- Email: [email protected]
Optimization of desiccant wheel speed and area ratio of regeneration to dehumidification as a function of regeneration temperature
Available online at www.sciencedirect.com
Solar Energy 83 (2009) 625–635 www.elsevier.com/locate/solener
Optimization of desiccant wheel speed and area ratio of regeneration to dehumidification as a function of regeneration temperature Jae Dong Chung a,*, Dae-Young Lee b, Seok Mann Yoon c a
Department of Mechanical Engineering, Sejong University, 98 Kunja-Dong, Kwangjin-Gu, Seoul 143-747, Republic of Korea b Korea Institute of Science and Technology, Seoul 136-791, Republic of Korea c District Heating Technology Research Institute, Korea District Heating Corporation, Seoul 135-886, Republic of Korea Received 13 November 2007; received in revised form 9 October 2008; accepted 16 October 2008 Available online 4 November 2008 Communicated by: Associate Editor P. Gandhidasan
Abstract Numerical simulation has been conducted for the desiccant wheel, which is the crucial component of a desiccant cooling system. The mathematical model has been validated by comparing with previous experimental data and numerical results. The calculation results are in reasonable agreement with both, experimental and numerical results. As the key operating/design parameters, the wheel speed and the area ratio of regeneration to dehumidification have been examined for a range of regeneration temperatures from 60 °C to 150 °C. Optimization of these parameters is conducted based on the wheel performance evaluated by means of its moisture removal capacity (MRC) which is more appropriate than effectiveness as a performance index of unbalanced flows. Also the effects of the outdoor air temperature and humidity on the optimum design parameters are examined. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Desiccant wheel; Optimization; MRC (moisture removal capacity); Wheel speed; Area ratio of regeneration/dehumidification; Regeneration temperature
1. Introduction The design of heating, ventilating and air-conditioning (HVAC) systems for thermal comfort requires increasing attention, especially matters arising from recent regulations and standards on ventilation (Mazzei et al., 2005). The optimum level of indoor humidity is desired to be reached and maintained to ensure a comfortable and healthy environment. Desiccant cooling systems have advantages in environmental-conscious operation and separate control of 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
*
Corresponding author. Tel.: +82 2 3408 3776; fax: +82 2 3408 4333. E-mail address: [email protected] (J.D. Chung).
0038-092X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.solener.2008.10.011
being able to use low density energy such as natural gas, waste heat and solar energy. This study is motivated for the use of low energy provided by Korea District Heating Corporation in summer season. The US-Department of Energy estimated 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 and Beggs, 2002; Halliday et al., 2002), performance predictions (Dai et al., 2001a; Mazzei et al., 2002), wheel optimization (Maclaine-Cross, 1988; Collier and Cohen, 1991; Zheng et al., 1995; Dai et al., 2001b; Kodama et al., 2001) and development of new materials (Cui et al., 2005; Jia et al., 2006).
626
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
Nomenclature a A b c Cp Dh fm h hm Hsor k L Le m_ MRC Nu P P, Ps r t tp, tr T
channel height (m) area (m2) channel width (m) channel wall thickness (m) specific heat (J kg1 K1) hydraulic diameter (m) mass fraction of desiccant in the wheel convective heat transfer coefficient (W m2 K1) mass transfer coefficient (kg m2 s1) heat of adsorption (J kg1) thermal conductivity (W m1 K1) channel length (m) Lewis number, h/(hmCpa) mass flow rate (kg h1) moisture removal capacity (kg h1) Nusselt number, hDh/ka perimeter of flow channel (m) pressure, saturated pressure (Pa) radial coordinate time (s) time required for the process and the regeneration per one wheel revolution (s/rev) temperature (K)
The wheel is the most crucial component of the desiccant cooling system. Mathematical modeling of the wheel therefore plays an important role in enhancing the overall system performance. The optimum wheel speed and thickness, and operating parameters such as air flow rate, relative humidity of inlet air and regeneration air temperature on the wheel performance have all been examined (Dai et al., 2001b; Zheng and Worek, 1993; Zhang and Niu, 2002; Zhang et al., 2003; Ahmed et al., 2005). However, most of the research has studied balanced flow, i.e. the wheel is split equally between the process and regeneration air flows. It is commonly accepted that as the regeneration temperature decreases the regeneration section becomes a larger portion of the wheel. According to the manufacturer’s catalog, the 1:3 split between regeneration and dehumidification is generally used at high regeneration temperatures and a 1:1 split used for low regeneration temperatures. However, it is doubtful that each area ratio effectively covers such a broad temperature range. There have been few studies of this issue. In this study, a numerical model is used to study and discuss the performance of a desiccant cooling system in terms of its moisture removal capacity (MRC) considering operating and design parameters such as the area ratio of regeneration to dehumidification, wheel speed and regeneration temperature. Also variations of the outdoor air temperature and humidity effects on the optimum design parameters are examined.
u Y z W Wmax a ed / h q
velocity (m s1) humidity ratio (kg kg1) axial coordinate (m) water content of the desiccant material (kg kg1) maximum water content (kg kg1) dimensionless channel area ration, b/a wheel effectiveness relative humidity angular coordinate density (kg m3)
Subscripts a air in inlet H constant heat flux l liquid out outlet p process r regeneration T constant temperature v vapor w desiccant
2. Analysis 2.1. Model description Fig. 1 shows a typical desiccant cooling system compared with conventional air-conditioning systems using vapor compression refrigeration. In the conventional system air must be dehumidified by cooling it below its dew point to meet the latent load (s ? Ó) and reheating is often required (Ó ? v) to satisfy the sensible heat factor
Fig. 1. Psychrometric processes of desiccant cooling system and conventional air-conditioning system using vapor compression refrigeration.
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
(SHF) needed. This implies very poor energy efficiency, especially for low SHF, i.e. high latent cooling load. Also, a very low temperature for the supplied air can create situations of draft in the air conditioned space. In the desiccant cooling system, the heated and dehumidified supply air exits from the humidification section of the rotor (s ? t). Afterwards the hot and dry air is cooled by a sensible heat exchanger (t ? u). Finally, this flow is subjected to evaporative cooling to get cold and humid air (u ? v) to be supplied to the conditioned space. The sensible heat exchanger acts as a pre-cooler after the desiccant and also as a preheater before the regeneration section, which results in enhanced performance of the whole system. The desiccant wheel is a rotating cylindrical wheel divided into two sections, the adsorption section and the regeneration section. The wheel rotates slowly to expose one portion to the process air stream while the other portion simultaneously passes through the regeneration air stream. The wheel consists of numerous flow channels. Moist air enters the process side and passes over the desiccant and is dehumidified (s ? t). Regeneration occurs on the other side of the partition where heated air enters, usually from the opposite direction, then passes over the desiccant and finally exhausts from the dehumidifier 10 ? 11 ). The ideal outlets of process and regeneration ( are the points of intersection between lines of constant relative humidity and enthalpy passing through the inlets of 10 ), respectively. A detailed process (s) and regeneration ( description of the desiccant wheel can be found in Daou et al. (2006). 2.2. Numerical simulation Assessing the great number of available options and their optimum combinations involved in the design of a desiccant wheel is a time-intensive task if using an experimental approach. Thus, modeling and numerical simulation can become highly effective tools in designing a desiccant wheel by effectively isolating one variable at a time and examining trends and causes. Fig. 2 shows the schematics of a desiccant wheel. Because of geometric similarity and to avoid prohibitive computation costs, it is reasonable to represent the multiple annular layers of straight slots in the desiccant wheel by a ‘‘representative annulus” whose cross sectional view is presented in Fig. 2. In this way, the three cylindrical coordinates of (r, h, z) can reasonably be reduced to a steady two-dimensional (h, z) or unsteady one-dimensional ( t, z) problem. In this study, the unsteady onedimensional model (t, z) is chosen for the coupled heat and mass transfer process in the rotary desiccant wheel. The numerical analysis is based on the following assumptions: (1) The air flow is one-dimensional. (2) The axial heat conduction and mass diffusion in the fluid are neglected.
627
Fig. 2. Schematics of desiccant wheel and computational domains (Zhang et al., 2003).
(3) There is no leakage of fluid in the desiccant wheel. (4) All ducts are impermeable and adiabatic. (5) The thermodynamic properties are constant and uniform. (6) The heat and mass transfer coefficient between the air flow and the desiccant wall is constant along the channel. Based on the above assumptions, the energy and mass conservation equations can be obtained as follows (Zheng and Worek, 1993; Sphaier and Worek, 2004). Mass conservation for the process air oY a hm P p ¼ ðY w Y a Þ oz ua qa Ap
ð1Þ
where left-hand and right-hand sides mean sorbate influx by fluid flow and sorbate transfer rate to felt, respectively. Energy conservation for the process air ðC pa þ Y a C pv Þ
oT a hP p ¼ ðT w T a Þ oz ua qa Ap
ð2Þ
where left-hand side means sum of energy transferred by fluid flow and decreased by sorbate transfer to felt and right-hand side means conduction heat transfer to felt. Conservation of water content for the absorbent oW hm P w ¼ ðY a Y w Þ qw fm Aw ot
ð3Þ
Conservation of energy for the absorbent ðC pw þ fm WC pl Þ
oT a hP w hm H sor P w ¼ ðT a T w Þ þ ðY a Y w Þ ot qw Aw qw Aw ð4Þ
628
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
where the first term of left-hand side is energy transfer by heat conduction and the second term is energy transfer by mass transfer. The above equations are subject to the following boundary and initial conditions which are easily obtained considering periodic nature of desiccant wheel. For the process section T ap ðt; 0Þ ¼ T ap;in Y ap ðt; 0Þ ¼ Y ap;in
ð5aÞ ð5bÞ
T wp ð0; zÞ ¼ T wr ðtr ; L zÞ
ð5cÞ
Y wp ð0; zÞ ¼ Y wr ðtr ; L zÞ
ð5dÞ
For the regeneration section T ar ðt; 0Þ ¼ T ar;in
ð6aÞ
Y ar ðt; 0Þ ¼ Y ar;in T wr ð0; zÞ ¼ T wp ðtp ; L zÞ
ð6bÞ ð6cÞ
Y wr ð0; zÞ ¼ Y wp ðtp ; L zÞ
ð6dÞ
The governing Eqs. (1)–(4) have five unknowns T a , T w , Y a , Y w , and W . In order to solve this set of equations, i.e. to close the problem, it is necessary to relate the equilibrium composition Y w to the water content W and temperature of the adsorbent T w . Here we employ silica gel as a desiccant material, so the water content in the desiccant is governed by the following isotherm (Pesaran and Mills, 1984; Harshe et al., 2005) /w ¼ 0:0078 0:0576W þ 24:2W 2 124W 3 þ 204W 4
ð7Þ
where /w is the equilibrium relative humidity over the desiccant with water content W . The relationship between the humidity ratio and the relative humidity is expressed as Yw ¼
0:622/w P s P /w P
Table 1 Input data used in simulations. Channel shape height, a width, b wall thickness, c Rotor length facing area facing air velocity, ua Desiccant material mass fraction of sorbent, fm capacity, C pw density, qw Air density, qa capacity, C pa thermal conductivity, k a Water vapor capacity, C pv liquid capacity, C pl
Sinusoidal 1.75 10-3 m 3.5 10-3 m 0.15 10-3 m 0.2 m 1 m2 2.0 m/s Silica gel 0.7 921 J/kg K 720 kg/m3 1.1614 kg/m3 1007 J/kg K 0.0263 W/m K 1872 J/kg K 4186 J/kg K
Dh =b ¼ ð1:0542 0:4660a 0:1180a2 þ 0:1794a3 0:0436a4 Þa
ð9Þ
NuT ¼ 1:1791 ð1 þ 2:7701a 3:1901a
2
1:9975a3 0:4966a4 Þ
NuH ¼ 1:903 ð1 þ 0:4556a þ 1:2111a
ð10Þ 2
1:6805a3 þ 0:7724a4 0:1228a5 Þ Nu ¼ ðNuT þ NuH Þ=2
ð11Þ ð12Þ
where a ¼ a=b. The mass transfer coefficient is obtained on the assumption of Le = 1. Table 1 includes the data for all properties and geometries employed in the simulations.
ð8Þ
The aforementioned governing equations are discretized into finite difference equations by implicit, forward difference scheme. All the simulations in this paper are performed with a grid of 800 for time and 50 for space. The grid independence has been proved to be valid within a tolerable limit. The iteration is terminated when the error 2 index D/ < 0:00001, where D/ ¼ f½ðT a T old a Þ=T a þ 2 2 2 1=2 old old old ½ðT w T w Þ=T w þ½ðY a Y a Þ=Y a þ½ðY w Y w Þ= Y w g . It is found that the termination condition is achieved after 10 iterations. The relative error of the uptake mass with respect to the release mass is less than 0.5%. Simulations have been conducted for a desiccant wheel of width 0.2 m with silica gel wall thickness 0.15 mm. The geometry of the channels in the wheel shown in Fig. 2 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 periods. The convective heat transfer coefficient and hydraulic diameter are calculated from the Nusselt number in the sinusoidal shaped channels (Kakac et al., 1987):
Fig. 3. Validation by experimental data and previous numerical results.
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
2.3. Performance index The primary performance indicator is assessed by the moisture removal capacity (MRC), as described in the ASHRAE proposed national standard method of test (ASHRAE, 1998). The MRC is calculated using the following equation: MRC ¼ m_ p ðY ap;in Y ap;out Þ
ð13Þ
If the humidity is fixed at the process and regeneration inlets, the behavior of the MRC is usually similar to that of the effectiveness defined in Eq. (14) which is used in other literature (Zhang and Niu, 2002; Sphaier and Worek, 2004). Based on the effectiveness, the best performance is obtained at the lowest adsorption-side outlet humidity, Y ap;out : ed ¼
Y ap:in Y ap;out Y ap;in
629
profiles from the current formulation are compared with experimental data (Brillhart, 1997) and published numerical results (Sphaier and Worek, 2004) in Fig. 3. Detailed simulation conditions are presented in Sphaier and Worek (2004). Our simulation results are in good agreement with the results of Sphaier and Worek and are very close to the experimental data. The agreement with Sphaier and Worek (2004) whose work is based on a two-dimensional model also gives confidence in the one-dimensional assumption used in present model. In the initial process
ð14Þ
But in the case of an unbalanced flow, the mass flow rate m_ p will change according to the ratio of the area of regeneration to dehumidification, Ar =Ap . Thus the MRC is more appropriate than ed as a performance index of unbalanced flows. At the optimum rotational speed, the MRC reaches a maximum but this does not mean a minimum value of Y ap;out since the mass of moisture removed can become larger not only by decreasing outlet humidity of the process section but also by increasing the area of the process section. This is explained in Section 3.2. 3. Discussion 3.1. Model validation
Fig. 5. Variations of adsorption-side outlet humidity for various wheel speeds at each regeneration temperature, 60 °C, 90 °C, 120 °C and 150 °C.
In order to check the validity and resolvability of the present numerical approach, temperature and humidity
Fig. 4. Profiles of the air humidity along the channel in the dehumidification process.
Fig. 6. Variations of the MRC for various wheel speeds at each regeneration temperature, 60 °C, 90 °C, 120 °C and 150 °C.
630
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
Fig. 7. Variations of the optimum wheel speed as a function of regeneration temperature and outdoor conditions.
Fig. 8. Variations of the optimum value of Ar =Ap as a function of regeneration temperature and outdoor conditions.
Fig. 9. Profiles of the MRC as a function of the optimum wheel speed and Ar =Ap for outdoor conditions T p;in ¼ 30 C and /p;in ¼ 40%; (a) T r;in ¼ 60 C, (b) T r;in ¼ 90 C, (c) T r;in ¼ 120 C and (d) T r;in ¼ 150 C.
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
region, however, considerable discrepancies with the experimental data are found, as also observed by Sphaier and Worek (2004). This is explained by the fact that experimental uncertainties are always greatest in the beginning of an adsorption period (Sphaier and Worek, 2004). Fig. 4 shows the profiles of the air humidity along the channel in the dehumidification process after t = 0 s, 0.8 s, 28.5 s, 57 s, and 82.8 s, respectively. When the air channel just enters the dehumidification zone from the regeneration zone, the process air may be humidified instead of being dehumidified. The reason for this can be explained by the fact that the equilibrium water vapor pressure at the desiccant surface is high, hence the adsorbent cannot adsorb water vapor effectively since the adsorbent at high temperature rotates out of the regeneration section at the beginning stage of process section. Afterwards the dehumidification capacity of the desiccant wheel increases gradually until it reaches its maximum, and then reduces gradually. This qualitative behavior is commonly found in desiccant wheels, supporting the validity of our present approach.
631
3.2. Optimum wheel speed Kang and Maclaine-Cross (1989) showed that the desiccant wheel is the key component of a desiccant cooling system and the cooling coefficient of performance of the system can be significantly enhanced by improving the performance of this component. Optimum rotation speed is one of the most important factors that can improve the wheel performance. The rotation speed should be low enough for complete regeneration but also high enough to keep the adsorbent far from equilibrium. This conflict yields the optimum rotation speed. The existence of an optimum rotation speed, at which the humidity of the product air becomes minimized, has already been reported (Zheng and Worek, 1993; Zhang and Niu, 2002). However, no further analysis has been carried out of its dependence on the regeneration temperature and Ar =Ap . For outdoor conditions when T p;in ¼ 30 C and /p;in ¼ 40%, Fig. 5 shows the variation of the adsorption-side outlet humidity for various wheel speeds (=3600/(tp + tr ) rph) for each regeneration temperature, 60 °C, 90 °C, 120 °C and
Fig. 10. Profiles of MRC as a function of the optimum wheel speed and Ar =Ap for outdoor conditions T p;in ¼ 30 C and /p;in ¼ 55%; (a) T r;in ¼ 60 C, (b) T r;in ¼ 90 C, (c) T r;in ¼ 120 C and (d) T r;in ¼ 150 C.
632
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
150 °C. The solid lines show the results for Ar =Ap ¼ 0.784 and the dashed lines are for the results at the optimum values of Ar =Ap (=0.784, 0.623, 0.521 and 0.424) corresponding to the regeneration temperatures 60 °C, 90 °C, 120 °C and 150 °C, respectively. Note that the optimum value of Ar =Ap is obtained by maximizing the MRC. If the wheel diameter is fixed, the mass of moisture removed can be increased not only by decreasing the outlet humidity of the process section but also by increasing the process section area. Thus we cannot say that the optimum performance is found at the minimum value of the outlet humidity of the process section if the value of Ar =Ap is the design parameter. Fig. 6 clearly shows this. If the MRC is chosen as an indicator of wheel performance, the maximum mass of moisture removed is found at the optimum value of Ar =Ap but in Fig. 5, the minimum humidity in found at other values of Ar =Ap . The variation of optimum wheel speed as a function of regeneration temperature and outdoor conditions is shown in Fig. 7. For convenience, four outdoor conditions are nominated. Regardless of the outdoor conditions, as the regeneration temperature becomes higher the optimum wheel speed decreases and then approaches a constant
value. At low regeneration temperatures, the optimum speed is not so sensitive to the outdoor conditions as compared with high regeneration temperatures. Also we find that there is a dependence of the optimum wheel speed on the outdoor humidity. The highest optimum wheel speed is observed for outdoor conditions T p;in ¼ 30 C and /p;in ¼ 40% for which the humidity has the smallest value (10.6 g/kga) in our studies. The optimum wheel speeds are almost same for the outdoor conditions T p;in ¼ 30 C, /p;in ¼ 55% and T p;in ¼ 35 C, /p;in ¼ 40%. Note that the humidities of these two outdoor conditions are very close (14.6 and 14.1, respectively). The humidity 19.6 g/kgs corresponding to the outdoor condition T p;in ¼ 35 C, /p;in ¼ 40% shows the lowest optimum wheel speed. For the optimization procedure to maximize the MRC, the IMSL routine ZXMWD is linked to the program for evaluating the wheel performance. 3.3. Optimum area ratio of regeneration to dehumidification section It is generally accepted that as the regeneration temperature decreases the regeneration section becomes a larger
Fig. 11. Profiles of MRC as a function of the optimum wheel speed and Ar =Ap for outdoor conditions T p;in ¼ 35 C and /p;in ¼ 40%; (a) T r;in ¼ 60 C, (b) T r;in ¼ 90 C, (c) T r;in ¼ 120 C and (d) T r;in ¼ 150 C.
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
fraction of the wheel. According to the manufacturer’s catalog, the 1:3 split between regeneration and dehumidification is generally used for high regeneration temperatures and a 1:1 split for low regeneration temperatures. However, it is doubtful that each area ratio effectively covers such wide temperature ranges. The effect of the value of Ar =Ap on the wheel performance is investigated at different regeneration temperatures. Variations of the optimum value of Ar =Ap as a function of the regeneration temperature are shown in Fig. 8 for four outdoor conditions. We have determined the dependency of Ar =Ap on the outdoor humidity. As the humidity increases the optimum value of Ar =Ap becomes larger. The results of Ahmed et al. (2005) also show this dependency but they did not provide any discussion of this issue. The lowest value of Ar =Ap is observed for the outdoor condition of T p;in ¼ 30 C and /p;in ¼ 40% for which the humidity is 10.6 g/kga. The humidities for the outdoor conditions T p;in ¼ 30 C, /p;in ¼ 55% and T p;in ¼ 35 C, /p;in ¼ 40% are almost the same and show similar optimum values of Ar =Ap over the entire range of regeneration temperature. The outdoor condition with the highest humidity,
633
i.e. T p;in ¼ 35 C, /p;in ¼ 40% shows the largest optimum value of Ar =Ap . Ahmed et al. (2005) ascertained that the optimum values of Ar =Ap are between 0.7 and 0.9 at T r;in ¼ 60 C and between 0.25 and 0.4 at T r;in ¼ 90 C depending on the inlet humidity. In Fig. 8, the optimum values of Ar =Ap are between 0.78 and 0.90 at T r;in ¼ 60 C for the inlet conditions studied; these are close to those of Ahmed et al. (2005). However, the optimum values of Ar =Ap are between 0.62 and 0.79 for T r;in ¼ 90 C, which shows considerable deviation from Ahmed et al. However, considering that Ar =Ap ¼ 0:33 is usually used for regeneration temperatures much higher than 90 °C and the regeneration section becomes a larger portion of the wheel as the regeneration temperature decreases, the results of Ahmed et al. (2005) seem to under-predict the value of Ar =Ap . Note that in our study the optimum value of Ar =Ap for T r;in ¼ 150 C is found to be 0.42 for outdoor conditions T p;in ¼ 30 C and /p;in ¼ 40%, which is slightly greater than the value from the manufacturer’s catalog (0.33). Figs. 9–12 show the profiles of the MRC as a function of the optimum wheel speed and Ar =Ap for regeneration
Fig. 12. Profiles of MRC as a function of the optimum wheel speed and Ar =Ap for outdoor conditions T p;in ¼ 35 C and /p;in ¼ 55%; (a) T r;in ¼ 60 C, (b) T r;in ¼ 90 C, (c) T r;in ¼ 120 C and (d) T r;in ¼ 150 C.
634
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
temperatures of 60 °C, 90 °C, 120 °C, and 150 °C and four different outdoor conditions. The optimum behavior, i.e. maximum MRC, wheel speed and optimum Ar =Ap is discussed in previous sections and here the behavior of offdesign points is shown. With the sacrifice of a slight decrease of the MRC, the design point can be determined in the range of wheel speed and Ar =Ap . We can also see that the desiccant wheel is highly effective for dehumidification if the regeneration temperature becomes high (Zhang et al., 2003). 3.4. Sensitivity to outdoor weather condition Figs. 13–16 represent the effect of outdoor conditions (20 °C < T < 35 °C, 0.005 < Y < 0.025) on the MRC for regeneration temperatures of 60 °C, 90 °C, 120 °C and 150 °C by means of psychrometric charts. The performance decreases as the outdoor temperature increases or the outdoor humidity decreases. As the regeneration temperature becomes higher, the effect of the outdoor temperature
Fig. 15. Psychrometric chart representation of the effect of outdoor conditions on the MRC for T r;in ¼ 120 C.
Fig. 16. Psychrometric chart representation of the effect of outdoor conditions on the MRC for T r;in ¼ 150 C. Fig. 13. Psychrometric chart representation of the effect of outdoor conditions on the MRC for T r;in ¼ 60 C.
weakens. In particular, for low values of the outdoor humidity and a high regeneration temperature, the wheel performance is nearly independent of the outdoor temperature. 4. Conclusions
Fig. 14. Psychrometric chart representation of the effect of outdoor conditions on the MRC for T r;in ¼ 90 C.
Unsteady one-dimensional numerical simulations have been carried out for the desiccant wheel, which is the most crucial component of the desiccant cooling system. The mathematical model has been validated with experimental data and previous numerical results; our calculations are in reasonable agreement with both of these. As the key operating and design parameters, the wheel speed and Ar =Ap have been examined for a range of regeneration temperatures from 60 °C to 150 °C. The optimum rotation speed is determined from the condition that the moisture removal capacity (MRC) is maximized. Regard-
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
less of the outdoor conditions, as the regeneration temperature becomes higher the optimum wheel speed decreases and then approaches a constant value. The lower the outdoor humidity, the higher is the optimum wheel speed. However, at low regeneration temperatures, the optimum speed is not particularly sensitive to outdoor conditions compared with observations at high regeneration temperatures. The optimum value of Ar =Ap at a low regeneration temperature shows close agreement with other simulations and the value from the manufacturer’s catalog. At high regeneration temperatures there is considerable over-prediction of the optimum Ar =Ap compared with other simulations although it is in reasonable agreement with the manufacturer’s catalog. The effect of outdoor condition on the optimum value of Ar =Ap shows that as the outdoor humidity increases the value of Ar =Ap becomes larger. Also the wheel performance decreases as the outdoor temperature increases and the humidity decreases. However, as the regeneration temperature increases, the effect of the outdoor temperature weakens. Acknowledgement This research was supported by a grant from Korea District Heating Corporation. References Ahmed, M.H., Kattab, N.M., Fouad, M., 2005. Evaluation and optimization of solar desiccant wheel performance. Renewable Energy 30, 305–325. ASHRAE, 1998. Standard 139-1998. Methods of testing for rating desiccant dehumidifiers utilizing heat for the regeneration process, Atlanta. ASHARAE Inc. Brillhart, P.L., 1997. Evaluation of Desiccant Rotor Matrices using an Advanced Fixed-bed Test System. Ph.D. Thesis. University of Illinois at Chicago. Collier, R.K., Cohen, B.M., 1991. An analytic investigation of methods for improving the performance of desiccant cooling system. ASME Journal of Solar Energy Science and Engineering 113, 157–163. Cui, Q., Chen, H., Tao, G., Yao, H., 2005. Performance study of new adsorbent for solid desiccant cooling. Energy 30, 273–279. Dai, Y.J., Wang, R.Z., Zhang, H.F., Yu, J.D., 2001a. Use of liquid desiccant cooling to improve the performance of vapour compression air conditioning. Applied Thermal Engineering 21, 1185–1205.
635
Dai, Y.J., Wang, H.F., Zhang, H.F., 2001b. Parameter analysis to improve rotary desiccant dehumidification using a mathematical model. International Journal of Thermal Science 40, 400–408. Daou, K., Wang, R.Z., Xia, Z.Z., 2006. Desiccant cooling air conditioning: a review. Renewable and Sustainable Energy Reviews 10, 55–77. Halliday, S.P., Beggs, C.B., Sleigh, P.A., 2002. The use of solar desiccant cooling in the UK: a feasibility study. Applied Thermal Engineering 22, 1327–1338. Harshe, Y.M., Utikar, R.P., Ranade, V.V., Pahwa, D., 2005. Modeling of rotary desiccant wheels. Chemical Engineering & Technology 28, 1473–1479. 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 Conversion and Management 47, 2523–2534. Kakac, S., Shah, R.K., Aung, W. (Eds.), 1987. Handbook of Single-phase Convective Heat Transfer. John Wiley and Sons Inc. Kang, T.S., Maclaine-Cross, I.L., 1989. High performance, solid desiccant, open cooling cycles. Journal of Solar Energy Engineering 111, 176–183. Kodama, A., Hirayama, T., Goto, M., Hirose, T., Critoph, R.E., 2001. The use of psychrometric charts for the optimization of a thermal swing desiccant wheel. Applied Thermal Engineering 21, 1657–1674. Maclaine-Cross, I.L., 1988. Performance of a solar desiccant cooling system. ASHRAE Transaction 94, 1997–2009. Mavroudaki, P., Beggs, C.B., 2002. Sleigh PA, Halliday SP. The potential for solar powered single-stage desiccant cooling in southern Europe. Applied Thermal Engineering 22, 1129–1140. Mazzei, P., Minichiello, F., Palma, D., 2002. Desiccant HVAC systems for commercial buildings. Applied Thermal Engineering 22, 545–560. Mazzei, P., Minichiello, F., Palma, D., 2005. HVAC dehumidification systems for thermal comfort: a critical review. Applied Thermal Engineering 25, 677–707. Pesaran, A.A., Mills, A., 1984. Modeling of solid-side transfer in desiccant particle beds. In: Proceedings of the 6th Annual ASME Solar Energy Division Conference. Solar Engineering, New York, USA, pp. 177–185. Pesaran, A.A., Penney, T.R., Czandern, A.W., 1992. Desiccant Cooling: State-of-the-art Assessment. NREL Report No. TP-254-4147, p. 221. Sphaier, L.A., Worek, W.M., 2004. Analysis of heat and mass transfer in porous sorbents used in rotary regenerators. International Journal of Heat and Mass Transfer 47, 3415–3430. Zhang, L.Z., Niu, J.L., 2002. Performance comparisons of desiccant wheels for air dehumidification and enthalpy recovery. Applied Thermal Engineering 22, 1347–1367. Zhang, X.J., Dai, Y.J., Wang, R.Z., 2003. A simulation study of heat and mass transfer in a honeycombed rotary desiccant dehumidifier. Applied Thermal Engineering 23, 989–1003. Zheng, W., Worek, W.M., 1993. Numerical simulation of combined heat and mass transfer process in a rotary dehumidifier. Numerical Heat Transfer Part A 23, 211–232. Zheng, W., Worek, W.M., Novesel, D., 1995. Performance optimization of rotary dehumidifiers. Journal of Solar Energy Engineering 117, 40–44.
Solar Energy 83 (2009) 625–635 www.elsevier.com/locate/solener
Optimization of desiccant wheel speed and area ratio of regeneration to dehumidification as a function of regeneration temperature Jae Dong Chung a,*, Dae-Young Lee b, Seok Mann Yoon c a
Department of Mechanical Engineering, Sejong University, 98 Kunja-Dong, Kwangjin-Gu, Seoul 143-747, Republic of Korea b Korea Institute of Science and Technology, Seoul 136-791, Republic of Korea c District Heating Technology Research Institute, Korea District Heating Corporation, Seoul 135-886, Republic of Korea Received 13 November 2007; received in revised form 9 October 2008; accepted 16 October 2008 Available online 4 November 2008 Communicated by: Associate Editor P. Gandhidasan
Abstract Numerical simulation has been conducted for the desiccant wheel, which is the crucial component of a desiccant cooling system. The mathematical model has been validated by comparing with previous experimental data and numerical results. The calculation results are in reasonable agreement with both, experimental and numerical results. As the key operating/design parameters, the wheel speed and the area ratio of regeneration to dehumidification have been examined for a range of regeneration temperatures from 60 °C to 150 °C. Optimization of these parameters is conducted based on the wheel performance evaluated by means of its moisture removal capacity (MRC) which is more appropriate than effectiveness as a performance index of unbalanced flows. Also the effects of the outdoor air temperature and humidity on the optimum design parameters are examined. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Desiccant wheel; Optimization; MRC (moisture removal capacity); Wheel speed; Area ratio of regeneration/dehumidification; Regeneration temperature
1. Introduction The design of heating, ventilating and air-conditioning (HVAC) systems for thermal comfort requires increasing attention, especially matters arising from recent regulations and standards on ventilation (Mazzei et al., 2005). The optimum level of indoor humidity is desired to be reached and maintained to ensure a comfortable and healthy environment. Desiccant cooling systems have advantages in environmental-conscious operation and separate control of 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
*
Corresponding author. Tel.: +82 2 3408 3776; fax: +82 2 3408 4333. E-mail address: [email protected] (J.D. Chung).
0038-092X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.solener.2008.10.011
being able to use low density energy such as natural gas, waste heat and solar energy. This study is motivated for the use of low energy provided by Korea District Heating Corporation in summer season. The US-Department of Energy estimated 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 and Beggs, 2002; Halliday et al., 2002), performance predictions (Dai et al., 2001a; Mazzei et al., 2002), wheel optimization (Maclaine-Cross, 1988; Collier and Cohen, 1991; Zheng et al., 1995; Dai et al., 2001b; Kodama et al., 2001) and development of new materials (Cui et al., 2005; Jia et al., 2006).
626
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
Nomenclature a A b c Cp Dh fm h hm Hsor k L Le m_ MRC Nu P P, Ps r t tp, tr T
channel height (m) area (m2) channel width (m) channel wall thickness (m) specific heat (J kg1 K1) hydraulic diameter (m) mass fraction of desiccant in the wheel convective heat transfer coefficient (W m2 K1) mass transfer coefficient (kg m2 s1) heat of adsorption (J kg1) thermal conductivity (W m1 K1) channel length (m) Lewis number, h/(hmCpa) mass flow rate (kg h1) moisture removal capacity (kg h1) Nusselt number, hDh/ka perimeter of flow channel (m) pressure, saturated pressure (Pa) radial coordinate time (s) time required for the process and the regeneration per one wheel revolution (s/rev) temperature (K)
The wheel is the most crucial component of the desiccant cooling system. Mathematical modeling of the wheel therefore plays an important role in enhancing the overall system performance. The optimum wheel speed and thickness, and operating parameters such as air flow rate, relative humidity of inlet air and regeneration air temperature on the wheel performance have all been examined (Dai et al., 2001b; Zheng and Worek, 1993; Zhang and Niu, 2002; Zhang et al., 2003; Ahmed et al., 2005). However, most of the research has studied balanced flow, i.e. the wheel is split equally between the process and regeneration air flows. It is commonly accepted that as the regeneration temperature decreases the regeneration section becomes a larger portion of the wheel. According to the manufacturer’s catalog, the 1:3 split between regeneration and dehumidification is generally used at high regeneration temperatures and a 1:1 split used for low regeneration temperatures. However, it is doubtful that each area ratio effectively covers such a broad temperature range. There have been few studies of this issue. In this study, a numerical model is used to study and discuss the performance of a desiccant cooling system in terms of its moisture removal capacity (MRC) considering operating and design parameters such as the area ratio of regeneration to dehumidification, wheel speed and regeneration temperature. Also variations of the outdoor air temperature and humidity effects on the optimum design parameters are examined.
u Y z W Wmax a ed / h q
velocity (m s1) humidity ratio (kg kg1) axial coordinate (m) water content of the desiccant material (kg kg1) maximum water content (kg kg1) dimensionless channel area ration, b/a wheel effectiveness relative humidity angular coordinate density (kg m3)
Subscripts a air in inlet H constant heat flux l liquid out outlet p process r regeneration T constant temperature v vapor w desiccant
2. Analysis 2.1. Model description Fig. 1 shows a typical desiccant cooling system compared with conventional air-conditioning systems using vapor compression refrigeration. In the conventional system air must be dehumidified by cooling it below its dew point to meet the latent load (s ? Ó) and reheating is often required (Ó ? v) to satisfy the sensible heat factor
Fig. 1. Psychrometric processes of desiccant cooling system and conventional air-conditioning system using vapor compression refrigeration.
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
(SHF) needed. This implies very poor energy efficiency, especially for low SHF, i.e. high latent cooling load. Also, a very low temperature for the supplied air can create situations of draft in the air conditioned space. In the desiccant cooling system, the heated and dehumidified supply air exits from the humidification section of the rotor (s ? t). Afterwards the hot and dry air is cooled by a sensible heat exchanger (t ? u). Finally, this flow is subjected to evaporative cooling to get cold and humid air (u ? v) to be supplied to the conditioned space. The sensible heat exchanger acts as a pre-cooler after the desiccant and also as a preheater before the regeneration section, which results in enhanced performance of the whole system. The desiccant wheel is a rotating cylindrical wheel divided into two sections, the adsorption section and the regeneration section. The wheel rotates slowly to expose one portion to the process air stream while the other portion simultaneously passes through the regeneration air stream. The wheel consists of numerous flow channels. Moist air enters the process side and passes over the desiccant and is dehumidified (s ? t). Regeneration occurs on the other side of the partition where heated air enters, usually from the opposite direction, then passes over the desiccant and finally exhausts from the dehumidifier 10 ? 11 ). The ideal outlets of process and regeneration ( are the points of intersection between lines of constant relative humidity and enthalpy passing through the inlets of 10 ), respectively. A detailed process (s) and regeneration ( description of the desiccant wheel can be found in Daou et al. (2006). 2.2. Numerical simulation Assessing the great number of available options and their optimum combinations involved in the design of a desiccant wheel is a time-intensive task if using an experimental approach. Thus, modeling and numerical simulation can become highly effective tools in designing a desiccant wheel by effectively isolating one variable at a time and examining trends and causes. Fig. 2 shows the schematics of a desiccant wheel. Because of geometric similarity and to avoid prohibitive computation costs, it is reasonable to represent the multiple annular layers of straight slots in the desiccant wheel by a ‘‘representative annulus” whose cross sectional view is presented in Fig. 2. In this way, the three cylindrical coordinates of (r, h, z) can reasonably be reduced to a steady two-dimensional (h, z) or unsteady one-dimensional ( t, z) problem. In this study, the unsteady onedimensional model (t, z) is chosen for the coupled heat and mass transfer process in the rotary desiccant wheel. The numerical analysis is based on the following assumptions: (1) The air flow is one-dimensional. (2) The axial heat conduction and mass diffusion in the fluid are neglected.
627
Fig. 2. Schematics of desiccant wheel and computational domains (Zhang et al., 2003).
(3) There is no leakage of fluid in the desiccant wheel. (4) All ducts are impermeable and adiabatic. (5) The thermodynamic properties are constant and uniform. (6) The heat and mass transfer coefficient between the air flow and the desiccant wall is constant along the channel. Based on the above assumptions, the energy and mass conservation equations can be obtained as follows (Zheng and Worek, 1993; Sphaier and Worek, 2004). Mass conservation for the process air oY a hm P p ¼ ðY w Y a Þ oz ua qa Ap
ð1Þ
where left-hand and right-hand sides mean sorbate influx by fluid flow and sorbate transfer rate to felt, respectively. Energy conservation for the process air ðC pa þ Y a C pv Þ
oT a hP p ¼ ðT w T a Þ oz ua qa Ap
ð2Þ
where left-hand side means sum of energy transferred by fluid flow and decreased by sorbate transfer to felt and right-hand side means conduction heat transfer to felt. Conservation of water content for the absorbent oW hm P w ¼ ðY a Y w Þ qw fm Aw ot
ð3Þ
Conservation of energy for the absorbent ðC pw þ fm WC pl Þ
oT a hP w hm H sor P w ¼ ðT a T w Þ þ ðY a Y w Þ ot qw Aw qw Aw ð4Þ
628
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
where the first term of left-hand side is energy transfer by heat conduction and the second term is energy transfer by mass transfer. The above equations are subject to the following boundary and initial conditions which are easily obtained considering periodic nature of desiccant wheel. For the process section T ap ðt; 0Þ ¼ T ap;in Y ap ðt; 0Þ ¼ Y ap;in
ð5aÞ ð5bÞ
T wp ð0; zÞ ¼ T wr ðtr ; L zÞ
ð5cÞ
Y wp ð0; zÞ ¼ Y wr ðtr ; L zÞ
ð5dÞ
For the regeneration section T ar ðt; 0Þ ¼ T ar;in
ð6aÞ
Y ar ðt; 0Þ ¼ Y ar;in T wr ð0; zÞ ¼ T wp ðtp ; L zÞ
ð6bÞ ð6cÞ
Y wr ð0; zÞ ¼ Y wp ðtp ; L zÞ
ð6dÞ
The governing Eqs. (1)–(4) have five unknowns T a , T w , Y a , Y w , and W . In order to solve this set of equations, i.e. to close the problem, it is necessary to relate the equilibrium composition Y w to the water content W and temperature of the adsorbent T w . Here we employ silica gel as a desiccant material, so the water content in the desiccant is governed by the following isotherm (Pesaran and Mills, 1984; Harshe et al., 2005) /w ¼ 0:0078 0:0576W þ 24:2W 2 124W 3 þ 204W 4
ð7Þ
where /w is the equilibrium relative humidity over the desiccant with water content W . The relationship between the humidity ratio and the relative humidity is expressed as Yw ¼
0:622/w P s P /w P
Table 1 Input data used in simulations. Channel shape height, a width, b wall thickness, c Rotor length facing area facing air velocity, ua Desiccant material mass fraction of sorbent, fm capacity, C pw density, qw Air density, qa capacity, C pa thermal conductivity, k a Water vapor capacity, C pv liquid capacity, C pl
Sinusoidal 1.75 10-3 m 3.5 10-3 m 0.15 10-3 m 0.2 m 1 m2 2.0 m/s Silica gel 0.7 921 J/kg K 720 kg/m3 1.1614 kg/m3 1007 J/kg K 0.0263 W/m K 1872 J/kg K 4186 J/kg K
Dh =b ¼ ð1:0542 0:4660a 0:1180a2 þ 0:1794a3 0:0436a4 Þa
ð9Þ
NuT ¼ 1:1791 ð1 þ 2:7701a 3:1901a
2
1:9975a3 0:4966a4 Þ
NuH ¼ 1:903 ð1 þ 0:4556a þ 1:2111a
ð10Þ 2
1:6805a3 þ 0:7724a4 0:1228a5 Þ Nu ¼ ðNuT þ NuH Þ=2
ð11Þ ð12Þ
where a ¼ a=b. The mass transfer coefficient is obtained on the assumption of Le = 1. Table 1 includes the data for all properties and geometries employed in the simulations.
ð8Þ
The aforementioned governing equations are discretized into finite difference equations by implicit, forward difference scheme. All the simulations in this paper are performed with a grid of 800 for time and 50 for space. The grid independence has been proved to be valid within a tolerable limit. The iteration is terminated when the error 2 index D/ < 0:00001, where D/ ¼ f½ðT a T old a Þ=T a þ 2 2 2 1=2 old old old ½ðT w T w Þ=T w þ½ðY a Y a Þ=Y a þ½ðY w Y w Þ= Y w g . It is found that the termination condition is achieved after 10 iterations. The relative error of the uptake mass with respect to the release mass is less than 0.5%. Simulations have been conducted for a desiccant wheel of width 0.2 m with silica gel wall thickness 0.15 mm. The geometry of the channels in the wheel shown in Fig. 2 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 periods. The convective heat transfer coefficient and hydraulic diameter are calculated from the Nusselt number in the sinusoidal shaped channels (Kakac et al., 1987):
Fig. 3. Validation by experimental data and previous numerical results.
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
2.3. Performance index The primary performance indicator is assessed by the moisture removal capacity (MRC), as described in the ASHRAE proposed national standard method of test (ASHRAE, 1998). The MRC is calculated using the following equation: MRC ¼ m_ p ðY ap;in Y ap;out Þ
ð13Þ
If the humidity is fixed at the process and regeneration inlets, the behavior of the MRC is usually similar to that of the effectiveness defined in Eq. (14) which is used in other literature (Zhang and Niu, 2002; Sphaier and Worek, 2004). Based on the effectiveness, the best performance is obtained at the lowest adsorption-side outlet humidity, Y ap;out : ed ¼
Y ap:in Y ap;out Y ap;in
629
profiles from the current formulation are compared with experimental data (Brillhart, 1997) and published numerical results (Sphaier and Worek, 2004) in Fig. 3. Detailed simulation conditions are presented in Sphaier and Worek (2004). Our simulation results are in good agreement with the results of Sphaier and Worek and are very close to the experimental data. The agreement with Sphaier and Worek (2004) whose work is based on a two-dimensional model also gives confidence in the one-dimensional assumption used in present model. In the initial process
ð14Þ
But in the case of an unbalanced flow, the mass flow rate m_ p will change according to the ratio of the area of regeneration to dehumidification, Ar =Ap . Thus the MRC is more appropriate than ed as a performance index of unbalanced flows. At the optimum rotational speed, the MRC reaches a maximum but this does not mean a minimum value of Y ap;out since the mass of moisture removed can become larger not only by decreasing outlet humidity of the process section but also by increasing the area of the process section. This is explained in Section 3.2. 3. Discussion 3.1. Model validation
Fig. 5. Variations of adsorption-side outlet humidity for various wheel speeds at each regeneration temperature, 60 °C, 90 °C, 120 °C and 150 °C.
In order to check the validity and resolvability of the present numerical approach, temperature and humidity
Fig. 4. Profiles of the air humidity along the channel in the dehumidification process.
Fig. 6. Variations of the MRC for various wheel speeds at each regeneration temperature, 60 °C, 90 °C, 120 °C and 150 °C.
630
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
Fig. 7. Variations of the optimum wheel speed as a function of regeneration temperature and outdoor conditions.
Fig. 8. Variations of the optimum value of Ar =Ap as a function of regeneration temperature and outdoor conditions.
Fig. 9. Profiles of the MRC as a function of the optimum wheel speed and Ar =Ap for outdoor conditions T p;in ¼ 30 C and /p;in ¼ 40%; (a) T r;in ¼ 60 C, (b) T r;in ¼ 90 C, (c) T r;in ¼ 120 C and (d) T r;in ¼ 150 C.
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
region, however, considerable discrepancies with the experimental data are found, as also observed by Sphaier and Worek (2004). This is explained by the fact that experimental uncertainties are always greatest in the beginning of an adsorption period (Sphaier and Worek, 2004). Fig. 4 shows the profiles of the air humidity along the channel in the dehumidification process after t = 0 s, 0.8 s, 28.5 s, 57 s, and 82.8 s, respectively. When the air channel just enters the dehumidification zone from the regeneration zone, the process air may be humidified instead of being dehumidified. The reason for this can be explained by the fact that the equilibrium water vapor pressure at the desiccant surface is high, hence the adsorbent cannot adsorb water vapor effectively since the adsorbent at high temperature rotates out of the regeneration section at the beginning stage of process section. Afterwards the dehumidification capacity of the desiccant wheel increases gradually until it reaches its maximum, and then reduces gradually. This qualitative behavior is commonly found in desiccant wheels, supporting the validity of our present approach.
631
3.2. Optimum wheel speed Kang and Maclaine-Cross (1989) showed that the desiccant wheel is the key component of a desiccant cooling system and the cooling coefficient of performance of the system can be significantly enhanced by improving the performance of this component. Optimum rotation speed is one of the most important factors that can improve the wheel performance. The rotation speed should be low enough for complete regeneration but also high enough to keep the adsorbent far from equilibrium. This conflict yields the optimum rotation speed. The existence of an optimum rotation speed, at which the humidity of the product air becomes minimized, has already been reported (Zheng and Worek, 1993; Zhang and Niu, 2002). However, no further analysis has been carried out of its dependence on the regeneration temperature and Ar =Ap . For outdoor conditions when T p;in ¼ 30 C and /p;in ¼ 40%, Fig. 5 shows the variation of the adsorption-side outlet humidity for various wheel speeds (=3600/(tp + tr ) rph) for each regeneration temperature, 60 °C, 90 °C, 120 °C and
Fig. 10. Profiles of MRC as a function of the optimum wheel speed and Ar =Ap for outdoor conditions T p;in ¼ 30 C and /p;in ¼ 55%; (a) T r;in ¼ 60 C, (b) T r;in ¼ 90 C, (c) T r;in ¼ 120 C and (d) T r;in ¼ 150 C.
632
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
150 °C. The solid lines show the results for Ar =Ap ¼ 0.784 and the dashed lines are for the results at the optimum values of Ar =Ap (=0.784, 0.623, 0.521 and 0.424) corresponding to the regeneration temperatures 60 °C, 90 °C, 120 °C and 150 °C, respectively. Note that the optimum value of Ar =Ap is obtained by maximizing the MRC. If the wheel diameter is fixed, the mass of moisture removed can be increased not only by decreasing the outlet humidity of the process section but also by increasing the process section area. Thus we cannot say that the optimum performance is found at the minimum value of the outlet humidity of the process section if the value of Ar =Ap is the design parameter. Fig. 6 clearly shows this. If the MRC is chosen as an indicator of wheel performance, the maximum mass of moisture removed is found at the optimum value of Ar =Ap but in Fig. 5, the minimum humidity in found at other values of Ar =Ap . The variation of optimum wheel speed as a function of regeneration temperature and outdoor conditions is shown in Fig. 7. For convenience, four outdoor conditions are nominated. Regardless of the outdoor conditions, as the regeneration temperature becomes higher the optimum wheel speed decreases and then approaches a constant
value. At low regeneration temperatures, the optimum speed is not so sensitive to the outdoor conditions as compared with high regeneration temperatures. Also we find that there is a dependence of the optimum wheel speed on the outdoor humidity. The highest optimum wheel speed is observed for outdoor conditions T p;in ¼ 30 C and /p;in ¼ 40% for which the humidity has the smallest value (10.6 g/kga) in our studies. The optimum wheel speeds are almost same for the outdoor conditions T p;in ¼ 30 C, /p;in ¼ 55% and T p;in ¼ 35 C, /p;in ¼ 40%. Note that the humidities of these two outdoor conditions are very close (14.6 and 14.1, respectively). The humidity 19.6 g/kgs corresponding to the outdoor condition T p;in ¼ 35 C, /p;in ¼ 40% shows the lowest optimum wheel speed. For the optimization procedure to maximize the MRC, the IMSL routine ZXMWD is linked to the program for evaluating the wheel performance. 3.3. Optimum area ratio of regeneration to dehumidification section It is generally accepted that as the regeneration temperature decreases the regeneration section becomes a larger
Fig. 11. Profiles of MRC as a function of the optimum wheel speed and Ar =Ap for outdoor conditions T p;in ¼ 35 C and /p;in ¼ 40%; (a) T r;in ¼ 60 C, (b) T r;in ¼ 90 C, (c) T r;in ¼ 120 C and (d) T r;in ¼ 150 C.
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
fraction of the wheel. According to the manufacturer’s catalog, the 1:3 split between regeneration and dehumidification is generally used for high regeneration temperatures and a 1:1 split for low regeneration temperatures. However, it is doubtful that each area ratio effectively covers such wide temperature ranges. The effect of the value of Ar =Ap on the wheel performance is investigated at different regeneration temperatures. Variations of the optimum value of Ar =Ap as a function of the regeneration temperature are shown in Fig. 8 for four outdoor conditions. We have determined the dependency of Ar =Ap on the outdoor humidity. As the humidity increases the optimum value of Ar =Ap becomes larger. The results of Ahmed et al. (2005) also show this dependency but they did not provide any discussion of this issue. The lowest value of Ar =Ap is observed for the outdoor condition of T p;in ¼ 30 C and /p;in ¼ 40% for which the humidity is 10.6 g/kga. The humidities for the outdoor conditions T p;in ¼ 30 C, /p;in ¼ 55% and T p;in ¼ 35 C, /p;in ¼ 40% are almost the same and show similar optimum values of Ar =Ap over the entire range of regeneration temperature. The outdoor condition with the highest humidity,
633
i.e. T p;in ¼ 35 C, /p;in ¼ 40% shows the largest optimum value of Ar =Ap . Ahmed et al. (2005) ascertained that the optimum values of Ar =Ap are between 0.7 and 0.9 at T r;in ¼ 60 C and between 0.25 and 0.4 at T r;in ¼ 90 C depending on the inlet humidity. In Fig. 8, the optimum values of Ar =Ap are between 0.78 and 0.90 at T r;in ¼ 60 C for the inlet conditions studied; these are close to those of Ahmed et al. (2005). However, the optimum values of Ar =Ap are between 0.62 and 0.79 for T r;in ¼ 90 C, which shows considerable deviation from Ahmed et al. However, considering that Ar =Ap ¼ 0:33 is usually used for regeneration temperatures much higher than 90 °C and the regeneration section becomes a larger portion of the wheel as the regeneration temperature decreases, the results of Ahmed et al. (2005) seem to under-predict the value of Ar =Ap . Note that in our study the optimum value of Ar =Ap for T r;in ¼ 150 C is found to be 0.42 for outdoor conditions T p;in ¼ 30 C and /p;in ¼ 40%, which is slightly greater than the value from the manufacturer’s catalog (0.33). Figs. 9–12 show the profiles of the MRC as a function of the optimum wheel speed and Ar =Ap for regeneration
Fig. 12. Profiles of MRC as a function of the optimum wheel speed and Ar =Ap for outdoor conditions T p;in ¼ 35 C and /p;in ¼ 55%; (a) T r;in ¼ 60 C, (b) T r;in ¼ 90 C, (c) T r;in ¼ 120 C and (d) T r;in ¼ 150 C.
634
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
temperatures of 60 °C, 90 °C, 120 °C, and 150 °C and four different outdoor conditions. The optimum behavior, i.e. maximum MRC, wheel speed and optimum Ar =Ap is discussed in previous sections and here the behavior of offdesign points is shown. With the sacrifice of a slight decrease of the MRC, the design point can be determined in the range of wheel speed and Ar =Ap . We can also see that the desiccant wheel is highly effective for dehumidification if the regeneration temperature becomes high (Zhang et al., 2003). 3.4. Sensitivity to outdoor weather condition Figs. 13–16 represent the effect of outdoor conditions (20 °C < T < 35 °C, 0.005 < Y < 0.025) on the MRC for regeneration temperatures of 60 °C, 90 °C, 120 °C and 150 °C by means of psychrometric charts. The performance decreases as the outdoor temperature increases or the outdoor humidity decreases. As the regeneration temperature becomes higher, the effect of the outdoor temperature
Fig. 15. Psychrometric chart representation of the effect of outdoor conditions on the MRC for T r;in ¼ 120 C.
Fig. 16. Psychrometric chart representation of the effect of outdoor conditions on the MRC for T r;in ¼ 150 C. Fig. 13. Psychrometric chart representation of the effect of outdoor conditions on the MRC for T r;in ¼ 60 C.
weakens. In particular, for low values of the outdoor humidity and a high regeneration temperature, the wheel performance is nearly independent of the outdoor temperature. 4. Conclusions
Fig. 14. Psychrometric chart representation of the effect of outdoor conditions on the MRC for T r;in ¼ 90 C.
Unsteady one-dimensional numerical simulations have been carried out for the desiccant wheel, which is the most crucial component of the desiccant cooling system. The mathematical model has been validated with experimental data and previous numerical results; our calculations are in reasonable agreement with both of these. As the key operating and design parameters, the wheel speed and Ar =Ap have been examined for a range of regeneration temperatures from 60 °C to 150 °C. The optimum rotation speed is determined from the condition that the moisture removal capacity (MRC) is maximized. Regard-
J.D. Chung et al. / Solar Energy 83 (2009) 625–635
less of the outdoor conditions, as the regeneration temperature becomes higher the optimum wheel speed decreases and then approaches a constant value. The lower the outdoor humidity, the higher is the optimum wheel speed. However, at low regeneration temperatures, the optimum speed is not particularly sensitive to outdoor conditions compared with observations at high regeneration temperatures. The optimum value of Ar =Ap at a low regeneration temperature shows close agreement with other simulations and the value from the manufacturer’s catalog. At high regeneration temperatures there is considerable over-prediction of the optimum Ar =Ap compared with other simulations although it is in reasonable agreement with the manufacturer’s catalog. The effect of outdoor condition on the optimum value of Ar =Ap shows that as the outdoor humidity increases the value of Ar =Ap becomes larger. Also the wheel performance decreases as the outdoor temperature increases and the humidity decreases. However, as the regeneration temperature increases, the effect of the outdoor temperature weakens. Acknowledgement This research was supported by a grant from Korea District Heating Corporation. References Ahmed, M.H., Kattab, N.M., Fouad, M., 2005. Evaluation and optimization of solar desiccant wheel performance. Renewable Energy 30, 305–325. ASHRAE, 1998. Standard 139-1998. Methods of testing for rating desiccant dehumidifiers utilizing heat for the regeneration process, Atlanta. ASHARAE Inc. Brillhart, P.L., 1997. Evaluation of Desiccant Rotor Matrices using an Advanced Fixed-bed Test System. Ph.D. Thesis. University of Illinois at Chicago. Collier, R.K., Cohen, B.M., 1991. An analytic investigation of methods for improving the performance of desiccant cooling system. ASME Journal of Solar Energy Science and Engineering 113, 157–163. Cui, Q., Chen, H., Tao, G., Yao, H., 2005. Performance study of new adsorbent for solid desiccant cooling. Energy 30, 273–279. Dai, Y.J., Wang, R.Z., Zhang, H.F., Yu, J.D., 2001a. Use of liquid desiccant cooling to improve the performance of vapour compression air conditioning. Applied Thermal Engineering 21, 1185–1205.
635
Dai, Y.J., Wang, H.F., Zhang, H.F., 2001b. Parameter analysis to improve rotary desiccant dehumidification using a mathematical model. International Journal of Thermal Science 40, 400–408. Daou, K., Wang, R.Z., Xia, Z.Z., 2006. Desiccant cooling air conditioning: a review. Renewable and Sustainable Energy Reviews 10, 55–77. Halliday, S.P., Beggs, C.B., Sleigh, P.A., 2002. The use of solar desiccant cooling in the UK: a feasibility study. Applied Thermal Engineering 22, 1327–1338. Harshe, Y.M., Utikar, R.P., Ranade, V.V., Pahwa, D., 2005. Modeling of rotary desiccant wheels. Chemical Engineering & Technology 28, 1473–1479. 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 Conversion and Management 47, 2523–2534. Kakac, S., Shah, R.K., Aung, W. (Eds.), 1987. Handbook of Single-phase Convective Heat Transfer. John Wiley and Sons Inc. Kang, T.S., Maclaine-Cross, I.L., 1989. High performance, solid desiccant, open cooling cycles. Journal of Solar Energy Engineering 111, 176–183. Kodama, A., Hirayama, T., Goto, M., Hirose, T., Critoph, R.E., 2001. The use of psychrometric charts for the optimization of a thermal swing desiccant wheel. Applied Thermal Engineering 21, 1657–1674. Maclaine-Cross, I.L., 1988. Performance of a solar desiccant cooling system. ASHRAE Transaction 94, 1997–2009. Mavroudaki, P., Beggs, C.B., 2002. Sleigh PA, Halliday SP. The potential for solar powered single-stage desiccant cooling in southern Europe. Applied Thermal Engineering 22, 1129–1140. Mazzei, P., Minichiello, F., Palma, D., 2002. Desiccant HVAC systems for commercial buildings. Applied Thermal Engineering 22, 545–560. Mazzei, P., Minichiello, F., Palma, D., 2005. HVAC dehumidification systems for thermal comfort: a critical review. Applied Thermal Engineering 25, 677–707. Pesaran, A.A., Mills, A., 1984. Modeling of solid-side transfer in desiccant particle beds. In: Proceedings of the 6th Annual ASME Solar Energy Division Conference. Solar Engineering, New York, USA, pp. 177–185. Pesaran, A.A., Penney, T.R., Czandern, A.W., 1992. Desiccant Cooling: State-of-the-art Assessment. NREL Report No. TP-254-4147, p. 221. Sphaier, L.A., Worek, W.M., 2004. Analysis of heat and mass transfer in porous sorbents used in rotary regenerators. International Journal of Heat and Mass Transfer 47, 3415–3430. Zhang, L.Z., Niu, J.L., 2002. Performance comparisons of desiccant wheels for air dehumidification and enthalpy recovery. Applied Thermal Engineering 22, 1347–1367. Zhang, X.J., Dai, Y.J., Wang, R.Z., 2003. A simulation study of heat and mass transfer in a honeycombed rotary desiccant dehumidifier. Applied Thermal Engineering 23, 989–1003. Zheng, W., Worek, W.M., 1993. Numerical simulation of combined heat and mass transfer process in a rotary dehumidifier. Numerical Heat Transfer Part A 23, 211–232. Zheng, W., Worek, W.M., Novesel, D., 1995. Performance optimization of rotary dehumidifiers. Journal of Solar Energy Engineering 117, 40–44.