- Email: [email protected]
Simulations of the performance of open cycle desiccant systems using solar energy
Solar Energy VoL 21, pp. ~3-~8
0038-092x/78/1001-0273/$02.00/0
© PergamonPressLtd.,1978. PrintedinGreatBritain
SIMULATIONS OF THE PERFORMANCE OF OPEN CYCLE DESICCANT SYSTEMS USING SOLAR ENERGY J. S. NELSON, W. A. BECKMAN,J. W. MITCHELLand D. J. CLOSE Solar Energy Laboratory Universityof Wisconsin-Madison,Madison,WI 53706, U.S.A. (Received 16 January 1978; revision accepted 15 June 1978)
Abstract- A feasibilitystudy of open cycle air conditioningsystems that use solid desiccants and solar energy has been performed. The two configurationsevaluated are the ventilation mode, in which ambient air is continually introduced into the room, and the recirculationmode, in which room air is recirculated. Seasonal simulationsfor Miami, Florida, show that the auxiliaryenergy requirement for the ventilationmode is about one half that for the recirculation mode. The seasonal COP for the system using solar energy as the auxiliary is approximately0.75. A conventional flat plate solar energy system of moderate size can provide a large fraction of the energy required to meet the sensible and latent loads of a typical house.
describe the operation of a unit a design conditions, but do not evaluate performance over an entire cooling season. This paper describes a computer simulation of the seasonal performance of two operating modes for an open cycle cooling system. In the ventilation mode fresh air is continually introduced into the conditioned space, while in the recirculation mode exhaust air from the conditioned space is reconditioned and reintroduced into the room. The operation is similar to that of the SolarMEC system [2]. A detailed description of the study is given in Ref. [5]. Schematics of the equipment and psychrometric diagrams of the ventilation and recirculation modes are shown in Figs. 1-4. In Figs. 1 and 3, the temperature and humidity are described relative to comfort levels in the
INTRODUCTION
Open cycle desiccant air conditioning systems utilize a desiccant to remove moisture from humid air. Air thus dried is cooled in evaporative coolers and sensible heat exchangers to meet both sensible and latent air conditioning loads. Solar energy or other energy sources are used to regenerate the desiccant. The emphasis in this study is on systems that use solid desiccants. A variety of open cycle systems using solid desiccants and solar energy regeneration have been proposed. Dunkle[l] discusses cooling and dehumidification of air using solar energy and a rotary desiccant wheel. A design point calculation is made to establish that the system can meet the sensible and latent loads of a specific building. Rush et aL [2] report on a test installation in Los Angeles which uses both solar energy and natural gas to regenerate the desiccant wheel. The cooling COP based on total energy supplied was found to be 0.53 at one operating point. Ruder and Rousseau[3] describe a design of a compact system for residential use. A desiccant system is described by Lunde [4] which uses several stages of intercooling in the dehumidification process. Design point calculations are made to establish optimum air flow rates and wheel thickness. These studies
room.
In the ventilation mode, Figs. I and 2, ambient air is dried and heated by the dehumidifier, regeneratively cooled by exhaust air, evaporatively cooled, and introduced into the house. The process can be controlled so that the temperature and humidity of the supply air are lower than that in the house, and this allows the sensible and latent loads to be met. The exhaust air is first
Open Cycle Desiccant Cooling Ventilation Mode WATER
ROOM
Tem~eret,.=re Comfortable HJmJdity: Comfortable
HEAT
l
" ~
War
EXHAUST
Hot
--
Domp
Regenerator WATER
o C Moist ~
~: ~
~
~
AMBIENT
Dehumidifier . ~ W e t Dry -..
Hot Dry
Fig. 1. Schematicof ventilationmode. 273
_
2?4
J. S. NELSOI~et
Ventilation
/
/EXHAUST
Mode
al.
Recirculotlon /
EXHAUST ~'"
Humidity
®/
%
Humidity /
.
IC~AMBIENT
Temperature
Temperature
Fig. 2. Psychrometricdiagramfor ventilationmode,
Fig. 4. Psychrometricdiagram for recirculation mode.
evaporatively cooled to provide a low temperature sink for heat transfer from the supply air in the regenerator. The air is then heated by an energy supply which could be either a conventional fuel, a solar source, or both. Passage of the heated air through the dehumidifier regenerates the desiccant and cools the air. The psychrometric diagram shows the states of the air for typical operating conditions. Ambient conditions change over the course of the day, with resulting changes in the sensible and latent loads and in the psychrometric diagram. The recirculation mode shown in Figs: 3 and 4 employs the same components as the ventilation mode. However, room air is recirculated and ambient air is used only for regeneration. Room air is dehumidified and heated by the desiccant wheel, regeneratively cooled, and then evaporatively cooled prior to reentering the room. Ambient ak is evaporatively cooled, regeneratively heated, and then heated by an energy supply. The heated ambient air passes through the dehumidifier and regenerates the desiccant. For both systems, the sensible heat regenerator and dehumidifier are assumed to be rotary elements, but direct transfer exchangers and fixed beds could equally well be used. The objectives of this study are: 1. To develop computer models for the various components of a desiccant system. 2. To determine the annual thermal performance of the
system and the ability to meet a combination of latent and sensible loads. 3. To evaluate the potential of a solar energy system to provide regeneration of the dehumidifier under typical weather conditions. 4. To develop a control strategy to allow the system to provide comfort under varying ambient conditions.
COMPONENT MODELS
Computer models of the various components were developed for use with the transient simulation program TRNSYS [6]. Evaporative coolers were modeled using the analytical relations developed by Hollands [7]. The major assumption is that the Lewis number equals unity for the air-water vapor mixture. This allows the saturation effectiveness to be formulated in terms of an overall Ntu similar to that for a sensible heat exchanger. The sensible heat regenerator was modeled using the relationships of Kays and London[8]. The development does not account for fluid leakage or carryover from one stream to the other, and assumes a uniform temperature across the inlet air streams. The specific rotary regenerator modeled was one constructed of polypropylene ribbon wound spirally around the rotation axis. The air flow is through the parallel passages between successive layers of the ribbon. This type of regenerator has been shown by Dunlde and Maclaine-Cross[9] to
Open Cycle Desiccant Cooling Recirculotion Mode AMBIENT
/
HEAT
WATER
EXHAUST
I II
ROOM
Temperature: Comfortable
Humidity:
Comfortable ~
°"-LI
:'°u=t';*l I
" °'' U )
Fig. 3. Schematicof recirculation mode.
275
Simulations of the performance of open cycle desiccant systems using solar energy have satisfactory heat exchange and pressure drop properties for air conditioning applications. The model for the rotary dehumidfier is based on an analysis developed in a series of papers by Banks, Close and Maclalne--Cross[10-13]. This analysis allows the performance of a rotary dehumidifier to be calculated by analogy with that for heat transfer in rotary sensible heat regenerators. Characteristic potentials and specific capacity ratios that are analogous to temperature and specific heat ratio are formulated, and nondimensional parameters analogous to effectiveness, Ntu, and capacity rate ratio are formed. Then sensible heat regenerator theory [8] is used to evaluate the inlet and outlet states and the performance. As shown by Banks and MaclaineCross, the process line for a well designed dehumidifier approximates a constant enthalpy process, as indicated in Figs. 2 and 4. The specific dehumidifier modeled was a parallel passage dehumidifier with construction similar to that of the sensible heat regenerator. The desiccant material was taken to be silica gel, with properties as given in Ref. [12]. The simulations were performed using 1955 weather data for Miami, Florida. Miami was chosen as there are high sensible and latent loads which vary significantly over the course of the day. Hourly sensible and latent loads were evaluated using the transfer function approach for a house of 140m2 floor area that was insulated to ASHRAE 90-75 standards [14]. At Miami outdoor design conditions, the house model generates sensible and latent loads of 3800 W and 1000W, respectively. SYSTEM ~
C
E
CHARACTERISTICS
Systems with ideal (100 per cent effectiveness) components were studied in order to establish upper limits on performance and provide insight into system operation. The load was that from the house model for July. The air flow through the system was 0.34kg/s, which was based on accepted practice for vapor compression machines of 400 cfm/ton. The regenerating stream temperature (state point 8 in Fig. 2 and state point 4 in Fig. 4) was held constant. The latent and sensible loads met by the system and the system total and sensible coefficients of performance were evaluated. The coefficient of performance is the load met divided by the energy supplied by the external energy supply. The variation of the ideal coefficient of performance with regenerating stream temperature is shown in Fig. 5. Although these results are valid only for the Miami climate, several general conclusions may be drawn. High values of COP may be obtained at low regenerating stream temperatures. However, the capacity of the system is low even though energy is used effectively. At regenerating stream temperatures above about 80°C, the COP for the ventilation mode decreases significantly. A high regenerating stream temperature produces a very low humidity in the supply stream (state point 2, Fig. 2), but the resulting wet bulb temperature for state 3 is not correspondingly reduced. As a result, only a slightly greater load is met with the greatly increased supply energy.
Vgnfila!ion Mode COP--..,,
20 15
cop 1.0 0.5
60
0 90 I00 Trecjeneratio n (°C)
liO
120
Fig. 5. Coefficient of performance for ideal systems as a function
of regenerating stream temperature, for July 1955,Miami. The ideal COP for the ventilation mode is about twice that of the recirculation mode. The reason for this can be seen from a psychrometric diagram for a given operating condition. Both cycles are represented in Fig. 6 for the same ambient and room states and for the same house latent and sensible loads. The sensible heat regenerator in the ventilation mode is able to increase the temperature of the air entering the energy supply T7 to a higher value than that in the recirculation mode, T3. Correspondingly, in the ventilaiton mode, the dried air stream is regeneratively cooled to a lower temperature, T3, than in the recirculating mode, Ts. To achieve the same inlet state to the room, state 9, the absolute humidity of the dried air in the recirculation mode must be lower than that in the ventilation mode. This requires a higher temperature of the air entering the regenerator in the recirculation mode, T4, than in the ventilation mode, Ts. As a result, the amount of auxiliary energy that must be supplied in the recirculation mode is greater than for the ventilation mode. Further examination of Fig. 6 shows that this relation Ventilation Mode /
/
EXHAUST 9
Hum=dily ~AM BIEN T " ~
~
"I
~ zSupplieO
Temperature
Recirculotion /
T7
// EXHAUST 5~\
,K,
l"emperature
TB
-x. Hom,o'.,
T~
T~
Fig. 6. Comparison of ventilation and recirculafion mode, with ideal components.
2%
J. S. NELSONet
between the two cycles depends on ambient conditions. For conditions where the ambient wet bulb is lower than the room wet bulb, the recirculation system would have a higher COP. In the Miami climate, this condition does not occur often, and over the season the ventilation mode would require less energy. Figures 5 and 6 were based on systems with ideal regenerators and dehumidifiers. The results with less than ideal effectiveness are similar. For the Miami climate, the ventilation mode consistently met the same load as the recirculation mode with about one-half the auxiliary energy requirement. Because of the better performance of the ventilation mode in this climate, the remaining simulations were for this mode only. A parametric analysis showed that system performance was most sensitive to the effectiveness of the sensible heat regenerator. T h e effect of reduced effectiveness is shown schematically in Fig. 7. Lowered effectiveness reduces the temperature of the stream entering the energy supply, T3. In order for the evaporative cooling to produce the same end point and allow the same load to be met, the supply stream must be at lower humidity. This requires a higher regenerating temperature, Ts, and a corresponding increase in energy supplied. The latent and sensible capacities of a given system were found to depend on the regenerating stream temperature and the degree of humidification (effectiveness) in the evaporative coolers. Performance maps for specified ambient and room conditions were generated to determine the interactive effects of temperature and evaporative cooling. A typical map for the ventilation mode for the ASHRAE outdoor design conditions [15] for Miami is shown in Fig. 8 as a plot of the system sensible capacity as a function of latent capacity. The evaporative cooler in the air stream leaving the conditioned space is turned off (i.e. eo = 0) for this particular VENTILATION MODE Ideal Components
7 /t:G I ~ R
Humidity
Temperature
Re0enerotor Effectiveness = 8 0 %
,. I-
Regen~/erotive X~Auxiiiory /Heating .L. Heoti.n9 . Humidity "1" r
Temperature
Fig. 7. Ventilationmode with ideal and 80 per cent effectiveness regenerator.
al. P e r f o r m a n c e Mop Design Ambient
4000[ inlet ~voporotlve =
[ Cooler Effechveness
o,o ~
m --~
OO~<"~
m
".4
~ -~ I ~ , 5
\
% 7~ ~
~
~Regenerotion ~Temperoture
\--
~.~""~\
~
Outlet ~
~
Cooler
,S HUmidifyeKj~
~ Dehumidi fying
o
4OJOO
LATENT LOAD (W)
Fig. 8. Performance map for the ventilation mode for Miami.
map. Increased regenerating stream temperature iT, which is Ta of Fig. 2) increases both sensible and latent capacity. Reduced effectiveness of the evaporative cooler in the supply stream (e~) allows a greater latent load to be met at lower sensible load. Heating and/or humidification are possible at low regenerating stream temperatures and inlet evaporator cooler effectiveness. Maps for increased effectiveness of the evaporative cooler in the exhaust stream show that the system sensible capacity increases at a given operating condition. These maps demonstrate that a wide variety of load combinations can be met. The maps further indicate how the system may be controlled. For example, under daytime conditions of high sensible and moderate latent load, the system may be operated with high values of e~ and variable 7",. However, at night with low sensible but significant latent loads, the system could operate with low e~. Alternatively, operation with eo = 0.5, for example, would be possible with a different combination of 7", and ~. The goal of the control strategy is to meet the required loads at a minimum expenditure of energy, which usually means operating with the lowest possible regenerating stream temperature. The performance maps demonstrate that the system is capable of meeting both sensible and latent loads through proper control of the operating variables. A control system model was developed to allow the system to meet any combination of loads at a minimum energy expenditure. The control strategy evolved from study of the performance maps, and the resulting approach is shown schematically on a psychrometric diagram in Fig. 9. In the comfort zone [15], indicated as zone l, the humidity is below a specified maximum and the room temperature is within acceptable limits. If internal or ambient gains drive the room state into zone 2, cooling only is required, and the system should operate with increased sensible capacity and low, constant latent capacity. If the room state is forced into zone 3, both increased sensible and latent capacity are required, while for zone 4 only increased latent capacity is needed. For zone 5, the system is required to provide heat (operate with negative sensible capacity) and dehumidify, while for zone 6, heating is required. The system could provide heating, but it is more efficient to supply
Simulationsof the performanceof open cycle desiccant systems usingsolar energy
~Control
/
j~Zone 5 IncreaseTR DecreaseE i ¢o:0 Zone 6
Strategy
Zone 4 IncreaseTel I 4, :tow i | %:0
Zone :3 IncreaseTn ~i :high Eo=high
/r / , / / , ' / /
[~ Zone I ~ Zone 2 i~Mochine ~ Increase TR Machine off ~. off ~ ¢i :high r,COMFORT/ Go: high I/,/, ,// Trnln (22.2C) Tmox(2~.~) ROOM TEMPERATURE
ROOM HUMIDITY Wmax I0,0112~G/~I
Fig. 9. Psychrometric diagram for control of the ventilation mode.
heat directly. The control of regenerating stream temperature and evaporative cooler effectiveness needed to return the room state from any zone back to zone 1 are also shown in Fig. 9. The control strategy requires both a thermostat and a humidistat to sense the room state. The zone that the room is in is identified, and the three control variables, T,~s,. . . . t a r , e~ and Co, are incremented appropriately. In the simulations, these variables were maintained at the new levels until the room state was sensed after a specified time increment. This is similar to a continous feedback system with time lags. An upper practical limit was placed on the regenerating stream temperature of 85C and on the evaporative cooler effectiveness of 90 per cent.
277
per cent effectiveness heat exchanger was placed between the collector loop and the storage tank. The water storage tank size was 0.075 m3 per m2 of collector area. The collector and tank flow rates were 0.0139kg/s-m 2 and 0.0222kg/s-m2, respectively. No domestic hot water load was considered in these simulations. The results of a six month simulation of the combined solar-desiccant system are given in Table 1 and presented graphically in Fig. 10. As collector area increases, the amount of auxiliary energy required decreases. At 45 m2, which is about one-third of the house floor area, the solar system provides 95 per cent of the energy required. In an actual system, this would probably be sufficient to meet the air conditioning load with acceptable diurnal variations in room temperature. The system total COP decreases with increased collector area. At large areas, the temperatures available are, at times, higher than required to meet the load. Under these conditions, the system does not use energy efficiently. In addition, there is more energy supplied at large areas than can be stored and energy is dumped to the environment. An examination of the simulation results showed that the regeneration stream was below 650C approximately 60 per cent of the time. The temperature was required to be above 80*(2 only about 5 per cent of the time. This shows that low temperature solar energy can be used in these systems for air conditioning. CONCLUSION
This paper has described some of the significant factors involved in the design, control and operation of RESULTS
The feasibility of using solar energy to regenerate the desiccant was evaluated by adding a standard liquid based flat plate solar system to supply a portion of the auxiliary energy. An air to water heat exchanger with a constant 85 per cent effectiveness was located at point 7 in Figs. I and 2 to provide solar heat to the air entering the auxiliary source. The solar collectors were modeled as having two glass covers, and a non-selective absorbing surface. The collectors have an Fe (ra) of 0.65 and FRUL of 3.47 W/m2*C. A water-ethylene glycol mixture was used as the working fluid in the collector, and a constant 85
60
I
"~ ~ ~ 40 m ~ , ~ - -
I
]
Solar Energy S upplied-~...~ ~ J " ~ - ~ Total Lo..9.. ad
b.I 7 20 W 0
t
0
I0 20 :30 40 COLLECTOR AREA ( m 2)
5o
Fig. 10. Performance of ventilation mode as a function of collector are for Miami,April-September1955.
Table 1. Ventilation s y s t e m p e r f o r m a n c e A p r i l - S e p t e m b e r 1955, Miami, Florida
Collector Area (m2)
7.5
15
30
45
Solar Energy Supplied (GJ)
12.4
23.0
40.8
51.7
A u x i l i a r y Energy Supplied to System (GJ)
33.6
24.2
11.4
4.2
Sensible COP
0.45
0.44
Total COP
0.79
0.77
0.70
0.67
COP Based on A u x i l i a r y Energy
1.09
1.51
3.20
8.72
Sensible Load = 20.8 GJ Latent Load = 15.8 GJ
r
0.40
0.37
278
J.S. NEt.sos et al.
solar assisted, open cycle desiccant systems. These systems are adaptable to solar thermal energy supplies, and have overall coefficients of performance similar to those of absorption machines. Since these systems require energy at lower temperatures than absorption machines, they are better matched to a solar thermal energy supply. Further studies are needed using different system configurations and control strategies for various localities to assess the potential of these systems. Acknowledgement-This work was supported by the Energy Research and Development Administration under contract E (1 I-1)-2588. ItEFI~RI[~CK$ I. R. V. Dunkle. A method of solar air conditioning. Mech. Chem. Engr Trans., lnst Engng, Australia, MCI, 73 (1961). 2. W. Rush, J. Wurm, L. Wright, R. A. Ashworth, A description of the solar-MEC field test installation. ISES Conf., Los Angeles (July-Aug. 1975). 3. J. M. Ruder and J. Rousseau, Development of a SolarPowered Residential Air Conditioner. Garrett Corporation (1975). 4. P. Lunde, Solar desiccant air conditioning with silicagel. Proc. of Workshop on "Use of Solar Energy for Cooling of Buildings",ERDA SAN/1122-76/2 (Aug. 1975). 5. J. S. Nelson, An Investigationof Solar Powered Open Cycle Air Conditioners. M.S. thesis, Solar Energy Laboratory, University of Wisconsin-Madison (Dec. 1976). 6. S. A. Klein, et al.,TRNSY, A transientsimulation program.
University of Wisconsin-Madison, Engng Experiment Station Rep. No. 38 (1977). 7. K. G. T. Hollands, Analysis and design of evaporative cooler pads. Mech. Chem. Engng Trans., Inst. of Engrs., Australia, MC6, 55, (1970). 8. W. M. Kays, and A. L. London, Compact Heat Exchangers, McGraw-Hill, New York (1958). 9. R. V. Dunkle and I. L. Maclaine-Cross, Theory and design of rotary regenerators for air conditioning. Mech. and Chem. Engng Trans., Inst. of Engr., Australia, MC6, 1 (1970). 10. P. J. Banks, D. J. Close and I. L. Maclaine-Cross, Coupled heat and mass transfer in fluid flow through porous media-an analogy with heat transfer. Proc. 4th Int. Heat Trans. Conf., VII, CT3.1, Elsevier, Amsterdam (1970). 11. P. J. Banks, Coupled Equilibrium Heat and Single Adsorbate Transfer in Fluid Flow Through a Porous Medium-I Characteristic Potentials and Specific Capacity Britain, 27 1143-1154 f 1972~. 12. P. J. Banks and D. J. Close, Coupled heat and mass transfer and fluid flow through a porous medium-II prediction for a silica gel air drier using characteristic charts. Chem. Engng Sci., 27, 1157-1168 (1972). 13. I. L. Maclaine-Cross, and P. J. Banks, Coupled heat and mass transfer in regenerators-prediction using an analogy with heat transfer. Int. J. Heat and Mass Trans. Pergamon 15, 1225-1241 (1972). 14. ASHRAE Standards 90-75, energy conservation in New building design. Am. Soc. Heat., Refrig. and Air Cond. Engrs, New York (1975). 15. American Society of Heating, Refrigeration and Air Conditioning Engineers, Handbook of Fundamentals, ASHRAE, New York (1972).
0038-092x/78/1001-0273/$02.00/0
© PergamonPressLtd.,1978. PrintedinGreatBritain
SIMULATIONS OF THE PERFORMANCE OF OPEN CYCLE DESICCANT SYSTEMS USING SOLAR ENERGY J. S. NELSON, W. A. BECKMAN,J. W. MITCHELLand D. J. CLOSE Solar Energy Laboratory Universityof Wisconsin-Madison,Madison,WI 53706, U.S.A. (Received 16 January 1978; revision accepted 15 June 1978)
Abstract- A feasibilitystudy of open cycle air conditioningsystems that use solid desiccants and solar energy has been performed. The two configurationsevaluated are the ventilation mode, in which ambient air is continually introduced into the room, and the recirculationmode, in which room air is recirculated. Seasonal simulationsfor Miami, Florida, show that the auxiliaryenergy requirement for the ventilationmode is about one half that for the recirculation mode. The seasonal COP for the system using solar energy as the auxiliary is approximately0.75. A conventional flat plate solar energy system of moderate size can provide a large fraction of the energy required to meet the sensible and latent loads of a typical house.
describe the operation of a unit a design conditions, but do not evaluate performance over an entire cooling season. This paper describes a computer simulation of the seasonal performance of two operating modes for an open cycle cooling system. In the ventilation mode fresh air is continually introduced into the conditioned space, while in the recirculation mode exhaust air from the conditioned space is reconditioned and reintroduced into the room. The operation is similar to that of the SolarMEC system [2]. A detailed description of the study is given in Ref. [5]. Schematics of the equipment and psychrometric diagrams of the ventilation and recirculation modes are shown in Figs. 1-4. In Figs. 1 and 3, the temperature and humidity are described relative to comfort levels in the
INTRODUCTION
Open cycle desiccant air conditioning systems utilize a desiccant to remove moisture from humid air. Air thus dried is cooled in evaporative coolers and sensible heat exchangers to meet both sensible and latent air conditioning loads. Solar energy or other energy sources are used to regenerate the desiccant. The emphasis in this study is on systems that use solid desiccants. A variety of open cycle systems using solid desiccants and solar energy regeneration have been proposed. Dunkle[l] discusses cooling and dehumidification of air using solar energy and a rotary desiccant wheel. A design point calculation is made to establish that the system can meet the sensible and latent loads of a specific building. Rush et aL [2] report on a test installation in Los Angeles which uses both solar energy and natural gas to regenerate the desiccant wheel. The cooling COP based on total energy supplied was found to be 0.53 at one operating point. Ruder and Rousseau[3] describe a design of a compact system for residential use. A desiccant system is described by Lunde [4] which uses several stages of intercooling in the dehumidification process. Design point calculations are made to establish optimum air flow rates and wheel thickness. These studies
room.
In the ventilation mode, Figs. I and 2, ambient air is dried and heated by the dehumidifier, regeneratively cooled by exhaust air, evaporatively cooled, and introduced into the house. The process can be controlled so that the temperature and humidity of the supply air are lower than that in the house, and this allows the sensible and latent loads to be met. The exhaust air is first
Open Cycle Desiccant Cooling Ventilation Mode WATER
ROOM
Tem~eret,.=re Comfortable HJmJdity: Comfortable
HEAT
l
" ~
War
EXHAUST
Hot
--
Domp
Regenerator WATER
o C Moist ~
~: ~
~
~
AMBIENT
Dehumidifier . ~ W e t Dry -..
Hot Dry
Fig. 1. Schematicof ventilationmode. 273
_
2?4
J. S. NELSOI~et
Ventilation
/
/EXHAUST
Mode
al.
Recirculotlon /
EXHAUST ~'"
Humidity
®/
%
Humidity /
.
IC~AMBIENT
Temperature
Temperature
Fig. 2. Psychrometricdiagramfor ventilationmode,
Fig. 4. Psychrometricdiagram for recirculation mode.
evaporatively cooled to provide a low temperature sink for heat transfer from the supply air in the regenerator. The air is then heated by an energy supply which could be either a conventional fuel, a solar source, or both. Passage of the heated air through the dehumidifier regenerates the desiccant and cools the air. The psychrometric diagram shows the states of the air for typical operating conditions. Ambient conditions change over the course of the day, with resulting changes in the sensible and latent loads and in the psychrometric diagram. The recirculation mode shown in Figs: 3 and 4 employs the same components as the ventilation mode. However, room air is recirculated and ambient air is used only for regeneration. Room air is dehumidified and heated by the desiccant wheel, regeneratively cooled, and then evaporatively cooled prior to reentering the room. Ambient ak is evaporatively cooled, regeneratively heated, and then heated by an energy supply. The heated ambient air passes through the dehumidifier and regenerates the desiccant. For both systems, the sensible heat regenerator and dehumidifier are assumed to be rotary elements, but direct transfer exchangers and fixed beds could equally well be used. The objectives of this study are: 1. To develop computer models for the various components of a desiccant system. 2. To determine the annual thermal performance of the
system and the ability to meet a combination of latent and sensible loads. 3. To evaluate the potential of a solar energy system to provide regeneration of the dehumidifier under typical weather conditions. 4. To develop a control strategy to allow the system to provide comfort under varying ambient conditions.
COMPONENT MODELS
Computer models of the various components were developed for use with the transient simulation program TRNSYS [6]. Evaporative coolers were modeled using the analytical relations developed by Hollands [7]. The major assumption is that the Lewis number equals unity for the air-water vapor mixture. This allows the saturation effectiveness to be formulated in terms of an overall Ntu similar to that for a sensible heat exchanger. The sensible heat regenerator was modeled using the relationships of Kays and London[8]. The development does not account for fluid leakage or carryover from one stream to the other, and assumes a uniform temperature across the inlet air streams. The specific rotary regenerator modeled was one constructed of polypropylene ribbon wound spirally around the rotation axis. The air flow is through the parallel passages between successive layers of the ribbon. This type of regenerator has been shown by Dunlde and Maclaine-Cross[9] to
Open Cycle Desiccant Cooling Recirculotion Mode AMBIENT
/
HEAT
WATER
EXHAUST
I II
ROOM
Temperature: Comfortable
Humidity:
Comfortable ~
°"-LI
:'°u=t';*l I
" °'' U )
Fig. 3. Schematicof recirculation mode.
275
Simulations of the performance of open cycle desiccant systems using solar energy have satisfactory heat exchange and pressure drop properties for air conditioning applications. The model for the rotary dehumidfier is based on an analysis developed in a series of papers by Banks, Close and Maclalne--Cross[10-13]. This analysis allows the performance of a rotary dehumidifier to be calculated by analogy with that for heat transfer in rotary sensible heat regenerators. Characteristic potentials and specific capacity ratios that are analogous to temperature and specific heat ratio are formulated, and nondimensional parameters analogous to effectiveness, Ntu, and capacity rate ratio are formed. Then sensible heat regenerator theory [8] is used to evaluate the inlet and outlet states and the performance. As shown by Banks and MaclaineCross, the process line for a well designed dehumidifier approximates a constant enthalpy process, as indicated in Figs. 2 and 4. The specific dehumidifier modeled was a parallel passage dehumidifier with construction similar to that of the sensible heat regenerator. The desiccant material was taken to be silica gel, with properties as given in Ref. [12]. The simulations were performed using 1955 weather data for Miami, Florida. Miami was chosen as there are high sensible and latent loads which vary significantly over the course of the day. Hourly sensible and latent loads were evaluated using the transfer function approach for a house of 140m2 floor area that was insulated to ASHRAE 90-75 standards [14]. At Miami outdoor design conditions, the house model generates sensible and latent loads of 3800 W and 1000W, respectively. SYSTEM ~
C
E
CHARACTERISTICS
Systems with ideal (100 per cent effectiveness) components were studied in order to establish upper limits on performance and provide insight into system operation. The load was that from the house model for July. The air flow through the system was 0.34kg/s, which was based on accepted practice for vapor compression machines of 400 cfm/ton. The regenerating stream temperature (state point 8 in Fig. 2 and state point 4 in Fig. 4) was held constant. The latent and sensible loads met by the system and the system total and sensible coefficients of performance were evaluated. The coefficient of performance is the load met divided by the energy supplied by the external energy supply. The variation of the ideal coefficient of performance with regenerating stream temperature is shown in Fig. 5. Although these results are valid only for the Miami climate, several general conclusions may be drawn. High values of COP may be obtained at low regenerating stream temperatures. However, the capacity of the system is low even though energy is used effectively. At regenerating stream temperatures above about 80°C, the COP for the ventilation mode decreases significantly. A high regenerating stream temperature produces a very low humidity in the supply stream (state point 2, Fig. 2), but the resulting wet bulb temperature for state 3 is not correspondingly reduced. As a result, only a slightly greater load is met with the greatly increased supply energy.
Vgnfila!ion Mode COP--..,,
20 15
cop 1.0 0.5
60
0 90 I00 Trecjeneratio n (°C)
liO
120
Fig. 5. Coefficient of performance for ideal systems as a function
of regenerating stream temperature, for July 1955,Miami. The ideal COP for the ventilation mode is about twice that of the recirculation mode. The reason for this can be seen from a psychrometric diagram for a given operating condition. Both cycles are represented in Fig. 6 for the same ambient and room states and for the same house latent and sensible loads. The sensible heat regenerator in the ventilation mode is able to increase the temperature of the air entering the energy supply T7 to a higher value than that in the recirculation mode, T3. Correspondingly, in the ventilaiton mode, the dried air stream is regeneratively cooled to a lower temperature, T3, than in the recirculating mode, Ts. To achieve the same inlet state to the room, state 9, the absolute humidity of the dried air in the recirculation mode must be lower than that in the ventilation mode. This requires a higher temperature of the air entering the regenerator in the recirculation mode, T4, than in the ventilation mode, Ts. As a result, the amount of auxiliary energy that must be supplied in the recirculation mode is greater than for the ventilation mode. Further examination of Fig. 6 shows that this relation Ventilation Mode /
/
EXHAUST 9
Hum=dily ~AM BIEN T " ~
~
"I
~ zSupplieO
Temperature
Recirculotion /
T7
// EXHAUST 5~\
,K,
l"emperature
TB
-x. Hom,o'.,
T~
T~
Fig. 6. Comparison of ventilation and recirculafion mode, with ideal components.
2%
J. S. NELSONet
between the two cycles depends on ambient conditions. For conditions where the ambient wet bulb is lower than the room wet bulb, the recirculation system would have a higher COP. In the Miami climate, this condition does not occur often, and over the season the ventilation mode would require less energy. Figures 5 and 6 were based on systems with ideal regenerators and dehumidifiers. The results with less than ideal effectiveness are similar. For the Miami climate, the ventilation mode consistently met the same load as the recirculation mode with about one-half the auxiliary energy requirement. Because of the better performance of the ventilation mode in this climate, the remaining simulations were for this mode only. A parametric analysis showed that system performance was most sensitive to the effectiveness of the sensible heat regenerator. T h e effect of reduced effectiveness is shown schematically in Fig. 7. Lowered effectiveness reduces the temperature of the stream entering the energy supply, T3. In order for the evaporative cooling to produce the same end point and allow the same load to be met, the supply stream must be at lower humidity. This requires a higher regenerating temperature, Ts, and a corresponding increase in energy supplied. The latent and sensible capacities of a given system were found to depend on the regenerating stream temperature and the degree of humidification (effectiveness) in the evaporative coolers. Performance maps for specified ambient and room conditions were generated to determine the interactive effects of temperature and evaporative cooling. A typical map for the ventilation mode for the ASHRAE outdoor design conditions [15] for Miami is shown in Fig. 8 as a plot of the system sensible capacity as a function of latent capacity. The evaporative cooler in the air stream leaving the conditioned space is turned off (i.e. eo = 0) for this particular VENTILATION MODE Ideal Components
7 /t:G I ~ R
Humidity
Temperature
Re0enerotor Effectiveness = 8 0 %
,. I-
Regen~/erotive X~Auxiiiory /Heating .L. Heoti.n9 . Humidity "1" r
Temperature
Fig. 7. Ventilationmode with ideal and 80 per cent effectiveness regenerator.
al. P e r f o r m a n c e Mop Design Ambient
4000[ inlet ~voporotlve =
[ Cooler Effechveness
o,o ~
m --~
OO~<"~
m
".4
~ -~ I ~ , 5
\
% 7~ ~
~
~Regenerotion ~Temperoture
\--
~.~""~\
~
Outlet ~
~
Cooler
,S HUmidifyeKj~
~ Dehumidi fying
o
4OJOO
LATENT LOAD (W)
Fig. 8. Performance map for the ventilation mode for Miami.
map. Increased regenerating stream temperature iT, which is Ta of Fig. 2) increases both sensible and latent capacity. Reduced effectiveness of the evaporative cooler in the supply stream (e~) allows a greater latent load to be met at lower sensible load. Heating and/or humidification are possible at low regenerating stream temperatures and inlet evaporator cooler effectiveness. Maps for increased effectiveness of the evaporative cooler in the exhaust stream show that the system sensible capacity increases at a given operating condition. These maps demonstrate that a wide variety of load combinations can be met. The maps further indicate how the system may be controlled. For example, under daytime conditions of high sensible and moderate latent load, the system may be operated with high values of e~ and variable 7",. However, at night with low sensible but significant latent loads, the system could operate with low e~. Alternatively, operation with eo = 0.5, for example, would be possible with a different combination of 7", and ~. The goal of the control strategy is to meet the required loads at a minimum expenditure of energy, which usually means operating with the lowest possible regenerating stream temperature. The performance maps demonstrate that the system is capable of meeting both sensible and latent loads through proper control of the operating variables. A control system model was developed to allow the system to meet any combination of loads at a minimum energy expenditure. The control strategy evolved from study of the performance maps, and the resulting approach is shown schematically on a psychrometric diagram in Fig. 9. In the comfort zone [15], indicated as zone l, the humidity is below a specified maximum and the room temperature is within acceptable limits. If internal or ambient gains drive the room state into zone 2, cooling only is required, and the system should operate with increased sensible capacity and low, constant latent capacity. If the room state is forced into zone 3, both increased sensible and latent capacity are required, while for zone 4 only increased latent capacity is needed. For zone 5, the system is required to provide heat (operate with negative sensible capacity) and dehumidify, while for zone 6, heating is required. The system could provide heating, but it is more efficient to supply
Simulationsof the performanceof open cycle desiccant systems usingsolar energy
~Control
/
j~Zone 5 IncreaseTR DecreaseE i ¢o:0 Zone 6
Strategy
Zone 4 IncreaseTel I 4, :tow i | %:0
Zone :3 IncreaseTn ~i :high Eo=high
/r / , / / , ' / /
[~ Zone I ~ Zone 2 i~Mochine ~ Increase TR Machine off ~. off ~ ¢i :high r,COMFORT/ Go: high I/,/, ,// Trnln (22.2C) Tmox(2~.~) ROOM TEMPERATURE
ROOM HUMIDITY Wmax I0,0112~G/~I
Fig. 9. Psychrometric diagram for control of the ventilation mode.
heat directly. The control of regenerating stream temperature and evaporative cooler effectiveness needed to return the room state from any zone back to zone 1 are also shown in Fig. 9. The control strategy requires both a thermostat and a humidistat to sense the room state. The zone that the room is in is identified, and the three control variables, T,~s,. . . . t a r , e~ and Co, are incremented appropriately. In the simulations, these variables were maintained at the new levels until the room state was sensed after a specified time increment. This is similar to a continous feedback system with time lags. An upper practical limit was placed on the regenerating stream temperature of 85C and on the evaporative cooler effectiveness of 90 per cent.
277
per cent effectiveness heat exchanger was placed between the collector loop and the storage tank. The water storage tank size was 0.075 m3 per m2 of collector area. The collector and tank flow rates were 0.0139kg/s-m 2 and 0.0222kg/s-m2, respectively. No domestic hot water load was considered in these simulations. The results of a six month simulation of the combined solar-desiccant system are given in Table 1 and presented graphically in Fig. 10. As collector area increases, the amount of auxiliary energy required decreases. At 45 m2, which is about one-third of the house floor area, the solar system provides 95 per cent of the energy required. In an actual system, this would probably be sufficient to meet the air conditioning load with acceptable diurnal variations in room temperature. The system total COP decreases with increased collector area. At large areas, the temperatures available are, at times, higher than required to meet the load. Under these conditions, the system does not use energy efficiently. In addition, there is more energy supplied at large areas than can be stored and energy is dumped to the environment. An examination of the simulation results showed that the regeneration stream was below 650C approximately 60 per cent of the time. The temperature was required to be above 80*(2 only about 5 per cent of the time. This shows that low temperature solar energy can be used in these systems for air conditioning. CONCLUSION
This paper has described some of the significant factors involved in the design, control and operation of RESULTS
The feasibility of using solar energy to regenerate the desiccant was evaluated by adding a standard liquid based flat plate solar system to supply a portion of the auxiliary energy. An air to water heat exchanger with a constant 85 per cent effectiveness was located at point 7 in Figs. I and 2 to provide solar heat to the air entering the auxiliary source. The solar collectors were modeled as having two glass covers, and a non-selective absorbing surface. The collectors have an Fe (ra) of 0.65 and FRUL of 3.47 W/m2*C. A water-ethylene glycol mixture was used as the working fluid in the collector, and a constant 85
60
I
"~ ~ ~ 40 m ~ , ~ - -
I
]
Solar Energy S upplied-~...~ ~ J " ~ - ~ Total Lo..9.. ad
b.I 7 20 W 0
t
0
I0 20 :30 40 COLLECTOR AREA ( m 2)
5o
Fig. 10. Performance of ventilation mode as a function of collector are for Miami,April-September1955.
Table 1. Ventilation s y s t e m p e r f o r m a n c e A p r i l - S e p t e m b e r 1955, Miami, Florida
Collector Area (m2)
7.5
15
30
45
Solar Energy Supplied (GJ)
12.4
23.0
40.8
51.7
A u x i l i a r y Energy Supplied to System (GJ)
33.6
24.2
11.4
4.2
Sensible COP
0.45
0.44
Total COP
0.79
0.77
0.70
0.67
COP Based on A u x i l i a r y Energy
1.09
1.51
3.20
8.72
Sensible Load = 20.8 GJ Latent Load = 15.8 GJ
r
0.40
0.37
278
J.S. NEt.sos et al.
solar assisted, open cycle desiccant systems. These systems are adaptable to solar thermal energy supplies, and have overall coefficients of performance similar to those of absorption machines. Since these systems require energy at lower temperatures than absorption machines, they are better matched to a solar thermal energy supply. Further studies are needed using different system configurations and control strategies for various localities to assess the potential of these systems. Acknowledgement-This work was supported by the Energy Research and Development Administration under contract E (1 I-1)-2588. ItEFI~RI[~CK$ I. R. V. Dunkle. A method of solar air conditioning. Mech. Chem. Engr Trans., lnst Engng, Australia, MCI, 73 (1961). 2. W. Rush, J. Wurm, L. Wright, R. A. Ashworth, A description of the solar-MEC field test installation. ISES Conf., Los Angeles (July-Aug. 1975). 3. J. M. Ruder and J. Rousseau, Development of a SolarPowered Residential Air Conditioner. Garrett Corporation (1975). 4. P. Lunde, Solar desiccant air conditioning with silicagel. Proc. of Workshop on "Use of Solar Energy for Cooling of Buildings",ERDA SAN/1122-76/2 (Aug. 1975). 5. J. S. Nelson, An Investigationof Solar Powered Open Cycle Air Conditioners. M.S. thesis, Solar Energy Laboratory, University of Wisconsin-Madison (Dec. 1976). 6. S. A. Klein, et al.,TRNSY, A transientsimulation program.
University of Wisconsin-Madison, Engng Experiment Station Rep. No. 38 (1977). 7. K. G. T. Hollands, Analysis and design of evaporative cooler pads. Mech. Chem. Engng Trans., Inst. of Engrs., Australia, MC6, 55, (1970). 8. W. M. Kays, and A. L. London, Compact Heat Exchangers, McGraw-Hill, New York (1958). 9. R. V. Dunkle and I. L. Maclaine-Cross, Theory and design of rotary regenerators for air conditioning. Mech. and Chem. Engng Trans., Inst. of Engr., Australia, MC6, 1 (1970). 10. P. J. Banks, D. J. Close and I. L. Maclaine-Cross, Coupled heat and mass transfer in fluid flow through porous media-an analogy with heat transfer. Proc. 4th Int. Heat Trans. Conf., VII, CT3.1, Elsevier, Amsterdam (1970). 11. P. J. Banks, Coupled Equilibrium Heat and Single Adsorbate Transfer in Fluid Flow Through a Porous Medium-I Characteristic Potentials and Specific Capacity Britain, 27 1143-1154 f 1972~. 12. P. J. Banks and D. J. Close, Coupled heat and mass transfer and fluid flow through a porous medium-II prediction for a silica gel air drier using characteristic charts. Chem. Engng Sci., 27, 1157-1168 (1972). 13. I. L. Maclaine-Cross, and P. J. Banks, Coupled heat and mass transfer in regenerators-prediction using an analogy with heat transfer. Int. J. Heat and Mass Trans. Pergamon 15, 1225-1241 (1972). 14. ASHRAE Standards 90-75, energy conservation in New building design. Am. Soc. Heat., Refrig. and Air Cond. Engrs, New York (1975). 15. American Society of Heating, Refrigeration and Air Conditioning Engineers, Handbook of Fundamentals, ASHRAE, New York (1972).