to its relative contribution to the whole system. Also, the contribution and optimum condition of other components of the desiccant system such as the regenerative evaporative cooler (REC) and sensible heat exchanger have not been scrutinized in detail. Besides the contribution of each system component, contributions of operating conditions such as outdoor conditions, regenerative temperature and rate of outdoor influx need to be examined. In the present study, two different configurations of the desiccant system are considered. The contribution of each component of the desiccant system and the contributions of operating conditions for each system configuration have been systematically examined. Numerical simulation has been conducted for 11 design parameters with 3 levels. Orthogonal array L27(313) is adopted for the analysis of variance.
2.
Analysis
2.1.
Model description
Fig. 1 shows two different configurations of the desiccant cooling system considered in the present work, and Fig. 2 is the psychrometric charts corresponding to Fig. 1(a) compared with
V Wmax a 3 4 r
923
mean square maximum humidity ration of dry air (kg kg1) significance level performance relative humidity or degree of freedom density (kg m3)
Subscripts 1, 2, 3 levels air air d desiccant e error h hot in inlet out outlet p process r regeneration REC regenerative evaporative cooler SHE sensible heat exchanger
the conventional air-conditioning system using vapor compression refrigeration. In the conventional system air must be dehumidified by cooling it below its dew point to meet the latent load (② / ª), and reheating is often required (ª / ⑤) to satisfy the sensible heat factor (SHF) needed. This implies very poor energy efficiency, especially for low SHF, i.e. high latent cooling load. Also, a very low temperature of the supplied air can create draft situations in the air conditioned space. In the desiccant cooling system, the heated and dehumidified supply air exits from the humidification section of the rotor (② / ③). Afterwards the hot and dry air is cooled by a sensible heat exchanger (③ / ④). Finally, this flow is subjected to evaporative cooling to get cold and humid air (④ / ⑤) to be supplied to the conditioned space. The sensible heat exchanger acts as a pre-cooler after the desiccant and also as a pre-heater before the regeneration section, which results in enhanced performance of the whole system. On the other hand, in system configuration B of Fig. 1(b) the sensible heat exchanger fulfills the role of pre-heater before the heating coil which reduces the required energy for regeneration and results in enhanced performance. The performances of each system are compared.
2.2.
Numerical simulation
Primary numerical burden is from the analysis of the rotary desiccant wheel. The unsteady one-dimensional coupled heat and mass transfer model is developed for the rotary desiccant wheel. Details regarding governing equations, assumptions, numerical approaches and its validation have been described recently in Chung et al. (2009), so they will not be discussed further here. Simulations have been conducted for a desiccant wheel of width 0.2 m with a wall thickness (c) of 0.15 mm. The geometry of the channels in the wheel shown in Fig. 3 is sinusoidal with a width (b) of 3.5 mm and a height (a) of 1.75 mm. The air velocity is 2 m/s in both the adsorption and regeneration processes. The convective heat transfer coefficient and hydraulic diameter are calculated from the Nusselt number in the sinusoidal shaped
924
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 4 ( 2 0 1 1 ) 9 2 2 e9 2 7
Fig. 2 e Psychrometric processes of the desiccant cooling system (corresponding to Fig. 1(a)) and the conventional air-conditioning system using vapor compression refrigeration cycle.
Fig. 1 e Schematics of desiccant cooling system (a) configuration A and (b) configuration B. Fig. 3 e Schematics of the desiccant wheel and computational domains. channels (Chung and Lee, 2009; Kakac et al., 1987) and the mass transfer coefficient is obtained based on the assumption of Le ¼ 1. The desired indoor condition is fixed at T1 ¼ 27 C and 41 ¼ 50%. The 1:1 split between regeneration and dehumidification sections of the desiccant wheel is assumed since the present work is focused on low regeneration temperatures. Table 1 summarizes the data for the base conditions employed in the simulations and the amount of variations for the analysis of variance. It needs to be noted that the regeneration temperature is set at T10 ¼ 75 C which is low enough to be acquired from low density energy resources such as local heating, solar energy and waste heat. Base conditions for desiccant material are from silica gel. Parameters related to desiccant material are mass fraction, heat capacity, density, and maximum water uptake. Wheel performance dependency on these parameters is discussed by Chung et al. (2010). Wheel speed is set at 150 s/rev which is near the optimum wheel speed obtained in the earlier work (Chung et al., 2009).
Table 1 e Base condition and 3 levels of design parameters employed in the simulations. Factor
BP Tair [ C] 4air [%] 3SHE [-] 3REC [-] Tr [ C] tp [s] fd [-] cp,d [J/kg K] rd [kg/m3] Wmax [-]
Level 1
2 (base condition)
3
10% 10% 10% 10% 10% 30% 30% 30% 30% 30% 30%
0.3 35 40 0.8 0.8 75 75 0.75 921 720 0.4
þ10% þ10% þ10% þ10% þ10% þ30% þ30% þ30% þ30% þ30% þ30%
925
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 4 ( 2 0 1 1 ) 9 2 2 e9 2 7
Table 2 e ANOVA for system configuration A when COP is chosen as the indicator of system performance.
Sum of squares Degree of freedom Mean square F0 (before pooling) F0 (after pooling) Contribution ratio
BP
Tair
4air
3SHE
0.01724 2 0.00862 0.42193
0.33964 2 0.16982 8.31197 13.2069 0.255
0.13678 2 0.06839 3.3474 5.31877 0.09
0.00131 2 0.00065 0.03199
3REC
tp
fd
cp,d
rd
0.01921 2 0.00960 0.47007
0.01930 2 0.00965 0.47226
0.05303 2 0.02652 1.29790
0.00610 2 0.00305 0.14927
Tr
0.10569 0.41927 2 2 0.05284 0.20964 2.58648 10.26078 4.10972 16.30360 0.065 0.319
Wmax
e
0.03354 0.08172(0.23145) 2 4(18) 0.01677 0.02043(0.012858) 0.82081 0.271
Table 3 e ANOVA for system configuration B when COP is chosen as the indicator of system performance. BP Sum of squares Degree of freedom Mean square F0 (before pooling) F0 (after pooling) Contribution ratio
Tair
0.02027 0.28244 2 2 0.01014 0.14122 1.12888 15.72880 9.654592 0.206
4air 0.13621 2 0.06810 7.585232 4.6559381 0.087
3SHE
3REC
0.03634 0.22273 0.32192 2 2 2 0.01817 0.11136 0.16096 2.023568 12.40336 17.92703 7.613386 11.00390 0.158 0.239
The cooling performance of the REC can be represented by the cooling effectiveness defined as 3REC ¼
Th;in Th;out Th;in Dh;in
(1)
where Dh,in is the inlet dew point temperature of the process air. It is the lowest outlet temperature that can be achieved in the REC. The base condition for 3REC is 0.8. The performance of the sensible heat exchanger can be evaluated by 3SHE ¼
T9 T8 for Cycle A T3 T8
(2)
3SHE ¼
T9 T8 for Cycle B T11 T8
(3)
The base condition for 3SHE is 0.8. System performance is the object function for ANOVA and can be evaluated by the amount of cooling capacity per unit air flow rate and/or COP defined as the ratio of cooling capacity to required energy for regeneration. Q ¼ ðh1 h6 Þð1 BPÞ COP ¼
Tr
(4)
Q Qr
(5)
Where the required energy for regeneration is obtained by _ p ðh10 h9 Þtr =tp . Qr ¼ m
3.
tp
fd
cp,d
rd
0.02512 2 0.01256 1.39873
0.00851 2 0.00426 0.47391
0.07827 2 0.03914 4.358865
0.03255 2 0.01627 1.812563
Wmax
e
0.02632 0.03591(0.26329) 2 4(18) 0.01316 0.00898(0.14627) 1.46586 0.310
Analysis of variance
The system design of the desiccant cooling system poses a number of questions associated with a single parameter and/ or their combinations. These include system components such as the sensible heat exchanger, regenerative evaporative cooler and desiccant wheel, as well as the operating conditions of outdoor conditions, regenerative temperature and rate of outdoor influx. Also, the system configuration, see Fig. 1, and parameters related to desiccant wheel such as wheel speed and properties of desiccant material, should be considered. Assessing the great number of parameters involved in the performance of a desiccant cooling system is a time-intensive task. It is therefore impractical to evaluate each individual case by exhaustive approach. In the present study, numerical simulation has been conducted for 11 design parameters with 3 levels. The considered design parameters and its variations (30% for the properties of desiccant material and 10% for the others) are summarized in Table 1. Orthogonal array L27(313) was adopted for the analysis of variance (ANOVA) (Park, 1996) to find the optimal condition of each parameter and give the individual impact ratio of the system performance. The sum of the square is defined as P ST ¼ x2ij ðT2 =NÞ, where x is the value of the numerical P experiment, Tð¼ xij is the summation of the numerical experiments, and N is the number of numerical experiments. We denote the degree of freedom by 4, with the appropriate
Table 4 e ANOVA for system configuration A when Q is chosen as the indicator of system performance. BP Sum of squares Degree of freedom Mean square F0 (before pooling) F0 (after pooling) Contribution ratio
Tair
RHair
8.33292 337.44301 98.83324 2 2 2 4.1665 168.7215 49.4166 0.8667 35.0976 10.2797 31.1154 9.11334 0.238 0.064
3SHE
3REC
Treg
spro
fd
cp,d
rd
Wmax
e
5.03098 88.22058 749.5578 8.44539 28.60372 2.39312 14.2557 11.3134 19.22886(97.60410) 2 2 2 2 2 2 2 2 4(18) 2.5155 44.1103 374.7789 4.2227 14.3019 1.1966 7.1279 5.6567 4.8072(5.42245) 0.52327 9.1759 77.9617 0.8784 2.9751 0.2489 1.4827 1.1767 8.13475 69.1162 0.056 0.539 0.103
926
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 4 ( 2 0 1 1 ) 9 2 2 e9 2 7
Table 5 e ANOVA of fixed regenerative temperature and fixed outdoor conditions for system configuration A when COP is chosen as the indicator of system performance.
Sum of squares Degree of freedom Mean square F0 Contribution ratio
BP
3SHE
3REC
tp
fd
cp,d
rd
Wmax
e
0.00455 2 0.002273
0.00929 2 0.004644
0.06778 2 0.033889 19.982616 0.325
0.02399 2 0.011996 7.0734978 0.104
0.03755 2 0.018774 11.069915 0.172
0.00032 2 0.00016
0.03826 2 0.019129 11.279383 0.176
0.00076 2 0.000379
0.01561 10(18) 0.001561(0.001696)
subscript i (¼BP, Tair, 4air, 3SHE, 3REC, Tr, tp, fd, cp,d, rd, Wmax), which is defined as the level of factor i minus 1. The mean square, V is obtained by ST/4. For a given level of significance, if F0(¼Vi/Ve) is greater than the corresponding F-distribution, (i.e., F(4i,4e; a)), we say that factor i is significant, where a means significance level. If we chose COP as the indicator of system performance, the results of ANOVA are summarized in Tables 2 and 3 for the configurations A and B of desiccant systems shown in Fig. 1. The F0 values of 4air and 3REC are 3.35 and 2.59, which is not small enough to be ignored in the case of significance level of 5%. Thus we may consider that the factors 4air and 3REC are also slightly significant. Regardless of the system configuration, it is found that the contribution of BP, 3SHE, tp and parameters related to desiccant material such as fd, cp,d, rd, and Wmax are negligible, which may be surprising since it is believed that the desiccant material is one of the dominant factors on the system performance. Recently, Lee et al. (Faust et al., 2007) developed a highpowered, SAP-based desiccant that can absorb 4e5 times more water than silica gel or zeolite, which showed apparent system performance enhancement. We can conclude that more than 30% variations of desiccant material properties are required for the development a new desiccant material for the meaningful enhancement of system performance. After pooling the contribution of the negligible parameters, the contribution ratio of each factor is re-evaluated. The procedure is summarized in the ANOVA in Tables 2 and 3, which shows that factors Tair, 4air, 3REC and Tr are significant at the 5% level. Note that the values in parentheses are after pooling. The details of the procedure on the ANOVA can be found in Park (1996). The result gives quantitative estimation of the various design parameters affecting the performance and helps to determine the main factors for optimum design of a desiccant cooling system. The regeneration temperature is found to be the most dominant parameter (31.9% for configuration A and 23.9% for system configuration B) and the outdoor conditions play a significant role on the performance of a desiccant cooling system (25.5% and 9% for Tair and 4air, respectively, for system configuration A and 20.6% and 8.7% for
0.223
Tair and 4air, respectively, for system configuration B). These results are completely in line with those if we chose cooling capacity Q as the indicator of system performance. For reference, the result for system configuration A is given in Table 4. If we assume the that errors are normally distributed, the 100(1 a)% confidence interval on the t-distribution is sffiffiffiffiffiffi a V e y t fe ; (6) ne 2 Where 4e and Ve are the degree of freedom and mean square for error, respectively. ne denotes the effective number of replications that can be obtained from the relationship ne ¼ P
n fi þ 1
(7)
The optimum levels of significant factors in system configuration A that maximize the COP are Tair,1, 4air,1, 3REC,3 and Tr,1 Note that the estimated COP at optimum conditions is 0.9896 within the confidence internal 1.0374 0.1375. The optimum levels of significant factors in system configuration B are same as system configuration A and the estimated COP at optimum conditions is 1.28577, which is also within the confidence internal 1.2042 0.1467. In order to focus on the application of a district cooling system to utilize the over-supplied heat in summer season or a solar system using medium-temperature collectors, the regeneration temperature is fixed at a typical value of 75 C. And outdoor conditions are also fixed as T7 ¼ 35 C, 47 ¼ 40%, the typical summer condition in Korea. The results of ANOVA are summarized in Tables 5 and 6 for both A and B system configurations. For system configuration A, it is found that only the contributions of 3REC, tp, fd, rd are meaningful. The most dominant parameter is 3REC whose contribution ratio to the system performance is 32.5%. General trends of wheel parameters fd and rd are same as earlier research by Chung et al. (2010) although the analysis in Chung et al. (2010) is confined only to wheel performance, i.e. wheel performance is increased as fd and rd become higher.
Table 6 e ANOVA of fixed regenerative temperature and fixed outdoor conditions for system configuration B when COP is chosen as the indicator of system performance.
Sum of squares Degree of freedom Mean square F0 Contribution ratio
BP
3SHE
3REC
tp
fd
cp,d
rd
Wmax
e
0.02896 2 0.014479 39.995087 0.088
0.05498 2 0.027491 75.937071 0.169
0.19084 2 0.095422 263.5757 0.593
0.00730 2 0.003651 10.084541 0.021
0.00694 2 0.003468 9.5791649 0.019
0.02217 2 0.011084 30.617604 0.067
0.00487 2 0.002433 6.7198643 0.013
0.00109 2 0.000543
0.00326(0.00434) 10(12) 0.000326(0.000362) 0.029
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 4 ( 2 0 1 1 ) 9 2 2 e9 2 7
Contrary to system configuration A, it is found that the sensible heat exchanger plays significant role in system configuration B whose contribution ratio is estimated as much as 16.9%. The impact of the cooling performance of REC is recognized as the most crucial factor also in this configuration. This suggests further research attention on regenerative evaporative cooler should be given to enhance the system performance. In this system configuration, the general trends of wheel parameters fd, cp,d and rd are also same as the earlier research by Chung et al. (2010).
4.
Conclusions
The analysis on the effect of various kinds of design parameters on the performance of a desiccant cooling system has been carried out for the two different desiccant cooling system configurations. The considered parameters include system components such as the sensible heat exchanger, regenerative evaporative cooler and desiccant wheel, as well as the operating conditions of outdoor conditions, regenerative temperature and rate of outdoor influx. Also, the parameters related to desiccant wheel such as wheel speed and properties of the desiccant material, are examined. Numerical simulation has been conducted for these 11 design parameters with 3 levels (30% for the properties of desiccant material and 10% for the others). Orthogonal array L27(313) is adopted for the analysis of variance. In the range of parameters considered, the regenerative temperature is found to be the most dominant parameter of contribution ratio of 31.9% and 23.9% for each system configuration. Regardless of the system configuration, the contribution of BP, 3SHE, tp and parameters related to desiccant material such as fd, cp,d, rd, and Wmax is found negligible at the 5% significance level, which may be surprising since it is believed that the desiccant material is one of the dominant factors on the system performance. Thus more than 30% variations of the properties of desiccant material would be required to develop a new desiccant material for the meaningful enhancement of system performance. In the case of the confined interest of the applications such as a district cooling system or solar system using medium-temperature collectors, the cooling performance of the regenerative evaporative cooler is the most crucial for the system performance. For all cases, the estimated COPs at optimum conditions are confirmed within the confidence internal.
Acknowledgement This research was supported by the Small & Medium Business Administration of Korean government.
927
references
Chung, J.D., Lee, D.Y., 2009. Effect of desiccant isotherm on the performance of desiccant wheel. Int. J. Refrigeration 32, 720e726. Chung, J.D., Lee, D.Y., Yoon, S.M., 2009. Optimization of desiccant wheel speed and area ratio of regeneration to dehumidification according to regeneration temperature. Solar Energy 83, 625e635. Chung, J.D., Lee, D.Y., Yoon, S.M., 2010. Effect of parameters related to channel and desiccant on the size of desiccant rotor. Int. J. Air-Conditioning Refrigeration 18, 201e211. Collier, R.K., Cohen, B.M., 1991. An analytic investigation of methods for improving the performance of desiccant cooling system. ASME J. Solar Energy Sci. Eng. 113, 157e163. Cui, Q., Chen, H., Tao, G., Yao, H., 2005. Performance study of new adsorbent for solid desiccant cooling. Energy 30, 273e279. Dai, Y.J., Wang, H.F., Zhang, H.F., 2001a. Parameter analysis to improve rotary desiccant dehumidification using a mathematical model. Int. J. Therm. Sci. 40, 400e408. Dai, Y.J., Wang, R.Z., Zhang, H.F., Yu, J.D., 2001b. Use of liquid desiccant cooling to improve the performance of vapour compression air conditioning. Appl. Therm. Eng. 21, 1185e1205. Faust S., Guido F., Lee D.Y., 2007. Desiccant element and manufacturing method of the same, Korea Patent 0704235. Halliday, S.P., Beggs, C.B., Sleigh, P.A., 2002. The use of solar desiccant cooling in the UK: a feasibility study. Appl. Therm. Eng. 22, 1327e1338. Jia, C.X., Dai, Y.J., Wu, J.Y., Wang, R.Z., 2006. Experimental comparison of two honeycombed desiccant wheels fabricated with silica gel and composite desiccant material. Energy Convers. Manage. 47, 2523e2534. Kakac, S., Shah, R.K., Aung, W. (Eds.), 1987. Handbook of SinglePhase Convective Heat Transfer. John Wiley & Sons, Inc. Maclaine-Cross, I.L., 1988. Performance of a solar desiccant cooling system. ASHRAE Trans. 94, 1997e2009. Mavroudaki, P., Beggs, C.B., Sleigh, P.A., Halliday, S.P., 2002. The potential for solar powered single-stage desiccant cooling in southern Europe. Appl. Therm. Eng. 22, 1129e1140. Mazzei, P., Minichiello, F., Palma, D., 2002. Desiccant HVAC systems for commercial buildings. Appl. Therm. Eng. 22, 545e560. Park, S.H., 1996. Robust Design and Analysis for Quality Engineering. Chapman & Hall, London, UK. Pesaran A.A., Penney T.R., Czanderna A.W., 1992 Desiccant cooling: state-of-the-art assessment. NREL Report No. TP-2544147, p. 221. Yu, B.F., Hu, Z.B., Liu, M., Yang, H.L., Kong, Q.X., Liu, Y.H., 2009. Review of research on air-conditioning systems and indoor air quality control for human health. Int. J. Refrigeration 32, 3e20. Zheng, W., Worek, W.M., Novesel, D., 1995. Performance optimization of rotary dehumidifiers. J. Solar. Energy Eng. 117, 40e44.
available at www.sciencedirect.com
w w w . i i fi i r . o r g
journal homepage: www.elsevier.com/locate/ijrefrig
Contributions of system components and operating conditions to the performance of desiccant cooling systems Jae Dong Chung a,*, Dae-Young Lee b a b
Department of Mechanical Engineering, Sejong University, Seoul 143-747, Republic of Korea Korea Institute of Science and Technology, Seoul 136-791, Republic of Korea
article info
abstract
Article history:
This study systematically analyzes the effect of various kinds of design parameters on the
Received 7 December 2010
performance of a desiccant cooling system under two different system configurations. The
Received in revised form
considered parameters include system components such as the sensible heat exchanger,
22 February 2011
regenerative evaporative cooler and desiccant wheel, as well as operating conditions of
Accepted 8 March 2011
outdoor conditions, regenerative temperature and rate of outdoor influx. Numerical
Available online 15 March 2011
simulation has been conducted for these 11 design parameters with 3 levels. The orthogonal array L27(313) is adopted for the analysis of variance. In the range of the parameters
Keywords:
considered, the regenerative temperature is found to be the most dominant parameter of
Cooling system
contribution ratio of 31.9% and 23.9% for each system configuration. In the case of confined
Desiccant wheel
interest of the applications such as a district cooling system or a solar system using
Regenerative temperature
medium-temperature collectors, the cooling performance of the regenerative evaporative
Evaporative cooler
cooler is the most crucial for the system performance. ª 2011 Elsevier Ltd and IIR. All rights reserved.
Impact des composants des syste`mes et des conditions de fonctionnement sur la performance des syste`mes de refroidissement a` de´shydratant Mots cle´s : Syste`me de refroidissement ; Roue de´shydratante ; Tempe´rature de re´ge´ne´ration ; Refroidisseur a` e´vaporation
1.
Introduction
Indoor air quality (IAQ) has received increasing attention in recent years (Yu et al., 2009). Besides temperature control to control IAQ, the optimum level of indoor humidity should be maintained to ensure a comfortable and healthy environment. The desiccant cooling system is one of the promising solutions
for these purposes. Desiccant cooling systems also have advantages in environmental-conscious operations and separate controls for sensible and latent cooling loads which lead to comfortable indoor air quality. In addition, the desiccant cooling system is a heat-driven cycle and so it has the promise of being able to use low density energy such as natural gas, waste heat and solar energy. The US-Department of Energy estimated
* Corresponding author. Tel.: þ82 2 3408 3776; fax: þ82 2 3408 4333. E-mail address: [email protected] (J.D. Chung). 0140-7007/$ e see front matter ª 2011 Elsevier Ltd and IIR. All rights reserved. doi:10.1016/j.ijrefrig.2011.03.003
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 4 ( 2 0 1 1 ) 9 2 2 e9 2 7
Nomenclature a ANOVA b BP c cp COP D f F0 h Le N _ m Q ST t T
channel height (m) analysis of variance channel width (m) rate of outdoor influx channel wall thickness (m) specific heat of dry air water (J kg1 K1) coefficient of performance dew point temperature ( C) mass fraction of desiccant in the wheel ratio of mean squares of factor i to error enthalpy (J kg1) Lewis number number of experiment mass flow rate (kg h1) cooling capacity per unit air flow rate (J kg1) sum of square of experiment time (s) temperature ( C)
that desiccant cooling systems could reduce annual energy consumption by 117.2 million MWh and carbon dioxide emissions by 6 million tons by 2010 (Pesaran et al., 1992). Various aspects of desiccant cooling systems have been intensively investigated by many researchers. The reported works are related to feasibility studies (Mavroudaki et al., 2002; Halliday et al., 2002), performance predictions (Dai et al., 2001b; Mazzei et al., 2002), wheel optimization (Maclaine-Cross, 1988; Collier and Cohen, 1991; Zheng et al., 1995; Dai et al., 2001a; Chung et al., 2009, 2010) and the development of new materials (Cui et al., 2005; Jia et al., 2006; Chung and Lee, 2009). The wheel is considered the most crucial component of the system, but comparatively scant attention has been given to its relative contribution to the whole system. Also, the contribution and optimum condition of other components of the desiccant system such as the regenerative evaporative cooler (REC) and sensible heat exchanger have not been scrutinized in detail. Besides the contribution of each system component, contributions of operating conditions such as outdoor conditions, regenerative temperature and rate of outdoor influx need to be examined. In the present study, two different configurations of the desiccant system are considered. The contribution of each component of the desiccant system and the contributions of operating conditions for each system configuration have been systematically examined. Numerical simulation has been conducted for 11 design parameters with 3 levels. Orthogonal array L27(313) is adopted for the analysis of variance.
2.
Analysis
2.1.
Model description
Fig. 1 shows two different configurations of the desiccant cooling system considered in the present work, and Fig. 2 is the psychrometric charts corresponding to Fig. 1(a) compared with
V Wmax a 3 4 r
923
mean square maximum humidity ration of dry air (kg kg1) significance level performance relative humidity or degree of freedom density (kg m3)
Subscripts 1, 2, 3 levels air air d desiccant e error h hot in inlet out outlet p process r regeneration REC regenerative evaporative cooler SHE sensible heat exchanger
the conventional air-conditioning system using vapor compression refrigeration. In the conventional system air must be dehumidified by cooling it below its dew point to meet the latent load (② / ª), and reheating is often required (ª / ⑤) to satisfy the sensible heat factor (SHF) needed. This implies very poor energy efficiency, especially for low SHF, i.e. high latent cooling load. Also, a very low temperature of the supplied air can create draft situations in the air conditioned space. In the desiccant cooling system, the heated and dehumidified supply air exits from the humidification section of the rotor (② / ③). Afterwards the hot and dry air is cooled by a sensible heat exchanger (③ / ④). Finally, this flow is subjected to evaporative cooling to get cold and humid air (④ / ⑤) to be supplied to the conditioned space. The sensible heat exchanger acts as a pre-cooler after the desiccant and also as a pre-heater before the regeneration section, which results in enhanced performance of the whole system. On the other hand, in system configuration B of Fig. 1(b) the sensible heat exchanger fulfills the role of pre-heater before the heating coil which reduces the required energy for regeneration and results in enhanced performance. The performances of each system are compared.
2.2.
Numerical simulation
Primary numerical burden is from the analysis of the rotary desiccant wheel. The unsteady one-dimensional coupled heat and mass transfer model is developed for the rotary desiccant wheel. Details regarding governing equations, assumptions, numerical approaches and its validation have been described recently in Chung et al. (2009), so they will not be discussed further here. Simulations have been conducted for a desiccant wheel of width 0.2 m with a wall thickness (c) of 0.15 mm. The geometry of the channels in the wheel shown in Fig. 3 is sinusoidal with a width (b) of 3.5 mm and a height (a) of 1.75 mm. The air velocity is 2 m/s in both the adsorption and regeneration processes. The convective heat transfer coefficient and hydraulic diameter are calculated from the Nusselt number in the sinusoidal shaped
924
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 4 ( 2 0 1 1 ) 9 2 2 e9 2 7
Fig. 2 e Psychrometric processes of the desiccant cooling system (corresponding to Fig. 1(a)) and the conventional air-conditioning system using vapor compression refrigeration cycle.
Fig. 1 e Schematics of desiccant cooling system (a) configuration A and (b) configuration B. Fig. 3 e Schematics of the desiccant wheel and computational domains. channels (Chung and Lee, 2009; Kakac et al., 1987) and the mass transfer coefficient is obtained based on the assumption of Le ¼ 1. The desired indoor condition is fixed at T1 ¼ 27 C and 41 ¼ 50%. The 1:1 split between regeneration and dehumidification sections of the desiccant wheel is assumed since the present work is focused on low regeneration temperatures. Table 1 summarizes the data for the base conditions employed in the simulations and the amount of variations for the analysis of variance. It needs to be noted that the regeneration temperature is set at T10 ¼ 75 C which is low enough to be acquired from low density energy resources such as local heating, solar energy and waste heat. Base conditions for desiccant material are from silica gel. Parameters related to desiccant material are mass fraction, heat capacity, density, and maximum water uptake. Wheel performance dependency on these parameters is discussed by Chung et al. (2010). Wheel speed is set at 150 s/rev which is near the optimum wheel speed obtained in the earlier work (Chung et al., 2009).
Table 1 e Base condition and 3 levels of design parameters employed in the simulations. Factor
BP Tair [ C] 4air [%] 3SHE [-] 3REC [-] Tr [ C] tp [s] fd [-] cp,d [J/kg K] rd [kg/m3] Wmax [-]
Level 1
2 (base condition)
3
10% 10% 10% 10% 10% 30% 30% 30% 30% 30% 30%
0.3 35 40 0.8 0.8 75 75 0.75 921 720 0.4
þ10% þ10% þ10% þ10% þ10% þ30% þ30% þ30% þ30% þ30% þ30%
925
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 4 ( 2 0 1 1 ) 9 2 2 e9 2 7
Table 2 e ANOVA for system configuration A when COP is chosen as the indicator of system performance.
Sum of squares Degree of freedom Mean square F0 (before pooling) F0 (after pooling) Contribution ratio
BP
Tair
4air
3SHE
0.01724 2 0.00862 0.42193
0.33964 2 0.16982 8.31197 13.2069 0.255
0.13678 2 0.06839 3.3474 5.31877 0.09
0.00131 2 0.00065 0.03199
3REC
tp
fd
cp,d
rd
0.01921 2 0.00960 0.47007
0.01930 2 0.00965 0.47226
0.05303 2 0.02652 1.29790
0.00610 2 0.00305 0.14927
Tr
0.10569 0.41927 2 2 0.05284 0.20964 2.58648 10.26078 4.10972 16.30360 0.065 0.319
Wmax
e
0.03354 0.08172(0.23145) 2 4(18) 0.01677 0.02043(0.012858) 0.82081 0.271
Table 3 e ANOVA for system configuration B when COP is chosen as the indicator of system performance. BP Sum of squares Degree of freedom Mean square F0 (before pooling) F0 (after pooling) Contribution ratio
Tair
0.02027 0.28244 2 2 0.01014 0.14122 1.12888 15.72880 9.654592 0.206
4air 0.13621 2 0.06810 7.585232 4.6559381 0.087
3SHE
3REC
0.03634 0.22273 0.32192 2 2 2 0.01817 0.11136 0.16096 2.023568 12.40336 17.92703 7.613386 11.00390 0.158 0.239
The cooling performance of the REC can be represented by the cooling effectiveness defined as 3REC ¼
Th;in Th;out Th;in Dh;in
(1)
where Dh,in is the inlet dew point temperature of the process air. It is the lowest outlet temperature that can be achieved in the REC. The base condition for 3REC is 0.8. The performance of the sensible heat exchanger can be evaluated by 3SHE ¼
T9 T8 for Cycle A T3 T8
(2)
3SHE ¼
T9 T8 for Cycle B T11 T8
(3)
The base condition for 3SHE is 0.8. System performance is the object function for ANOVA and can be evaluated by the amount of cooling capacity per unit air flow rate and/or COP defined as the ratio of cooling capacity to required energy for regeneration. Q ¼ ðh1 h6 Þð1 BPÞ COP ¼
Tr
(4)
Q Qr
(5)
Where the required energy for regeneration is obtained by _ p ðh10 h9 Þtr =tp . Qr ¼ m
3.
tp
fd
cp,d
rd
0.02512 2 0.01256 1.39873
0.00851 2 0.00426 0.47391
0.07827 2 0.03914 4.358865
0.03255 2 0.01627 1.812563
Wmax
e
0.02632 0.03591(0.26329) 2 4(18) 0.01316 0.00898(0.14627) 1.46586 0.310
Analysis of variance
The system design of the desiccant cooling system poses a number of questions associated with a single parameter and/ or their combinations. These include system components such as the sensible heat exchanger, regenerative evaporative cooler and desiccant wheel, as well as the operating conditions of outdoor conditions, regenerative temperature and rate of outdoor influx. Also, the system configuration, see Fig. 1, and parameters related to desiccant wheel such as wheel speed and properties of desiccant material, should be considered. Assessing the great number of parameters involved in the performance of a desiccant cooling system is a time-intensive task. It is therefore impractical to evaluate each individual case by exhaustive approach. In the present study, numerical simulation has been conducted for 11 design parameters with 3 levels. The considered design parameters and its variations (30% for the properties of desiccant material and 10% for the others) are summarized in Table 1. Orthogonal array L27(313) was adopted for the analysis of variance (ANOVA) (Park, 1996) to find the optimal condition of each parameter and give the individual impact ratio of the system performance. The sum of the square is defined as P ST ¼ x2ij ðT2 =NÞ, where x is the value of the numerical P experiment, Tð¼ xij is the summation of the numerical experiments, and N is the number of numerical experiments. We denote the degree of freedom by 4, with the appropriate
Table 4 e ANOVA for system configuration A when Q is chosen as the indicator of system performance. BP Sum of squares Degree of freedom Mean square F0 (before pooling) F0 (after pooling) Contribution ratio
Tair
RHair
8.33292 337.44301 98.83324 2 2 2 4.1665 168.7215 49.4166 0.8667 35.0976 10.2797 31.1154 9.11334 0.238 0.064
3SHE
3REC
Treg
spro
fd
cp,d
rd
Wmax
e
5.03098 88.22058 749.5578 8.44539 28.60372 2.39312 14.2557 11.3134 19.22886(97.60410) 2 2 2 2 2 2 2 2 4(18) 2.5155 44.1103 374.7789 4.2227 14.3019 1.1966 7.1279 5.6567 4.8072(5.42245) 0.52327 9.1759 77.9617 0.8784 2.9751 0.2489 1.4827 1.1767 8.13475 69.1162 0.056 0.539 0.103
926
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 4 ( 2 0 1 1 ) 9 2 2 e9 2 7
Table 5 e ANOVA of fixed regenerative temperature and fixed outdoor conditions for system configuration A when COP is chosen as the indicator of system performance.
Sum of squares Degree of freedom Mean square F0 Contribution ratio
BP
3SHE
3REC
tp
fd
cp,d
rd
Wmax
e
0.00455 2 0.002273
0.00929 2 0.004644
0.06778 2 0.033889 19.982616 0.325
0.02399 2 0.011996 7.0734978 0.104
0.03755 2 0.018774 11.069915 0.172
0.00032 2 0.00016
0.03826 2 0.019129 11.279383 0.176
0.00076 2 0.000379
0.01561 10(18) 0.001561(0.001696)
subscript i (¼BP, Tair, 4air, 3SHE, 3REC, Tr, tp, fd, cp,d, rd, Wmax), which is defined as the level of factor i minus 1. The mean square, V is obtained by ST/4. For a given level of significance, if F0(¼Vi/Ve) is greater than the corresponding F-distribution, (i.e., F(4i,4e; a)), we say that factor i is significant, where a means significance level. If we chose COP as the indicator of system performance, the results of ANOVA are summarized in Tables 2 and 3 for the configurations A and B of desiccant systems shown in Fig. 1. The F0 values of 4air and 3REC are 3.35 and 2.59, which is not small enough to be ignored in the case of significance level of 5%. Thus we may consider that the factors 4air and 3REC are also slightly significant. Regardless of the system configuration, it is found that the contribution of BP, 3SHE, tp and parameters related to desiccant material such as fd, cp,d, rd, and Wmax are negligible, which may be surprising since it is believed that the desiccant material is one of the dominant factors on the system performance. Recently, Lee et al. (Faust et al., 2007) developed a highpowered, SAP-based desiccant that can absorb 4e5 times more water than silica gel or zeolite, which showed apparent system performance enhancement. We can conclude that more than 30% variations of desiccant material properties are required for the development a new desiccant material for the meaningful enhancement of system performance. After pooling the contribution of the negligible parameters, the contribution ratio of each factor is re-evaluated. The procedure is summarized in the ANOVA in Tables 2 and 3, which shows that factors Tair, 4air, 3REC and Tr are significant at the 5% level. Note that the values in parentheses are after pooling. The details of the procedure on the ANOVA can be found in Park (1996). The result gives quantitative estimation of the various design parameters affecting the performance and helps to determine the main factors for optimum design of a desiccant cooling system. The regeneration temperature is found to be the most dominant parameter (31.9% for configuration A and 23.9% for system configuration B) and the outdoor conditions play a significant role on the performance of a desiccant cooling system (25.5% and 9% for Tair and 4air, respectively, for system configuration A and 20.6% and 8.7% for
0.223
Tair and 4air, respectively, for system configuration B). These results are completely in line with those if we chose cooling capacity Q as the indicator of system performance. For reference, the result for system configuration A is given in Table 4. If we assume the that errors are normally distributed, the 100(1 a)% confidence interval on the t-distribution is sffiffiffiffiffiffi a V e y t fe ; (6) ne 2 Where 4e and Ve are the degree of freedom and mean square for error, respectively. ne denotes the effective number of replications that can be obtained from the relationship ne ¼ P
n fi þ 1
(7)
The optimum levels of significant factors in system configuration A that maximize the COP are Tair,1, 4air,1, 3REC,3 and Tr,1 Note that the estimated COP at optimum conditions is 0.9896 within the confidence internal 1.0374 0.1375. The optimum levels of significant factors in system configuration B are same as system configuration A and the estimated COP at optimum conditions is 1.28577, which is also within the confidence internal 1.2042 0.1467. In order to focus on the application of a district cooling system to utilize the over-supplied heat in summer season or a solar system using medium-temperature collectors, the regeneration temperature is fixed at a typical value of 75 C. And outdoor conditions are also fixed as T7 ¼ 35 C, 47 ¼ 40%, the typical summer condition in Korea. The results of ANOVA are summarized in Tables 5 and 6 for both A and B system configurations. For system configuration A, it is found that only the contributions of 3REC, tp, fd, rd are meaningful. The most dominant parameter is 3REC whose contribution ratio to the system performance is 32.5%. General trends of wheel parameters fd and rd are same as earlier research by Chung et al. (2010) although the analysis in Chung et al. (2010) is confined only to wheel performance, i.e. wheel performance is increased as fd and rd become higher.
Table 6 e ANOVA of fixed regenerative temperature and fixed outdoor conditions for system configuration B when COP is chosen as the indicator of system performance.
Sum of squares Degree of freedom Mean square F0 Contribution ratio
BP
3SHE
3REC
tp
fd
cp,d
rd
Wmax
e
0.02896 2 0.014479 39.995087 0.088
0.05498 2 0.027491 75.937071 0.169
0.19084 2 0.095422 263.5757 0.593
0.00730 2 0.003651 10.084541 0.021
0.00694 2 0.003468 9.5791649 0.019
0.02217 2 0.011084 30.617604 0.067
0.00487 2 0.002433 6.7198643 0.013
0.00109 2 0.000543
0.00326(0.00434) 10(12) 0.000326(0.000362) 0.029
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 4 ( 2 0 1 1 ) 9 2 2 e9 2 7
Contrary to system configuration A, it is found that the sensible heat exchanger plays significant role in system configuration B whose contribution ratio is estimated as much as 16.9%. The impact of the cooling performance of REC is recognized as the most crucial factor also in this configuration. This suggests further research attention on regenerative evaporative cooler should be given to enhance the system performance. In this system configuration, the general trends of wheel parameters fd, cp,d and rd are also same as the earlier research by Chung et al. (2010).
4.
Conclusions
The analysis on the effect of various kinds of design parameters on the performance of a desiccant cooling system has been carried out for the two different desiccant cooling system configurations. The considered parameters include system components such as the sensible heat exchanger, regenerative evaporative cooler and desiccant wheel, as well as the operating conditions of outdoor conditions, regenerative temperature and rate of outdoor influx. Also, the parameters related to desiccant wheel such as wheel speed and properties of the desiccant material, are examined. Numerical simulation has been conducted for these 11 design parameters with 3 levels (30% for the properties of desiccant material and 10% for the others). Orthogonal array L27(313) is adopted for the analysis of variance. In the range of parameters considered, the regenerative temperature is found to be the most dominant parameter of contribution ratio of 31.9% and 23.9% for each system configuration. Regardless of the system configuration, the contribution of BP, 3SHE, tp and parameters related to desiccant material such as fd, cp,d, rd, and Wmax is found negligible at the 5% significance level, which may be surprising since it is believed that the desiccant material is one of the dominant factors on the system performance. Thus more than 30% variations of the properties of desiccant material would be required to develop a new desiccant material for the meaningful enhancement of system performance. In the case of the confined interest of the applications such as a district cooling system or solar system using medium-temperature collectors, the cooling performance of the regenerative evaporative cooler is the most crucial for the system performance. For all cases, the estimated COPs at optimum conditions are confirmed within the confidence internal.
Acknowledgement This research was supported by the Small & Medium Business Administration of Korean government.
927
references
Chung, J.D., Lee, D.Y., 2009. Effect of desiccant isotherm on the performance of desiccant wheel. Int. J. Refrigeration 32, 720e726. Chung, J.D., Lee, D.Y., Yoon, S.M., 2009. Optimization of desiccant wheel speed and area ratio of regeneration to dehumidification according to regeneration temperature. Solar Energy 83, 625e635. Chung, J.D., Lee, D.Y., Yoon, S.M., 2010. Effect of parameters related to channel and desiccant on the size of desiccant rotor. Int. J. Air-Conditioning Refrigeration 18, 201e211. Collier, R.K., Cohen, B.M., 1991. An analytic investigation of methods for improving the performance of desiccant cooling system. ASME J. Solar Energy Sci. Eng. 113, 157e163. Cui, Q., Chen, H., Tao, G., Yao, H., 2005. Performance study of new adsorbent for solid desiccant cooling. Energy 30, 273e279. Dai, Y.J., Wang, H.F., Zhang, H.F., 2001a. Parameter analysis to improve rotary desiccant dehumidification using a mathematical model. Int. J. Therm. Sci. 40, 400e408. Dai, Y.J., Wang, R.Z., Zhang, H.F., Yu, J.D., 2001b. Use of liquid desiccant cooling to improve the performance of vapour compression air conditioning. Appl. Therm. Eng. 21, 1185e1205. Faust S., Guido F., Lee D.Y., 2007. Desiccant element and manufacturing method of the same, Korea Patent 0704235. Halliday, S.P., Beggs, C.B., Sleigh, P.A., 2002. The use of solar desiccant cooling in the UK: a feasibility study. Appl. Therm. Eng. 22, 1327e1338. Jia, C.X., Dai, Y.J., Wu, J.Y., Wang, R.Z., 2006. Experimental comparison of two honeycombed desiccant wheels fabricated with silica gel and composite desiccant material. Energy Convers. Manage. 47, 2523e2534. Kakac, S., Shah, R.K., Aung, W. (Eds.), 1987. Handbook of SinglePhase Convective Heat Transfer. John Wiley & Sons, Inc. Maclaine-Cross, I.L., 1988. Performance of a solar desiccant cooling system. ASHRAE Trans. 94, 1997e2009. Mavroudaki, P., Beggs, C.B., Sleigh, P.A., Halliday, S.P., 2002. The potential for solar powered single-stage desiccant cooling in southern Europe. Appl. Therm. Eng. 22, 1129e1140. Mazzei, P., Minichiello, F., Palma, D., 2002. Desiccant HVAC systems for commercial buildings. Appl. Therm. Eng. 22, 545e560. Park, S.H., 1996. Robust Design and Analysis for Quality Engineering. Chapman & Hall, London, UK. Pesaran A.A., Penney T.R., Czanderna A.W., 1992 Desiccant cooling: state-of-the-art assessment. NREL Report No. TP-2544147, p. 221. Yu, B.F., Hu, Z.B., Liu, M., Yang, H.L., Kong, Q.X., Liu, Y.H., 2009. Review of research on air-conditioning systems and indoor air quality control for human health. Int. J. Refrigeration 32, 3e20. Zheng, W., Worek, W.M., Novesel, D., 1995. Performance optimization of rotary dehumidifiers. J. Solar. Energy Eng. 117, 40e44.