You are on page 1of 18

Available online at www.sciencedirect.

com

Solar Energy 84 (2010) 227–244


www.elsevier.com/locate/solener

Comparative study of different solar cooling systems for buildings


in subtropical city
K.F. Fong *, T.T. Chow, C.K. Lee, Z. Lin, L.S. Chan
Division of Building Science and Technology, College of Science and Engineering, City University of Hong Kong, Tat Chee Avenue,
Kowloon Tong, Kowloon, Hong Kong, China

Received 11 June 2009; received in revised form 28 October 2009; accepted 7 November 2009
Available online 8 December 2009

Communicated by: Associated Editor Ruzhu Wang

Abstract

In recent years, more and more attention has been paid on the application potential of solar cooling for buildings. Due to the fact that
the efficiency of solar collectors is generally low at the time being, the effectiveness of solar cooling would be closely related to the avail-
ability of solar irradiation, climatic conditions and geographical location of a place. In this paper, five types of solar cooling systems were
involved in a comparative study for subtropical city, which is commonly featured with long hot and humid summer. The solar cooling
systems included the solar electric compression refrigeration, solar mechanical compression refrigeration, solar absorption refrigeration,
solar adsorption refrigeration and solar solid desiccant cooling. Component-based simulation models of these systems were developed,
and their performances were evaluated throughout a year. The key performance indicators are solar fraction, coefficient of performance,
solar thermal gain, and primary energy consumption. In addition, different installation strategies and types of solar collectors were com-
pared for each kind of solar cooling system. Through this comparative study, it was found that solar electric compression refrigeration
and solar absorption refrigeration had the highest energy saving potential in the subtropical Hong Kong. The former is to make use of
the solar electric gain, while the latter is to adopt the solar thermal gain. These two solar cooling systems would have even better per-
formances through the continual advancement of the solar collectors. It will provide a promising application potential of solar cooling
for buildings in the subtropical region.
Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Solar cooling; Photovoltaic; Solar thermal technology; Solar collectors; Air-conditioning; Plant simulation

1. Introduction fuels, leading to the non-stopping emission of global warm-


ing gases. In a typical subtropical city, like Hong Kong, the
In the field of heating, ventilating and air-conditioning summer is hot and humid, and the winter is generally tem-
(HVAC) engineering, electricity has been used to energize perate. Therefore air-conditioning is essential in a subtrop-
the air-conditioning and refrigeration systems for a cen- ical city throughout a year, particularly for the commercial
tury. It is undoubted that HVAC is one of the indispens- and institutional buildings. In the subtropical Hong Kong,
able provisions in modern life, no matter in working or air-conditioning and refrigeration are the largest area of
living environments. Normally one-third to half of the electricity consumption, accounting for 33% in 2006 (Hong
annual total electricity consumption is used for air-condi- Kong Energy End-use Data, 2008).
tioning and refrigeration in the metropolis worldwide. Many facts show that the climate change is getting
80% of electricity is still generated by burning the fossil worse. The changes of global surface temperature, global
average sea level, snow and ice over 1850–2000 are
*
Corresponding author. Tel.: +852 27888724; fax: +852 27889716. reported (IPCC, 2007). To have wider application of
E-mail address: bssquare@cityu.edu.hk (K.F. Fong). renewable energy is a crucial sustainable strategy to allevi-

0038-092X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.solener.2009.11.002
228 K.F. Fong et al. / Solar Energy 84 (2010) 227–244

Nomenclature

COP coefficient of performance Qsolar solar thermal gain from the solar collectors
COPab COP of solar absorption refrigeration (kW)
COPad COP of solar adsorption refrigeration SF solar fraction
COPdesi COP of solar solid desiccant cooling SFelect solar fraction of solar electric compression
COPelect COP of solar electric compression refrigera- refrigeration
tion SFthermal
COPmech COP of solar mechanical compression refrig- solar fraction of solar thermal cooling
eration W power input to DC motor (kW)
Ep primary energy consumption (kWh) WAHU electrical energy consumption of AHU (kWh)
Ep,desi primary energy consumption of solar solid des- Waux auxiliary electrical power from the city power
iccant cooling (kWh) grid (kW)
Ep,elect primary energy consumption of solar electric Wfans electrical energy consumption of supply air fan
compression refrigeration (kWh) and exhaust air fan (kWh)
Elp,therma Wparasitic
primary energy consumption of solar thermal electrical energy consumption of parasitic
refrigeration (kWh) equipment, including pumps and cooling tower
Gelect solar electric gain (kWh) (kWh)
Gsolar solar thermal gain (kWh) W para,desi electrical energy consumption of parasitic
Gsolar,eq solar thermal gain equivalent (kWh) equipment, including pumps, desiccant wheel,
ho specific enthalpy of outdoor air (kW/kg) heat wheel and evaporative coolers (kWh)
hs specific enthalpy of supply air (kW/kg) Wsolar solar electric gain from photovoltaic panels
m total number of time steps in a month for hot (kW)
water pump in operation y total number of time steps in a year for hot
m0 total number of time steps in a month for chiller water pump in operation
in operation y0 total number of time steps in a year for chiller in
m_ a mass flow rate of supply air (kg/s) operation
Qaux heat output of auxiliary heater (kWh)
Qboil heat input to boiler of heat engine (kW) Greek symbols
Qde heat input for desorption (kW) ge energy efficiency for electrical energy converted
Qe refrigeration effect (kW) into primary energy
Qgen heat input to generator (kW) gg energy efficiency for gas energy converted into
Qregen heat input for regeneration (kW) primary energy

ate the climate change. Global warming has caused the reviews have been made about the feasibility of wider
increasing trend of the environmental temperatures. This application of solar cooling technologies (Fan et al.,
gradual change would affect the future air-conditioning 2007; Hwang et al., 2008; Kim and Infante Ferreira,
requirements in Hong Kong (Lam et al., 2008). The 2008). A number of demonstration projects have been
demand of air-conditioning would be increasing due to launched to gain more experience in the design and opera-
the effect of climate change and global warming. If we still tion of solar refrigeration and air-conditioning (Eicker,
rely on the conventional electric-driven air-conditioning 2003; Henning, 2004; Zhai et al., 2008). Air-conditioning
but the electricity is generated from the fossil fuels, the is already part of our daily lives, solar cooling certainly
greenhouse gas emission would continuously worsen global has a positive and long-term impact. For the continual
warming, in turn the demand of air-conditioning would be population and economic growth, wider use of solar energy
further increasing. If this dead loop cannot be solved, even in air-conditioning would secure the increasing energy
‘global heating’ would come one day. Therefore it is urgent demand.
to minimize the consumption of fossil fuels and promote Solar cooling can be broadly categorized into solar elec-
wider use of solar energy, particularly in refrigeration tric refrigeration, solar thermal refrigeration and solar
and air-conditioning. thermal air-conditioning. In the first category, the solar
Solar cooling is envisioned as the sustainable means to electric compression refrigeration uses photovoltaic (PV)
provide air-conditioning and refrigeration, since solar panels to power conventional refrigeration machine. In
energy is considered to be the primary energy source. Tech- the second category, the refrigeration effect can be pro-
nically, solar cooling is a feasible way to replace the electric duced through solar thermal gain, the solar mechanical
refrigeration machines for buildings. In recent years, more compression refrigeration, solar absorption refrigeration
K.F. Fong et al. / Solar Energy 84 (2010) 227–244 229

and solar adsorption refrigeration are the three common Hong Kong located at 22.32°N and 114.17°E. The office
options. In the third category, the conditioned air can be was single storey with floor area of 200 m2. The daily occu-
directly provided through the solar thermal gain by means pancy schedule was from 08:00 to 18:00, totally 10 h. The
of desiccant cooling. Both solid and liquid sorbents are wall-fenestration ratio was 0.5. The internal and external
available, like silica gel and lithium chloride, respectively. shading factors of fenestration were 0.8 and 0.2, respec-
However there is a risk of carry-over of the liquid sorbent tively. The internal heat gains included 24 persons seated
into the indoor space by the conditioned air, so it is more at work, 230 W/person of personal computer, and 17 W/
suitable to apply the solar solid desiccant cooling. As a m2 lighting with 30% convective part. The provisions and
result, the following five types of solar cooling systems were schedules of the internal heat gains were based on the local
included in this study: design practice (Code of Practice for Energy Efficiency of
Air Conditioning Installations, 2007). The fresh air flow
 solar electric compression refrigeration; rate was 10 litre/s/person. The design indoor conditions
 solar mechanical compression refrigeration; were 25.5 °C and 60% RH, and the year-round weather
 solar absorption refrigeration; data followed the Typical Meteorological Year of Hong
 solar adsorption refrigeration; and Kong (Chan et al., 2006). The design entering hot water,
 solar solid desiccant cooing. condenser water and chilled water temperatures were
90 °C, 30 °C and 13 °C, respectively. The chiller was sized
The first one uses the solar electric approach, while the based on the design entering condenser and chilled water
rest apply the solar thermal technology. temperatures at 30 °C and 13 °C, respectively. The con-
The installation strategy and types of solar collectors denser and chilled water flow rates were designed based
would directly affect the system performance since the on the 5 °C change between entering and leaving tempera-
amount of solar irradiation acquired would be different. tures. The component-based simulation program TRNSYS
The flat plate collectors and evacuated tubes are the com- (2006) and its component library TESS(2006) were used to
mon types of solar collectors used in the solar thermal sys- build and run the system simulations. The simulation time
tems. The parabolic concentrators are also getting more step was 6 min. Table 1 summarizes the common parame-
popular in building applications. It is worthy to understand ters used in modeling different solar cooling systems.
their performances in response to different system design
and operating requirements based on the solar irradiation, 2.2. Solar electric compression refrigeration
climatic and loading conditions in the subtropical region.
In the previous research works (Hwang et al., 2008; Kim 2.2.1. Model building
and Infante Ferreira, 2008), simulation and experimental Generally the solar electric compression refrigeration,
study on a certain type of solar cooling system has been the air side system and the corresponding control provi-
pinpointed, but not across a variety of solar cooling sys- sions were built by using TRNSYS and TESS, as shown
tems with different installation strategies and types of solar in Fig. 1. The major equipment included the PV panels,
collectors. direct current (DC) motor, power regulator, vapour com-
The rest of this paper is organized as follows: Section 2 pression chiller, chilled water pump, cooling tower, con-
describes development of the simulation models of the five denser water pump, air handling unit (AHU). A detailed
types of solar cooling systems serving a typical local office. chiller model was developed separately. The modeling
Section 3 addresses the framework of the comparative method of DC motor was based on Hughes (1993), that
study of different scenarios, including different installation of condenser and evaporator based on Lee (2008), and that
strategies and types of solar collectors associated with the of the compressor followed the parameter estimation tech-
solar cooling systems. Section 4 presents the definitions nique adopted by Jin and Spitler(2002). The total and space
of the performance indicators, including solar fraction, cooling load were 29 kW and 20 kW, respectively. The gen-
coefficient of performance, solar thermal gain, and primary eral modeling parameters of the solar electric compression
energy consumption. Section 5 discusses the results of the refrigeration is shown in Table 1.
preliminary study for the installation strategies and types
of solar collectors. Section 6 discusses results of the full 2.2.2. Part-load control of chiller
comparative study based on the performance indicators. The PV panels were designed to furnish the necessary
The feasible choice of solar cooling systems would be iden- electric power to drive the chiller. When the supply from
tified and recommended. Section 7 is the conclusion. the PV panels was insufficient due to the weather condition,
a power regulator was used to draw auxiliary electrical
2. Simulation of the five solar cooling systems power from the city power grid. The power regulator was
capable of tracking the maximum power point of the PV
2.1. General design information panels at any circumstances, so that the auxiliary electrical
power requirement could be minimized. To prevent fre-
In this study, each type of the solar cooling systems was quent on/off of the chiller in low-load condition, a part-
designed to serve a common typical office in the subtropical load controller for the chiller was used to modulate the
230 K.F. Fong et al. / Solar Energy 84 (2010) 227–244

Table 1
General modeling parameters for the solar cooling systems.
Solar electric Solar mechanical Solar Solar Solar solid
compression compression absorption adsorption desiccant
refrigeration refrigeration refrigeration refrigeration cooling
Air handling unit
Supply air mass flow rate (kg sÿ1) 2.15 2.15 2.15 2.15 3.36
Supply fan power (kW) 2.15 2.15 2.15 2.15 3.36
Supply fan efficiency 70% 70% 70% 70% 70%
Face area of cooling coil (m2) 0.9 0.9 0.9 0.9 N/A
Chiller
Chilled water mass flow rate (kg sÿ1) 1.4 1.4 1.4 1.4 N/A
Condenser water mass flow rate (kg sÿ1) 1.9 1.9 1.6 1.8 N/A
Specific heat capacity of liquid water (kJ kgÿ1 Kÿ1) 4.19 4.19 4.19 4.19 N/A
Specific heat capacity of water vapour (kJ kgÿ1 Kÿ1) N/A N/A 1.866 1.866 2.028
Overall heat transfer value of evaporator coil (kW Kÿ1) 2.1 2.1 4.3 6.0 N/A
Overall heat transfer value of condenser coil (kW Kÿ1) 3.1 3.1 4.8 6.5 N/A
Condenser/cooling water system
Cooling tower air volume flow rate (m3 sÿ1) 1.17 11.11 2.22 2.78 N/A
Cooling tower fan power (kW) 0.36 3.42 0.683 0.855 N/A
Cooling tower fan efficiency 65% 65% 65% 65% N/A
Condenser water pump power (kW) 0.131 3.23 0.093 0.115 N/A
Condenser water pump efficiency 60% 60% 60% 60% N/A
Chilled water system
Chilled water pump power (kW) 0.084 0.084 0.084 0.084 N/A
Chilled water pump efficiency 60% 60% 60% 60% N/A
Hot/regenerative water system
Hot water mass flow rate (kg sÿ1) N/A 2.8 2.0 2.8 1.92
Hot water pump power (kW) N/A 0.233 0.2 0.233 0.192
Hot water pump efficiency N/A 60% 60% 60% 60%
Regenerative water mass flow rate (kg sÿ1) N/A 15 2.0 2.8 1.92
Regenerative water pump power (kW) N/A 1.55 0.247 0.341 0.141
Regenerative water pump efficiency N/A 60% 60% 60% 60%
General
Area of solar collectors (m2) 100 100 100 100 100
Minimum driving temperature (°C) N/A 82 67 64 60
Capacity of auxiliary heater (kW) N/A 323 42 69 89
Remark: N/A means ‘not applicable’.

set point of the power regulator between 1.5 and 7.5 kW in return temperature of the chiller exceeded 12 °C, and
five steps. The set point started at 1.5 kW (the minimum stepped up if the water temperature dropped less than
required power input for the chiller) when the chilled water 1 °C after one simulation time step. On the contrary, when

Cooling
Tower

Thermostat with
Cooling Tower
PV Panels Controller T Condenser Water Pump Return Air

Condenser
Power AHU
Regulator
DC Fresh
PR Compressor Expansion -
Air
Motor Device Supply
Air
Vapour T Indoor Space
City Power 3-way
Grid Evaporator Compression C
Valve
Chiller Cooling Coil
Controller
C T
Part-load
controller
Chilled Water Pump

Fig. 1. Schematic diagram of HVAC system designed with solar electric compression refrigeration.
K.F. Fong et al. / Solar Energy 84 (2010) 227–244 231

the chilled water return temperature was below 12 °C, the A number of validated models were involved in the solar
set point would be stepped down if the water temperature mechanical compression refrigeration. Modeling approach
dropped more than 1 °C after one simulation time step. of the Rankine cycle followed that proposed by Putten and
By employing the part-load control, the chiller capacity Colonna (2007). The performance of a multistage turbine
could be adjusted through the input control of power at different operating conditions was approximated by Sto-
regulator. dola’s ‘law of the ellipse’ according to Dixon (1998). Mod-
eling methods of condenser, evaporator and compressor
2.2.3. Control and operation of chiller and associated were same as those for the solar electric compression refrig-
equipment eration in Section 2.2.1. The general modeling parameters
The chilled water entering the cooling coil of AHU was for the solar mechanical compression refrigeration are
controlled by a three-way valve based on space tempera- shown in Table 1, while the specific parameters are pre-
ture. The cooling coil controller generated a linear signal sented in Table 2.
from 0 to 1 corresponding to no to full flow to the cooling
coil when the space temperature varied from 24.5 to 2.3.2. Control and operation of Rankine-cycle-driven chiller
26.5 °C. To avoid instability when simulating the cooling The general provisions for the year-round control and
coil, the valve would open at least 30% when the space tem- operation of the chiller and associated equipment were
perature was above 24.5 °C. The operation of the chiller same as those in Section 2.2.3 for the solar electric com-
was determined by a controller that provided an on/off sig- pression refrigeration. However, a specific consideration
nal between 12 and 9 °C with hysteresis based on the was required for modeling the Rankine-cycle-driven chiller
chilled water return temperature at the chiller. When the in order to ensure its year-round operation. Chiller opera-
chiller was on, the condenser water pump would operate. tion at an insufficient driving temperature for the Rankine
The operation of the cooling tower would be controlled
by a thermostat with hysteresis between 20 and 15 °C based
Table 2
on the condenser water temperature leaving the chiller. Modeling parameters for the Rankine cycle of solar mechanical compres-
Basically the chilled water pump and AHU were in opera- sion refrigeration.
tion according to the daily occupancy schedule. Parameter Value
Cooler water mass flow rate (kg sÿ1) 14
2.3. Solar mechanical compression refrigeration Specific heat capacity of circulating fluid (kJ kgÿ1 Kÿ1) 4.19
Density of circulating fluid (kg mÿ3) 1000
2.3.1. Model building Volume of circulating fluid in boiler coil (m3) 0.08
Development of the solar mechanical compression Volume of circulating fluid in cooler coil (m3) 0.08
refrigeration was based on the schematic diagram shown Volume of circulating fluid in feed pump and connecting pipeline 0.004
(m3)
in Fig. 2. The essential equipment consisted of the solar Overall heat transfer value of boiler coil (kW Kÿ1) 50
collectors, hot water pump, hot water tank, auxiliary hea- Overall heat transfer value of cooler coil (kW Kÿ1) 50
ter, regenerative water pump, heat engine using Rankine Polytropic efficiency of turbine 80%
cycle, vapour compression chiller, chilled water pump, Mechanical efficiency of turbine 90%
cooling tower, condenser water pump and AHU. Ratio of specific heat for circulating fluid 1.3

Cooling
Tower

Hot Water Pump Thermostat with Cooling


Controller Tower Controller T Condenser Water Pump
C Regen. Feed
Cooler Condenser
Solar T Water Pump
Collectors T
Pump
Vapour
Hot Expansion Compression
Boiler Turbine Compressor
Hot Water Pump Water Device Chiller
Tank

Aux. Evaporator
Heater
Heat Engine
Town Gas (Rankine Cycle)
Supply
Chilled water supply and
return for the AHU serving the
indoor space (same as Fig. 1)

Fig. 2. Schematic diagram of solar mechanical compression refrigeration.


232 K.F. Fong et al. / Solar Energy 84 (2010) 227–244

cycle would cause heat loss from the hot water side to the collectors, hot water pump, hot water tank, auxiliary hea-
cooling water side with no cooling capacity produced. To ter, regenerative water pump, absorption chiller, chilled
avoid this, the chiller would start only when the regenera- water pump, cooling tower, condenser water pump and
tive water temperature entering the boiler exceeded a preset AHU.
minimum driving temperature, which was required by the The model of absorption refrigeration cycle was con-
Rankine-cycle-driven chiller according to the design enter- structed according to the validated single-effect absorption
ing cooling and chilled water conditions. This minimum chiller of Kim and Infante Ferreira (2008). In addition, the
driving temperature was 82 °C in this case. enthalpy of lithium bromide (LiBr) solution was given by
Florides et al. (2003) and the saturated vapour pressure
2.3.3. Control and operation of hot water pump and auxiliary of LiBr solution by Patek and Klomfar (2006). The various
heater properties of refrigerant (water) were determined according
A hot water tank was used to store the heat transferred by to Florides et al. (2003) and Zhang et al. (2003). The gen-
the solar collectors. To ensure fulfillment of the cooling load eral parameters for solar absorption refrigeration are
demand even when the solar irradiation was insufficient, shown in Table 1, while the specific parameters are listed
auxiliary heating was provided inside the storage tank as a in Table 3.
backup heat source. The set point for the auxiliary heater
was 90 °C. The temperature inside the hot water tank might 2.4.2. Control and operation of absorption chiller and
be higher than that at the outlet of the solar collectors in case auxiliary heater
of weak solar irradiation, then the solar collectors would dis- The general provisions for the year-round control and
sipate heat to the surrounding. To prevent this, the hot water operation of hot water pump and auxiliary heater were
pump, which was used to circulate hot water between the similar to those in Section 2.3.3 for solar mechanical com-
solar collectors and the hot water tank, was controlled by pression refrigeration, expect that the auxiliary heater
the hot water pump controller using a hysteresis loop con-
trol. Whenever the temperature at the solar collector outlet
Table 3
exceeded the water temperature inside the hot water tank Modeling parameters for solar absorption refrigeration.
by 5 °C, the hot water pump would be switched on until
Parameter Value
the temperature difference dropped to zero.
ÿ1
Absorber water mass flow rate (kg s ) 2.0
Overall heat transfer value of generator (kW Kÿ1) 4.5
2.4. Solar absorption refrigeration Overall heat transfer value of absorber (kW Kÿ1) 4.5
Overall heat transfer value of solution-to-solution heat exchanger 1.0
2.4.1. Model building (kW Kÿ1)
The entire solar absorption refrigeration was developed Degree of superheat at evaporator outlet (°C) 5.0
Solution volume flow rate at absorber outlet (L sÿ1) 0.1
as shown in Fig. 3. The major equipment covered the solar

Hot Water Pump


Controller
C
Solar T Regenerative
Collectors T Water Pump Cooling
Tower
Hot
Hot Water Pump Water
Tank
Condenser Water Pump
Town Gas Aux.
Supply Heater
T
C
Part-load
Controller
Generator Condenser

Heat
Exchanger
Expansion
Throttling Device
Valve

Absorber Evaporator

Solution Absorption
Pump Chiller

T
Chilled water supply and
return for the AHU serving the
indoor space (same as Fig. 1)

Fig. 3. Schematic diagram of solar absorption refrigeration.


K.F. Fong et al. / Solar Energy 84 (2010) 227–244 233

would be activated through an on/off signal between 12 nents were similar to those of the solar absorption
and 9 °C with hysteresis based on the chilled water return refrigeration, except the adsorption chiller was used
temperature at the chiller. The operation of chiller at an instead.
insufficient driving temperature would cause heat loss from For adsorption refrigeration, an economical adsorption
the hot water side to the condenser water side with no cool- pair is silica gel and water, where silica gel is the adsorbent
ing capacity produced. To avoid this, the chiller would and water the adsorbate (or refrigerant). For the configura-
start only when the auxiliary heater controller was on tion of adsorption cycle in the model, the two-chamber
and the regenerative water temperature entering the chiller design was used. In a working cycle, one of the chambers
exceeded a preset minimum, which was 5 °C lower than the was used for adsorption while another for desorption.
minimum driving temperature 67 °C of the absorption chil- The role of these two chambers was interchanged in
ler. When the chiller was on, the condenser water pump another cycle, so as to maintain a pseudo-continuous pro-
and the regenerative water pump would be in operation duction of refrigeration effect. Model development of the
as well. solar adsorption refrigeration cycle was based on the vali-
dated model of Cho and Kim (1992). The general modeling
2.4.3. Part-load control of absorption chiller parameters for the solar adsorption refrigeration are shown
To reduce the utilization of auxiliary heating, the hot in Table 1, while the specific parameters are outlined in
water temperature used to drive the chiller should be main- Table 4.
tained as low as possible. This was achieved by adopting a
part-load control for the chiller which modulated the set 2.5.2. Control and operation of solar adsorption refrigeration
point of the auxiliary heater between the minimum driving The details of control and operation for the chiller and
temperature 67 °C and the ceiling temperature 90 °C in 5 the associated equipment were same as those in Section
steps. The set point started at 67 °C when the chilled water 2.2.3 for the solar electric compression refrigeration. The
return temperature at the chiller exceeded 12 °C, and provisions of control and operation of the hot water pump
stepped up if the water temperature dropped less than
Table 4
1 °C after one simulation time step. Conversely, when the Modeling parameters for solar adsorption refrigeration.
chilled water return temperature was below 12 °C, the set
Parameter Value
point would step down if the water temperature dropped
Adsorption chamber water mass flow rate (kg sÿ1) 3.0
more than 1 °C after one simulation time step. By employ-
Specific heat capacity of metal (kJ kgÿ1 Kÿ1) 0.448
ing this part-load control, the chiller capacity could be reg- Specific heat capacity of silica gel (kJ kgÿ1 Kÿ1) 0.92
ulated through the control of driving temperature. Specific heat of adsorption (kJ kgÿ1) 2,800
Mass of metal in adsorption/desorption chamber (kg) 120
Mass of silica gel in adsorption/desorption chamber (kg) 80
2.5. Solar adsorption refrigeration Mass of metal in condenser coil (kg) 150
Mass of metal in evaporator coil (kg) 150
2.5.1. Model building Overall heat transfer value of adsorption chamber (kW Kÿ1) 16
The system simulation model of solar adsorption refrig- Overall heat transfer value of desorption chamber (kW Kÿ1) 16
Adsorption/desorption period (s) 360
eration was built according to Fig. 4. The system compo-

Cooling
Tower

T Condenser Water Pump

Hot Water Pump


Controller
Condenser
C Regen.
Solar T Water
Collectors T
Pump
Adsorption Adsorption
Hot
Compart- Compart-
Hot Water Pump Water
ment 1 ment 2
Tank

Town Gas Aux.


Supply Heater Adsorption
Evaporator
Part-load C Chiller
Controller
T

Chilled water supply and return


for the AHU serving the indoor
space (same as Fig. 1)

Fig. 4. Schematic diagram of solar adsorption refrigeration.


234 K.F. Fong et al. / Solar Energy 84 (2010) 227–244

and auxiliary heater were same as those in Section 2.4.2 for Table 5
the solar absorption refrigeration. The part-load control of Modeling parameters for the solar solid desiccant cooling.
the adsorption chiller was same as that in Section 2.4.3 for Parameter Value
the solar absorption refrigeration, except the minimum Process air stream
driving temperature was 64 °C for the solar adsorption Process air evaporative cooler saturation efficiency 90%
refrigeration. Process air evaporative cooler power consumption (kW) 0.1
Exhaust air stream
Exhaust air mass flow rate (kg sÿ1) 3.36
2.6. Solar solid desiccant cooling Exhaust air fan power (kW) 3.36
Exhaust air fan efficiency 70%
2.6.1. Model building Exhaust air evaporative cooler saturation efficiency 90%
In the solar solid desiccant cooling, the system simula- Exhaust air evaporative cooler power consumption (kW) 0.1
Overall heat transfer value of regenerative heating coil (kW Kÿ1) 9.72
tion model was constructed based on Fig. 5. It encom-
passed the solar collectors, hot water pump, hot water Heat wheel
Temperature effectiveness of heat wheel 80%
tank, auxiliary heater, regenerative water pump, regenera-
Heat wheel power consumption (kW) 0.2
tive heating coil, desiccant wheel, heat wheel, evaporative
coolers, supply air fan and exhaust air fan. Desiccant wheel
Specific heat capacity of air (kJ kgÿ1Kÿ1) 1.005
The model of desiccant wheel was developed based on the Specific heat capacity of metal (kJ kgÿ1Kÿ1) 0.9
validated results of Zhang et al. (2003), the equilibrium rela- Specific heat capacity of silica gel (kJ kgÿ1Kÿ1) 0.921
tive humidity was given by Pesaran and Mills (1987a,b). The Specific latent heat of vaporization of water (kJ kgÿ1) 2.400
general parameters for modeling the solar solid desiccant Thermal conductivity of air (W mÿ1 Kÿ1) 0.0321
cooling are shown in Table 1, while the specific parameters Density of air (kg mÿ3) 1.2
Diameter of desiccant wheel (m) 2
are presented in Table 5. Width of desiccant wheel (m) 0.2
Effective area ratio of desiccant wheel 0.744
Fraction of wheel area for regeneration 0.5
2.6.2. Additional design considerations of solar desiccant
Desiccant wheel speed (rph) 12
cooling Desiccant wheel power consumption (kW) 0.2
Unlike the other four types of solar cooling systems con-
taining chillers, the solar desiccant cooling would directly
treat the fresh air for air-conditioning purpose. Although perature at the design conditions. The flow rate of the
the indoor space was identical, the total cooling load was regenerative water was selected so that the temperature
increased to 109 kW because ventilation load was much drop after passing through the regenerative heating coil
higher for the process air. The desiccant wheel was sized would be 10 °C. The flow rate of the hot water pump
according to the design regeneration temperature of was same as that of the regenerative water pump. The
85 °C, with 5 °C temperature difference between the enter- capacity of the auxiliary heater was 89 kW based on the
ing regenerative water and the leaving exhaust air of the worst situation that all the regenerative heat needed by
regenerative heating coil. The desiccant wheel speed was the desiccant cooling system was provided by the auxiliary
set to a value which would yield the lowest supply air tem- heater.

Controller for Hot


Water Pump
C
Solar T Regenerative
Collectors Water Pump
T
Hot
T Hot Water Pump Water
Outdoor Tank
Air
Town Gas Aux.
Supply Heater
C
Auxiliary Heater
T
Controller

Evapor-
Exhaust Air
ative
Cooler 2
Exhaust Regenerative Supply
Air Fan Heating Coil Air Fan
Evapor-
Outdoor Air ative T
Desiccant Heat Cooler 1
Wheel Wheel
Indoor Space

Fig. 5. Schematic diagram of solar solid desiccant cooling.


K.F. Fong et al. / Solar Energy 84 (2010) 227–244 235

2.6.3. Control and operation of solar desiccant cooling there are flat plate collectors, evacuated tubes and para-
The auxiliary heater was controlled through a desiccant bolic concentrators. By accounting for these installation
cooling controller that provided an on/off signal between strategies and options of solar collectors, a variety of com-
26.5 and 24.5 °C with hysteresis based on the space temper- binations can be formed for the five types of solar cooling
ature. Moreover, the auxiliary heater only started when the systems as shown in Table 6.
ambient temperature was above 20 °C. The operation of From Table 6, there are altogether 26 combinations of
the desiccant cooling system at insufficient regeneration different installation strategies and solar collectors for the
temperature would cause heat loss from the hot water side five types of solar cooling systems. It is an extensive simu-
to the exhaust air with no cooling capacity produced. To lation study about the solar cooling application in the con-
avoid this, the desiccant cooling system would run only text of year-round operation. Since certain installation
when the following three conditions were satisfied: strategy and solar collector would have relatively inferior
effect to the system performances, a preliminary study
 The auxiliary heater controller was on. was carried out in order to remove the unnecessary combi-
 The ambient temperature was above 20 °C. nations in the later full study. As a result, the following
 The regenerative water temperature entering the regen- comparison works would be conducted first in the preli-
erative heating coil exceeded a preset minimum, which minary study:
was 5 °C lower than the minimum driving temperature
60 °C.  Effect of installation strategies: The performances of
solar electric compression refrigeration and solar
When the desiccant cooling system was on, the regener- absorption refrigeration with flat plate collectors based
ative water pump, desiccant wheel, heat wheel and the on the two installation strategies would be compared,
evaporative coolers would operate. The supply and exhaust so as to evaluate the more feasible option in view of
air fans would be in operation according to the daily occu- solar electric gain and solar thermal gain in the subtrop-
pancy schedule. ical region.
 Effect of solar collectors: The performances of the flat
3. Framework of comparative study plate collectors, evacuated tubes and parabolic concen-
trators applied to the absorption refrigeration would
3.1. Preliminary study and full study be compared, so as to find out the more feasible types
of solar collectors for the subtropical region.
In order to have a thorough comparative study of differ-
ent solar cooling systems, the types of installation strategies The models of solar collectors were based on those in
and solar collectors were also included. Generally, there are the TESS library for TRNSYS. The coefficients of the effi-
two feasible installation strategies for the solar collectors. ciency curve for the evacuated tubes were based on
The first one is to make use of any unobstructed open (Hochschul Rapperswil of Switzerland, 1997). To model
space, practically on the roof, to mount the solar collectors the building-integrated approach, the component model
at an inclination angle same as the local latitude. The sec- of building-integrated photovoltaic/thermal system (TESS
ond strategy is to apply the building-integrated approach Type 567, 2006) from TESS was used to interact to the
by mounting the solar collectors vertically on the building indoor space of the building system under study. In the
facades. For the high-rise buildings commonly found in building-integrated strategy, the solar collectors were verti-
Hong Kong, the former would be limited by the usable cally installed to the east, south and west. The north side
roof area, while the latter can make use of the available was not used since this study was about the subtropical
area of the external facades for solar collector installation. region in the northern hemisphere. In addition, three direc-
About the common choices of solar thermal collectors, tions were used instead of one because the three sides of

Table 6
Possible combinations of installation strategies and solar collectors for the solar cooling systems.
Installation Photovoltaic Flat plate Evacuated Parabolic
strategy panels collectors tubes concentrators
p
Solar electric compression refrigeration Roof-mounted
p
Building-integrated
p p p
Solar mechanical compression Roof-mounted
p p p
refrigeration Building-integrated
p p p
Solar absorption refrigeration Roof-mounted
p p p
Building-integrated
p p p
Solar adsorption refrigeration Roof-mounted
p p p
Building-integrated
p p p
Solar solid desiccant cooling Roof-mounted
p p p
Building-integrated
236 K.F. Fong et al. / Solar Energy 84 (2010) 227–244

external facades were all needed in order to sufficiently Qsolar


SF thermal ¼ ð2Þ
accommodate the design mounting area of solar collectors. Qsolar þ Qaux

3.2. Solar cooling systems against conventional electric- where SFthermal = solar fraction of solar thermal cooling
driven vapour compression refrigeration system; Qsolar = solar thermal gain from the solar collectors
(kW); Qaux = heat output of auxiliary heater (kW).
In order to promote wider application of the solar cooling The solar fraction would be averaged in monthly or
systems, it is necessary to benchmark their performances yearly basis, it is based on the SF acquired at each time step
against the conventional vapour compression refrigeration when solar thermal gain is produced due to the hot water
solely energized by electricity. There are two kinds of chiller pump in operation. The monthly averaged SF and yearly
plants commonly applied in the subtropical region: averaged SF are formulated below:
Xm
SF i
 The air-cooled packaged vapour compression refrigera- Monthly averaged SF ¼ ð3Þ
tion; and i¼1
m
 The water-cooled vapour compression refrigeration.
where m = total number of time steps in a month for hot
Based on the small and medium applications, the coeffi- water pump in operation.
cient of performance of air-cooled chiller was 2.4 and that Xy
SF i
of water-cooled chiller was 3.0 in this study. The general Yearly averaged SF ¼ ð4Þ
y
modeling parameters of the chillers and the associated i¼1

equipment are same as those in Table 1, so as to have a


common ground for comparison. where y = total number of time steps in a year for hot
water pump in operation.
4. Performance indicators For the dynamic simulation of the year-round operation
in this study, the simulation time step was 6 min as men-
A number of performance indicators are needed in order tioned in Section 2.1. However the number of time steps
to have a full understanding of different aspects of the solar involved finding the averaged solar fraction depended on
cooling systems. These performance indicators include the whether the related equipment of the solar cooling system
solar fraction, coefficient of performance, solar thermal was really in operation, since the indoor conditions could
gain and primary energy consumption. A year-round per- still be maintained at certain climatic and loading condi-
spective of these performance indicators was used, so that tions in a year. The number of time steps involved to deter-
the system performances could take into account the mine the monthly/yearly averaged solar fraction is based
changing loading and climatic conditions in this compara- on the total number of time steps for hot water pump in
tive study. The definitions, particularly the first two, would operation only. As discussed in Section 2.3.3, a hot water
depend on the types of solar cooling systems, therefore they pump controller was included to activate the hot water
are clearly described in this section. pump whenever the temperature at the solar collector out-
let exceeded the water temperature inside the hot water
tank. In this situation, solar thermal gain could be actually
4.1. Solar fraction, SF
acquired from the solar collectors. Therefore the solar frac-
tion was counted and averaged when the hot water pump
Solar fraction is commonly used to evaluate the perfor-
was in operation.
mance of a solar cooling system. The fraction shows the
portion of solar energy contribution as compared to the
total energy required to drive the refrigeration part of solar 4.2. Coefficient of performance, COP
cooling system. This total amount of energy input is com-
monly supplemented by the auxiliary provision. In this Coefficient of performance is an useful performance
study, the solar fractions are defined below. indicator to evaluate the key air-conditioning equipment
For solar electric compression refrigeration, of the solar cooling systems. In the solar electric refrigera-
W solar tion, solar mechanical refrigeration, solar absorption
SF elect ¼ ð1Þ refrigeration and solar adsorption refrigeration, the key
W solar þ W aux
air-conditioning equipment are the respective chillers. In
where SFelect = solar fraction of solar electric compression the solar solid desiccant cooling, it is the desiccant wheel.
refrigeration; Wsolar = solar electric gain from PV panels Therefore the COP of different solar cooling systems are
(kW); Waux = auxiliary electrical power from the city defined below.
power grid (kW). For solar electric compression refrigeration,
For the solar thermal cooling system including solar
mechanical compression refrigeration, absorption refriger- Qe
COP elect ¼ ð5Þ
ation, adsorption refrigeration and desiccant cooling, W
K.F. Fong et al. / Solar Energy 84 (2010) 227–244 237

where COPelect = COP of solar electric compression refrig- heater controller was on; the ambient temperature was
eration; Qe = refrigeration effect (kW); W = power input to above 20 °C and the regenerative water temperature enter-
DC motor (kW). ing the regenerative heating coil exceeded a preset
For solar mechanical compression refrigeration. minimum.
Qe
COP mech ¼ ð6Þ 4.3. Solar thermal gain, Gsolar
Qboil
where COPmech = COP of solar mechanical compression Solar thermal gain is the useful energy acquired through
refrigeration; Qboil = heat input to boiler of Rackine cycle the solar collectors to drive a solar cooling system. It is not
(kW). just related to the efficiency of the solar collectors, but also
For solar absorption refrigeration, the nature of energy demand of the solar cooling system. In
Qe the four solar thermal cooling systems under study, it is
COP ab ¼ ð7Þ about the gain of thermal energy from the flat plate collec-
Qgen
tors, evacuated tubes or parabolic concentrators. The
where COPab = COP of solar absorption refrigeration; amount of gain is related to both the type of solar collec-
Qgen = heat input to generator (kW). tors and system operation due to different climatic and
For solar adsorption refrigeration, loading conditions. Therefore it is also one of the perfor-
Qe mance indicators to evaluate the effectiveness of a solar
COP ad ¼ ð8Þ cooling system with certain type of solar collector.
Qde
However in the solar electric refrigeration, the PV panels
where COPad = COP of solar adsorption refrigeration; are used and the solar electric gain Gelect is produced
Qde = heat input for desorption (kW). instead of solar thermal gain. Since the solar electric gain
For solar solid desiccant cooling, and solar thermal gain are different in nature, a term called
m_ a ðho ÿ hs Þ solar thermal gain equivalent Gsolar,eq is defined for com-
COP desi ¼ ð9Þ parison purpose:
Qregen
Gsolar;eq ¼ Gelect =ge ð12Þ
where COPdesi = COP of solar solid desiccant cooling; m_ a =
mass flow rate of supply air (kg/s); ho = specific enthalpy of where ge is the energy efficiency used to convert the electri-
outdoor air (kW/kg); hs = specific enthalpy of supply air cal energy into the primary energy, and it is taken to be
(kW/kg); Qregen = heat input for regeneration (kW). 33% in this study.
The COP would be averaged in monthly or yearly basis,
it is based on the COP acquired at the time step when the 4.4. Primary energy consumption, Ep
key air-conditioning equipment (chiller or desiccant wheel)
is in operation. The monthly averaged COP and yearly When calculating the primary energy consumption, it is
averaged COP are formulated below: in a system approach. All the parasitic energy consump-
Xm0 tions from the associated equipment were considered upon
COP i those of the main equipment. Since two different types of
Monthly averaged COP ¼ ð10Þ
i¼1
m0 energy sources, electricity and gas, were involved, their
where m0 = total number of time steps in a month for chil- energy efficiencies were assumed below:
ler or desiccant wheel in operation.
 energy efficiency for electrical energy converted into pri-
Xy0
COP i mary energy, ge = 33%;
Yearly averaged COP ¼ ð11Þ  energy efficiency for gas energy converted into primary
i¼1
y0
energy, gg = 90%.
where y0 = total number of time steps in a year for chiller
or desiccant wheel in operation. The former was used to determine the primary energy
For the solar electric compression refrigeration, solar input of the electric-driven equipment, while the latter for
mechanical compression refrigeration, solar absorption the auxiliary heater operated by burning gas. Ep is there-
refrigeration and solar adsorption refrigeration, the total fore determined in different ways for the corresponding
number of time steps involved were based on the chiller solar cooling systems as follows.
in operation. As discussed in Section 2.2.3, the operation For solar electric compression refrigeration,
of the chiller was determined by a controller that provided
Ep;elect ¼ ðW aux þ W parasitic þ W AHU Þ=ge ð13Þ
an on/off signal based on the chilled water return tempera-
ture. For the solar desiccant cooling, the total number of where Ep,elect = primary energy consumption of solar elec-
time steps was counted according to the desiccant wheel tric compression refrigeration (kWh); Wparasitic = electrical
in operation. As mentioned in Section 2.6.3, the desiccant energy consumption of parasitic equipment, including
cooling system would be activated only when the auxiliary pumps and cooling tower (kWh); WAHU = electrical energy
238 K.F. Fong et al. / Solar Energy 84 (2010) 227–244

consumption of AHU (kWh).For solar mechanical com- the year-round totals of primary energy consumptions were
pression refrigeration, absorption refrigeration and adsorp- increased by 13.9% and 35.7% for the building-integrated
tion refrigeration, solar electric refrigeration and solar absorption refrigera-
tion respectively. The COP was about the same or slightly
Ep;thermal ¼ Qaux =gg þ ðW parasitic þ W AHU Þ=ge ð14Þ
increased, it was not due to the improvement of the build-
where Ep,thermal = primary energy consumption of solar ing-integrated approach, but just because of the auxiliary
thermal refrigeration (kWh). provisions operated whenever in need. Therefore, COP
For solar solid desiccant cooling, was not effective to reflect the performance of solar cooling
systems in this case.
Ep;desi ¼ Qaux =gg þ ðW para;desi þ W fan Þ=ge ð15Þ The poorer result of the building-integrated solar
where Ep,desi = primary energy consumption of solar solid absorption system was further due to the performances
desiccant cooling (kWh); Wpara,desi = electrical energy con- of the solar collectors installed at the three different direc-
sumption of parasitic equipment, including pumps, desic- tions. Based on the typical design, the hot water pump
cant wheel, heat wheel and evaporative coolers (kWh); would only operate when the water leaving temperature
Wfan = electrical energy consumption of supply air fan of the solar collectors was higher than that of the hot water
and exhaust air fan (kWh). storage tank. As the size of solar collector facing to each
direction was reduced to one-third, the hot water tempera-
ture that could be achieved would be lower, causing signif-
5. Preliminary study icant drops of solar fraction and solar thermal gain.
As a result, the building-integrated strategy was not
5.1. Effect of installation strategies of solar collectors involved in the ongoing comparative study for different
solar cooling systems. Such strategy is only recommended
The results of the two feasible installation strategies for when the roof or the available open space is not enough
the solar electric compression refrigeration and solar to accommodate the required number of solar collectors.
absorption refrigeration are summarized in Table 7. The
performances were evaluated on a year-round basis,
including the four performance indicators and the averaged 5.2. Effect of solar collectors
space cooling load. Due to the fact that the solar collectors
integrated to the building facades can reduce the transmis- Table 8 summarizes the year-round performances of
sion heat gains, it is found that the year-round averaged the solar absorption refrigeration system with different
space cooling load was reduced by 12.7% and 9.2% for types of solar collectors, including the flat plate collectors,
the building-integrated solar electric compression refrigera- evacuated tubes and parabolic concentrators. The four per-
tion and solar absorption refrigeration, as compared to the formance indicators were involved to evaluate the effective-
roof-mounted strategy. However, both cases could not out- ness of solar collectors.
weigh the significant drop of the corresponding solar frac- The type of flat plate collectors was used as the baseline
tions at 50.2% and 82.3%, as well as the decrease of the of comparison. It can be observed that both the evacuated
respective solar thermal gains at 65.8% and 83%. Finally tubes and parabolic concentrators had higher solar fraction

Table 7
Year-round performances of the two feasible installation strategies for the solar electric compression refrigeration and solar absorption refrigeration.
Solar cooling system Installation Year-round Year-round Year-round Year-round Year-round
strategy averaged space averaged SF averaged COP total of total of Ep (kWh)
load (kW) Gsolar (kWh)
Solar electric compression refrigeration Roof-mounted 10.99 0.687 4.599 31,678 44,589
Building-integrated 9.59 (;12.7%) 0.342 (;50.2%) 4.658 ("1.3%) 10,844 (;65.8%) 50,781 ("13.9%)
Solar absorption refrigeration Roof-mounted 11.11 0.497 0.769 37,234 72,797
Building-integrated 10.09 (;9.2%) 0.088 (;82.3%) 0.763 (;0.8%) 6,326 (;83.0%) 98,787 ("35.7%)

Table 8
Year-round performances of different types of solar collectors for the solar absorption refrigeration.
Type of solar Year-round averaged Year-round averaged Year-round total of Gsolar Year-round total of Ep (kWh)
collectors SF COP (kWh)
Flat plate 0.497 0.769 37,234 72,797
Evacuated tubes 0.818 ("64.6%) 0.763 (;0.8%) 67,383 ("81.0%) 49,425 (;32.1%)
Parabolic 0.596 ("19.9%) 0.777 ("1.0%) 47,929 ("28.7%) 67,450 (;7.3% vs. flat plate; ;36.5% vs.
concentrators evacuated tubes)
K.F. Fong et al. / Solar Energy 84 (2010) 227–244 239

and solar thermal gain, and lower primary energy 6.1. Overall year-round performances of different solar
consumption. However, since the parabolic concentrators cooling systems
only responded to the direct solar irradiation, their perfor-
mances were lower than that of the evacuated tubes despite The performance results of different solar cooling sys-
a higher collector temperature could be achieved. In the tems are summarized in Table 10. The four performance
year-round total of primary energy consumption, the par- indicators were involved in the analysis. The solar electric
abolic concentrators were 36.5% worse than the evacuated compression refrigeration was used as the baseline of this
tubes, and just 7.3% better than the flat plate collectors. In comparative study, except its COP due to different nature.
fact, the direct solar irradiation is only 42.4% of the total in In terms of the year-round total of primary energy con-
the subtropical Hong Kong (Fong et al., 2007). Moreover, sumption, the order of the solar cooling systems from the
within the same collector installation area, the total best to the worst is found as follows:
collector tubes of parabolic concentrators would be
fewer due to the substantial coverage of the parabolic  solar electric compression refrigeration
mirrors.  solar absorption refrigeration
The year-round averaged COP varies only within a very  solar adsorption refrigeration
narrow range, from 0.763 to 0.777, for these three types of  solar solid desiccant cooling
solar collectors. The presence of the auxiliary heater tended  solar mechanical compression refrigeration
to minimize the variation of effectiveness of different types
of solar collectors to the absorption chiller. In fact, the solar electric compression refrigeration and
As a result, the option of parabolic concentrators was solar absorption with evacuated tubes had comparable per-
excluded from the following comparative study, since it formances. Although the primary energy consumption of
could not have a distinctive performance as compared to the solar absorption refrigeration was 10.8% higher, its
the evacuated tubes in the subtropical Hong Kong. How- solar faction was 19.1% better. It is also observed that this
ever, the flat plate collectors were still involved because order is independent to the types of solar collectors
of their popularity and simple configuration. adopted. For the same kind of solar thermal cooling sys-
tem, it is clear that the evacuated tubes had better perfor-
6. Comparative study of solar cooling systems mances than the flat plate collectors. In this study, the
PV efficiency is based on the typical information of the
From the preliminary performance study, the options of PV component of TRNSYS (2006). The simulation result
building-integrated strategy and parabolic concentrators showed the yearly averaged efficiency was around 10%
were removed, therefore the variety of combinations for according to the climatic conditions of Hong Kong.
the full comparative study could be trimmed down as Although the solar electric compression refrigeration has
shown in Table 9. the best overall performances, particularly primary energy

Table 9
Comparative study of different solar cooling systems.
Roof-mounted PV panels Roof-mounted flat plate collectors Roof-mounted evacuated tubes
p
Solar electric compression refrigeration
p p
Solar mechanical compression refrigeration
p p
Solar absorption refrigeration
p p
Solar adsorption refrigeration
p p
Solar solid desiccant cooling

Table 10
Year-round performances of different solar cooling systems.
Solar cooling system Type of solar Year-round Year-round Year-round total of Gsolar Year-round total of Ep
collector averaged SF averaged COP (kWh) (kWh)
Solar electric compression PV 0.687 4.599 31,678, Gsolar,eq 44,589
refrigeration
Solar mechanical compression Flat plate 0.046 (;93.9%) 0.102 26,911 (;15.0%) 629,897 ("1313%)
refrigeration Evacuated tubes 0.112 (;83.7%) 0.102 65,513 ("107%) 592,292 ("1228%)
Solar absorption refrigeration Flat plate 0.497 (;27.7%) 0.769 37,234 ("17.5%) 72,797 ("63.3%)
Evacuated tubes 0.818 ("19.1%) 0.763 67,383 ("113%) 49,425 ("10.8%)
Solar adsorption refrigeration Flat plate 0.313 (;54.4%) 0.435 40,421 ("27.6%) 128,715 ("189%)
Evacuated tubes 0.577 (;16.0%) 0.437 72,885 ("130%) 93,692 ("110%)
Solar solid desiccant cooling Flat plate 0.336 (;51.1%) 1.066 40,435 ("27.6%) 149,356 ("235%)
Evacuated tubes 0.552 (;19.7%) 1.059 71,832 ("127%) 128,052 ("187%)
240 K.F. Fong et al. / Solar Energy 84 (2010) 227–244

consumption, they were acquired based on a typical PV mechanical compression refrigeration had exceptionally
panel, not due to a certain product with exceptionally high high year-round primary energy consumption, it was
efficiency. excluded in this part of evaluation. From Table 11, it is
The solar mechanical compression refrigeration had found that the average room air conditions of various solar
exceptionally high year-round primary energy consump- cooling systems are close to the design conditions at
tion, showing that the auxiliary heater, instead of solar col- 25.5 °C and 60% RH, showing that they could provide sat-
lectors, was the essential contributor for driving the isfactory indoor environment for the occupants. The per-
Rankine cycle of the heat engine. The narrow range of formances of solar electric compression refrigeration,
operative driving temperature of the Rankine-cycle-driven solar absorption refrigeration and solar adsorption refrig-
chiller maintained a high-temperature demand and the eration are very close (24.4–24.9 °C and 59–61%RH)
part-load control became useless. The solar collectors regardless the type of solar collectors. It is because they
could not have effective contribution due to its relatively provided air-conditioning in a similar way, by using the
low-temperature thermal gain. chilled water supply to cooling coil for cooling and dehu-
The solar adsorption refrigeration was the third in view midification purposes. Solar desiccant cooling had slight
of the primary energy consumption. Although its primary difference of room temperature and relative humidity from
energy consumption was about double of that of solar elec- the other three types because the supply air was handled in
tric compression refrigeration or solar absorption refriger- another way, by using the desiccant wheel, heat wheel and
ation, continual research is being carried out for more evaporative cooler.
effective adsorbent–adsorbate pairs (Fan et al., 2007; Khat-
tab, 2004; Wang et al., 2005), instead of the typical pair of 6.2. Monthly performances of different solar cooling systems
silica gel and water used in this model. On the other hand,
there are emerging configurations for enhancing the effec- Apart from the year-round performances, it is also wor-
tiveness of adsorption–desorption process (Wang et al., thy to understand the monthly variations of the perfor-
2005, 2008; Voyiatzis et al.,2007). Therefore the application mance indicators of different solar cooling systems. The
potential of solar adsorption refrigeration still exists. changes would be according to the climatic conditions,
The solar solid desiccant cooling had high year-round loading conditions and features of solar cooling systems.
primary energy consumption, but it was different from The annual profile of each performance indicator is there-
the other solar cooling systems because it had to handle fore presented and discussed in this section.
the extra ventilation load due to the inherent nature of full
fresh air design. Therefore the total cooling capacity of des- 6.2.1. Monthly averaged solar fraction
iccant cooling system would be much higher than those of Fig. 6 shows the annual profiles of solar fraction. It is
the other four solar cooling systems, hence higher primary found that the solar fraction was commonly low from
energy consumption. In addition, the parasitic energy con- May to September, which is the typical summer period
sumption was comparatively high, particularly the supply and the cooling load is around the peak. On the other
air fan and exhaust air fan. The feature of solar desiccant hand, the solar fraction was usually high from December
cooling in effect is more than only to satisfy the required to February, which is the typical winter season. Generally,
cooling load, it is also capable to supply the fresh air the solar fraction was less than one, indicating that just the
amount far above the minimum requirement of a func- solar electric or thermal gain acquired from the solar col-
tional area. This unique merit can guarantee a good indoor lectors was not sufficient to fully drive the respective solar
air quality and ventilation effectiveness. Therefore the cooling system (except the solar desiccant cooling), so the
application potential of solar solid desiccant cooling is still auxiliary provision was inevitably activated to supplement
valid. the deficit. As a whole, the solar absorption refrigeration
The overall performances of the average room and sup- had the highest solar fraction; the solar electric compres-
ply air conditions are summarized in Table 11. As the solar sion refrigeration the second; the solar adsorption refriger-

Table 11
Yearly average room and supply air conditions of various solar cooling systems.
Solar cooling system Type of solar Average room Average room Average supply air Average supply air
collector temperature (°C) relative humidity (%) temperature (°C) humidity ratio
(kg/kg dry air)
Solar electric compression refrigeration PV 24.9 61 19.4 11.7  10ÿ3
Solar absorption refrigeration Flat plate 24.5 59 19.7 11.1  10ÿ3
Evacuated tubes 24.4 59 19.6 11.0  10ÿ3
Solar adsorption refrigeration Flat plate 24.5 59 19.7 11.2  10ÿ3
Evacuated tubes 24.5 59 19.7 11.1  10ÿ3
Solar solid desiccant cooling Flat plate 23.6 67 20.5 12.2  10ÿ3
Evacuated tubes 23.6 67 20.4 12.2  10ÿ3
K.F. Fong et al. / Solar Energy 84 (2010) 227–244 241

Fig. 6. Annual profiles of solar fraction of different solar cooling systems. (Abbreviation: AB, absorption refrigeration; AD, adsorption refrigeration; DE,
desiccant cooling; EL, electric compression refrigeration; ME, mechanical compression refrigeration; et, evacuated tubes; fp, flat plate collectors; pv,
photovoltaic panels).

ation the third; the solar desiccant cooling the fourth; and (May to September), the order from the best to the worst
the solar mechanical compression refrigeration the last. is found as follows:
For the solar solid desiccant cooling, the solar fraction
was unity from December to February. This shows that  solar solid desiccant cooling
the solar thermal gain from solar collectors was enough  solar absorption refrigeration
to drive the system in such period, and free air cooling  solar adsorption refrigeration
might also be involved. On the other hand, the extremely  solar mechanical compression refrigeration
low solar fraction of solar mechanical compression refrig-
eration shows its heavy reliance on auxiliary heating and The solar solid desiccant cooling could have COP
very weak support from the solar collectors. greater than unity in such period. From Eq. (9), its COP
definition is different to the three solar thermal refrigera-
tion systems. The numerator is the enthalpy change of
6.2.2. Monthly averaged COP the process air, instead of refrigeration effect. For the three
Fig. 7 shows the monthly averaged COP of different solar thermal refrigeration systems, their patterns are sim-
solar cooling systems. The solar electric compression refrig- ilar and nearly flat throughout a year. This is because the
eration is excluded from this analysis because the nature of support from auxiliary heating would not stop whenever
its COP is different from that of the four solar thermal the required driving temperature is not sufficient. On the
cooling systems. For the COP in the summer period other hand, suitable control and operation of auxiliary

Fig. 7. Annual profiles of coefficient of performance of different solar cooling systems. (Abbreviation: AB, absorption refrigeration; AD, adsorption
refrigeration; DE, desiccant cooling; EL, electric compression refrigeration; ME, mechanical compression refrigeration; et, evacuated tubes; fp, flat plate
collectors; pv, photovoltaic panels).
242 K.F. Fong et al. / Solar Energy 84 (2010) 227–244

heating would lead to the COP independent to the type of in which the profile is also high from July to October, while
solar collectors. It is found that both flat plate collectors low in January and February (Fong et al., 2007).
and evacuated tubes had similar COP for the same type
of solar thermal systems.
6.2.4. Monthly total of primary energy consumption
For solar solid desiccant cooling system, the COP was
Fig. 9 shows the annual profiles of the primary energy
lower than one in the months other than the summer per-
consumption of different solar cooling systems. The perfor-
iod, in between 0.67 and 0.98. In January and February,
mance of the solar mechanical compression refrigeration
the COP was zero. It is because the operation mode of free
was excluded from the analysis since its year-round total
air cooling was effective to handle the space cooling load,
was far more higher than the other types of solar cooling
so the desiccant wheel and related equipment were not
systems as shown in Table 10. It is clear that the solar elec-
needed to operate.
tric compression refrigeration has the lowest primary
energy consumption. In fact, the annual profile of solar
absorption refrigeration with evacuated tubes was close
6.2.3. Monthly total of solar thermal gain
to that of the solar electric compression refrigeration, the
Fig. 8 shows the variation of monthly averaged solar
slight difference was mainly due to the additional energy
thermal gain of different solar cooling systems. It is found
required for the parasitic equipment for the solar thermal
that there are two sets of curves clustered together due to
part, particularly the hot water pump and regenerative
the type of solar collectors. The set of performance curves
pump. Generally these profiles are similar to the annual
of solar cooling systems using evacuated tubes are higher
cooling load variation. The primary energy consumption
than those using flat plate collectors. This clearly shows
is high from June to August, which is coincident to the
the evacuated tubes have a more superior solar thermal
peak cooling load demand in the summer period. The pri-
gain compared to the flat plate collectors. It is noted that
mary energy consumption is low from December to Febru-
the performance curve of solar electric compression refrig-
ary, which is the typical winter.
eration is within the lower cluster. Although it was trans-
formed to Gsolar,eq due to its nature of solar electric gain,
its performance was not as low as those systems with flat 6.3. Performances of solar cooling systems against
plate collectors. It can be found that the yearly averaged conventional electric-driven vapour compression refrigeration
solar fraction of solar electric compression refrigeration is
0.687, which is relatively high in Table 10, showing that Application potential of solar cooling systems depends on
the solar electric gain was quite enough to drive the refrig- their performances against the conventional electric-driven
eration cycle. vapour compression refrigeration, particularly the primary
Generally the solar thermal gain was high in the period energy consumption. Both the air-cooled and water-cooled
of July to October, which covers from the mid-summer to vapour compression chiller plants were involved in this
early autumn. It is not fully matched to the solar fraction study, it is found that their year-round totals of primary
profiles, which are mainly affected by the cooling load energy consumption were 86,289 kWh and 79,138 kWh,
demand in the hot season from May to September. The respectively, and their annual profiles are plotted together
variation of solar thermal gain primarily follows the avail- with those for the solar cooling systems in Fig. 9. From the
ability of solar irradiation in the subtropical Hong Kong, year-round totals of different solar cooling systems shown

Fig. 8. Annual profiles of solar thermal gain of different solar cooling systems. (Abbreviation: AB, absorption refrigeration; AD, adsorption refrigeration;
DE, desiccant cooling; EL, electric compression refrigeration; ME, mechanical compression refrigeration; et, evacuated tubes; fp, flat plate collectors; pv,
photovoltaic panels).
K.F. Fong et al. / Solar Energy 84 (2010) 227–244 243

Fig. 9. Annual profiles of primary energy consumption of different solar cooling systems (including those of electric-driven vapour compression
refrigeration). (Abbreviation: AB, absorption refrigeration; ACVCR, conventional air-cooled vapour compression refrigeration; AD, adsorption
refrigeration; DE, desiccant cooling; EL, electric compression refrigeration; ME, mechanical compression refrigeration; WCVCR, conventional water-
cooled vapour compression refrigeration; et, evacuated tubes; fp, flat plate collectors; pv, photovoltaic panels).

in Table 10, only the solar electric compression refrigeration  solar solid desiccant cooling, and
and solar absorption refrigeration (no matter with evacu-  solar mechanical compression refrigeration.
ated tubes or flat plate collectors) have lower primary energy
consumption. The same results can also be observed in In the preliminary study for the options of installation
Fig. 9. As compared to the conventional air-cooled and strategies and types of solar collectors, the alternatives of
water-cooled refrigeration, the primary energy savings of building-integrated approach and parabolic concentrators
solar electric compression refrigeration were decreased by were found not energy-efficient enough. Although the
48.3% and 43.7%, while those of solar absorption refrigera- building-integrated approach could reduce the space cool-
tion with evacuated tubes by 42.7% and 37.5%, and those of ing load by about 10%, but the overall system perfor-
solar absorption refrigeration with flat plate collectors by mances were worse than the typical roof-mounted
15.6% and 8.0%. Although the solar adsorption refrigera- approach. For the choice of solar collectors, the parabolic
tion with evacuated tubes had relatively higher primary concentrators had the primary energy consumption only
energy consumption, it was only 8.6% and 18.4% above slightly better than the flat plate collectors by 7.3% for
the conventional air-cooled and water-cooled refrigeration the absorption refrigeration, but worse than the evacuated
systems, respectively. tubes by 36.5%.
For the other types of solar cooling systems, their primary If the conventional electric-driven air-cooled and water-
energy consumptions were far higher than those of the two cooled vapour compression chiller plants were included in
conventional electric-driven refrigeration systems. There- the comparison, the solar electric compression refrigeration
fore they do not have any energy saving potential even the and solar absorption refrigeration (with evacuated tubes or
free solar energy is consumed. If life-cycle cost analysis is car- flat plate collectors) were the two types of solar cooling sys-
ried out, both initial and operating costs of such solar cool- tems that could have attractive energy saving potential.
ing systems do not have any advantages at all. The year-round energy savings would be from 15.6% to
48.3% compared to the conventional electric-driven air-
7. Conclusion cooled refrigeration, while 8.0% to 43.7% to the water-
cooled refrigeration. In fact, it is possible to have continual
A comparative study was thoroughly carried out for the improvement in the solar electric compression refrigeration
five types of solar cooling systems for a typical office in the and solar absorption refrigeration, particularly the technol-
subtropical Hong Kong. The results were worked out with ogy advancement of the solar collectors.
the emphasis of suitable system control and operation in The application potential of solar adsorption refrigera-
response to the year-round changing climatic and loading tion with evacuated tubes still exists, since its year-round
conditions. Based on the best year-round total of primary total of primary energy consumption is just slightly higher
energy consumption, the order of the five types of solar than the conventional electric-driven air-cooled vapour
cooling systems is: compression refrigeration. Continual advancement of the
adsorption pair and chiller configuration would enhance
 solar electric compression refrigeration, the use of its distinctive merit of low driving temperature.
 solar absorption refrigeration, Although the solar solid desiccant cooling has high year-
 solar adsorption refrigeration, round primary energy consumption, it is different from the
244 K.F. Fong et al. / Solar Energy 84 (2010) 227–244

other types of solar cooling systems because it has to tackle IPCC Fourth Assessment Report. Intergovernmental Panel on Climate
the high ventilation load from its inherent full fresh air Change, 2007.
Jin, H., Spitler, J.D., 2002. A parameter estimation based model of water-
provision. Subsequently it has high cooling capacity and to-water heat pumps for use in energy calculation programs. ASHRAE
large energy consumption of supply and exhaust air fans. Transactions 108, 3–17.
In fact, the solar desiccant cooling is able to supply Khattab, N.M., 2004. A novel solar-power adsorption refrigeration
excessive fresh air for a functional area, resulting a good module. Applied Thermal Engineering 24, 2747–2760.
indoor air quality and ventilation effectiveness. Therefore Kim, D.S., Infante Ferreira, C.A., 2008. Analytical modeling of steady
state single-effect absorption cycles. International Journal of Refrig-
application potential of solar solid desiccant cooling is still eration 31, 1012–1020.
valid. Kim, D.S., Infante Ferreira, C.A., 2008. Solar refrigeration options – a
This paper also describes the framework and useful per- state-of-the-art review. International Journal of Refrigeration 31, 3–
formance indicators for a comparative study of different 15.
solar cooling systems. The methodology can be applied Hong Kong Energy End-use Data 2008 Full Edition. Electrical and
Mechanical Services Department, Hong Kong SAR Government,
to evaluate the potential systems and configurations, and September 2008.
to benchmark the conventional electric-driven vapour com- Lam, J.C., Tang, H.L., Li, D.H.W., 2008. Seasonal variations in
pression refrigeration. residential and commercial sector electricity consumption in Hong
Kong. Energy 33, 513–523.
Acknowledgement Lee, C.K., 2008. Computer Modelling and Simulation of Geothermal
Heat Pump and Ground-coupled Liquid Desiccant Air Conditioning
Systems in Sub-tropical Regions, Doctoral Thesis, University of Hong
The work described in this paper was fully supported by Kong.
a grant from City University of Hong Kong (Project No. Patek, J., Klomfar, J., 2006. A computationally effective formulation of
7002310). the thermodynamic properties of LiBr–H2O solutions from 273 to 500
K over full composition range. International Journal of Refrigeration
29, 566–578.
References Pesaran, A.A., Mills, A.F., 1987a. Moisture transport in silica gel packed
beds – I. Theoretical study. International Journal of Heat and Mass
Chan, A.L.S., Chow, T.T., Fong, S.K.F., Fong, J.Z., 2006. Generation of Transfer 30, 1037–1049.
a typical meteorological year for Hong Kong. Energy Conversion and Pesaran, A.A., Mills, A.F., 1987b. Moisture transport in silica gel packed
Management 47, 87–96. beds – II. Experimental study. International Journal of Heat and Mass
Cho, S.H., Kim, J.N., 1992. Modeling of a silica gel/water adsorption- Transfer 30, 1051–1060.
cooling system. Energy 17, 829–839. Putten, H., Colonna, P., 2007. Dynamic modeling of steam power cycles.
Code of Practice for Energy Efficiency of Air Conditioning Installations, Part II – simulation of a small simple Rankine cycle system. Applied
Electrical and Mechanical Services Department, Government of Hong Thermal Engineering 27, 2566–2582.
Kong Special Administrative Region, 2007. TESS Library Documentation, Thermal Energy System Specialists, 2006.
Dixon, S.L., 1998. Fluid Mechanics and Thermodynamics of Turboma- TRNSYS 16, a Transient system simulation program, the solar energy
chinery. Buttworth-Heinemann, Boston. laboratory, University of Wisconsin-Madison, 2006.
Eicker, U., 2003. Solar Technologies for Buildings. Wiley, Chichester. TESS Type 567: Glazed Building-integrated Photovoltaic system (Inter-
Fan, Y., Luo, L., Souyri, B., 2007. Review of solar sorption refrigeration acts with Type 56), Green Building Library Technical Reference, TESS
technologies: development and applications. Renewable and Sustain- Library Documentation, Thermal Energy System Specialists, 2006.
able Energy Reviews 11, 1758–1775. Voyiatzis, E., Stefanakis, N., Palyvos, J., Papadopoulos, A., 2007.
Florides, G.A., Kalogirou, S.A., Tassou, S.A., Wrobel, L.C., 2003. Design Computational study of a novel continuous solar adsorption chiller:
and construction of a LiBr–water absorption machine. Energy performance prediction and adsorbent selection. International Journal
Conversion and Management 44, 2483–2508. of Energy Research 31, 931–946.
Fong, S.K.F., Chow, T.T., Lin, J.Z., Chan, A.L.S., 2007. Application Wang, S.G., Wang, R.Z., Li, X.R., 2005. Research and development of
potential of solar thermal technologies for buildings in Hong Kong. In: consolidated adsorbent for adsorption systems. Renewable Energy 30,
Conference Proceedings of Joint Hong Kong and Hangzhou Seminar 1425–1441.
for Sustainable Building, Hangzhou, China, September 2007, pp.118– Wang, D.C., Wang, Y.J., Zhang, J.P., Tian, X.L., Wu, J.Y., 2008.
126. Experimental study of adsorption chiller driven by variable heat
Henning, H-M., 2004. Solar-assisted air-conditioning in buildings. In: A source. Energy Conversion and Management 49, 1063–1073.
Handbook for Planners. Springer-Verlag Wien, New York. Zhai, X.Q., Wang, R.Z., Wu, J.Y., Dai, Y.J., Ma, Q., 2008. Design and
Hochschul Rapperswil of Switzerland, Test Report No. 264, 1997. performance of a solar-powered air-conditioning in a green building.
Hughes, A., 1993. Electric motors and drives: fundamentals. In: Types and Applied Energy 85, 297–311.
Applications. Newnes, Oxford. Zhang, X.J., Dai, Y.J., Wang, R.Z., 2003. A simulation study of heat and
Hwang, Y.H., Radermacher, R., Al Alili, A., Kubo, I., 2008. Review of mass transfer in a honeycombed rotary desiccant dehumidifier.
solar cooling technologies. HVAC&R Research 14, 507–528. Applied Thermal Engineering 23, 989–1003.

You might also like