Accepted Manuscript Advantages of variable driving temperature in solar absorption chiller Karolina Petela, Giampaolo Manfrida, Andrzej Szlek PII:
S0960-1481(17)30682-1
DOI:
10.1016/j.renene.2017.07.060
Reference:
RENE 9035
To appear in:
Renewable Energy
Received Date: 15 December 2016 Revised Date:
9 July 2017
Accepted Date: 12 July 2017
Please cite this article as: Petela K, Manfrida G, Szlek A, Advantages of variable driving temperature in solar absorption chiller, Renewable Energy (2017), doi: 10.1016/j.renene.2017.07.060. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
RI PT
Advantages of variable driving temperature in solar absorption chiller Karolina Petelaa,∗, Giampaolo Manfridab , Andrzej Szleka a Silesian
SC
University of Technology, Institute of Thermal Technology, Konarskiego 22, Gliwice 44-100, Poland b University of Florence, Department of Industrial Engineering, Viale G. Morgagni 40, Firenze 50134, Italy
M AN U
Abstract
The study described in this paper results from the observation that, when dealing with solar-driven absorption chillers, a compromise situation very likely exists between the collector efficiency and the coefficient of performance (COP) of the absorption cycle. An idea of a control strategy pursuing this objective, and thereby increasing the solar fraction for a solar absorption cooling cycle is consequently presented. The advantage of operating the solar chiller on a vari-
TE D
able driving temperature is demonstrated. The theoretical analysis considers a solar chiller operated under moderate climate conditions of Poland. The model of the ammonia-water absorption chiller is described in detail, and the main assumptions of the iterative control procedure are explained. The effects of its application are computationally tested for one location and one type of solar
EP
collector, for representative days of three months during the cooling season. The results of the simulation are compared with the reference case of fixed driving
AC C
temperature. According to the simulation results, it is possible to increase the cooling power yield at various levels, up to 34 W per 1 m2 of solar collector, by adjusting the chiller driving temperature throughout the cooling season. The same, it is possible to more than double the cooling effect for some hours, with respect to fixed-temperature control operation. ∗ Corresponding
author Email address:
[email protected] (Karolina Petela) URL: www.itc.polsl.pl/kpetela (Karolina Petela)
Preprint submitted to Renewable Energy
July 13, 2017
ACCEPTED MANUSCRIPT
RI PT
Keywords: solar cooling, absorption chiller, solar thermal, COP, control strategy
1. Introduction
In times when the environmental policy is becoming more and more rigorous
SC
about the primary energy consumption, one cannot expect that the population's energy demand will decrease, facilitating the savings. On the contrary, accord5
ing to the prognoses, the energy demand will keep increasing. One of the main
M AN U
reasons for that are higher requirements towards provision of thermal comfort. Having access to central heating and domestic hot water is no longer a sign of full thermal comfort satisfaction - this change of societys attitude is also noticeable in moderate climate locations. More and more energy is being consumed 10
for the purposes of air conditioning systems. Currently 30% of energy demand is assigned to cooling end uses on the European Heat and Cold market. Ac-
TE D
cording to the estimates,the share of energy demand for the cooling purposes on the services market of EU27 countries is constantly increasing and will be equal to the heat demand share by 2030.What is more, the Chiller global market has 15
risen from 7 to 8 billion Euro since 2012. [1] The competitively low investment cost of a vapour compression chiller (up to
EP
5 times smaller than that of an absorption chiller [2]) is constantly driving the consumer market to this choice. However, since thermal driven chillers may be powered by heat deriving from solar energy, constant efforts should be undertaken to make these systems more attractive and more frequently installed.
AC C 20
Their diffusion would contribute to a higher share of renewable energy consumption in national and global balance and, consequently, to important savings of non-renewable energy resources. Solar chillers rely on an energy source with time-dependent availability and therefore are usually considered unreliable.
25
Hence, it is important to elaborate methods for reasonable solar thermal cooling plant management, to prove those systems are effective even in locations of poor
2
ACCEPTED MANUSCRIPT
solar resource. Engineering efforts should be made to obtain high values of solar
to the total energy demand for given conditions [3]. 30
RI PT
fraction, defined as the ratio between energy obtained from the solar collector
Today's literature provides a rich overview of methods investigated for the pur-
pose of solar cooling systems reliability improvement. Shirazi et al. (2016) [4] were evaluating an effect on the solar fraction with various configurations of
SC
components within a solar heating and cooling system. Energetic, economic and environmental analysis considered independent cases: one where the solar 35
absorption chiller is backed up by a gas fired heater, one where it is bypassed
M AN U
by a vapour compression chiller and two where the size of the absorption chiller is reduced but cooperates with an auxiliary compression chiller. The analyses showed that the highest primary energy saving is obtained within a system where absorption chiller cooperates with an auxiliary compression chiller. Nonetheless, 40
none of these systems was proven economically profitable, if there is no financial incentive available. On the other hand, Eicker et al. (2015) [5] found that undersizing of the absorption chiller system with respect to the building cooling
TE D
load may be recommendable to increase the machine operating hours. They also admit that solar thermal cooling systems are more viable in hot than in moder45
ate climates. Eicker et al. (2009)[6] discuss main control strategy ideas applied to solar absorption chillers cycles. In the literature, the consideration of driving
EP
temperature most often appears. The majority of physically installed systems are operated at a fixed value of generator temperature. A complex study has been presented by Eicker et al. (2012) [7]. Authors investigated different control strategies applied for absorption chillers equipped with wet, dry or geothermal
AC C
50
heat rejection system- varying the cooling agent temperature and generator inlet temperature, however with the assumption of On/Off controlled solar collector pumps. The generator inlet temperature was adjusted to the temperature in the hot water storage tank. The highest possible value of electrical COP=13 was
55
achieved for the case with geothermal heat rejection. The analysis presented by Syed et al. (2005) [8] concerns a lithium bromide/water chiller and assumes that the nominal outlet temperature from the flat plate collectors field is 90◦ C, 3
ACCEPTED MANUSCRIPT
leading to 80◦ C temperature level at the top of the storage tank, from which
60
RI PT
the absorption chiller is directly driven. Daily average COP was achieved on the level of 0.42. The authors do see a COP improvement potential by considering a lower temperature driven absorption LiBr-H2 O chiller and undersizing the cooling capacity. Li et al. (2014) [9] were trying to evaluate optimum oper-
ation temperatures that would maximize average monthly system efficiency of
65
SC
a double stage Lithium Bromide chiller in a subtropical city. They found that depending on month and meteorological conditions, a different optimal outlet
temperature from the collector results. Lecuona et al. (2009) [10] created an
M AN U
algorithm whose intention was to find an instantaneous optimum hot water temperature based on the phenomenon of inverse relationship between the COP of a single-effect LiBr-water chiller and collector efficiency increase. The authors 70
evaluated an equation for optimum hot water temperature and concluded that to reach the optimum temperature, mass flow rate control is required. The analysis prepared by Albers (2013) [11] presents a strategy of simultaneous control of hot and cooling water temperature leading to decrease of operation cost.
75
TE D
The researchers state that for the purpose of minimizing the specific costs, the control strategy for thermally driven chillers should be independent of solar or district heating. As a result, a 5% reduction of operating cost can be achieved. However, the authors do notice that the economic optimum temperatures (of
EP
hot water and cooling water) will not be the same as for the maximum primary energy savings. An analysis of a complex combined cooling, heating and power 80
plant is offered by Al-Sulaiman (2012) in [12] where an ORC cycle cooperates
AC C
with solar collector field and an absorption chiller; in the same study, it was observed that controlling the ORC pump inlet temperature could be profitable as well.
Taking current accomplishments on this research field under consideration, the
85
aim of the present work is to indicate the potential that weather-dependent adapting of the driving stream temperature may have on the amount of solar cooling power produced by an absorption chiller under specific climate conditions. The idea described in this research, is to find a generator driving temper4
ACCEPTED MANUSCRIPT
ature in every simulation step that may provide a theoretically lower COP but will produce a higher solar collector output. The study focuses on an evacu-
RI PT
90
ated tube collector driven ammonia-water chiller demanding substantial driving
temperatures. The system operates under relatively difficult meteorological conditions (Polish climate with limited radiation).
95
SC
2. Methodology 2.1. Theoretical chiller model
M AN U
For the purposes of this analysis, a design model of a single stage ammoniawater chiller was prepared. The detailed assumptions for the model are presented in subsection 2.4. Once the size of the heat exchangers is defined, an off-design simulation of the chiller operation under variable external conditions 100
is run. A simplified scheme of the cycle is presented in Fig. 1. Cooling water
19
17
5
Q gen
(1) GENERATOR
18
High pressure flow
14
4
EP
Intermediate pressure flow
6
15
(6) PREABSORBER
2
20 7
(7) PUMP
13
1 8
(4) SUBCOOLER
11
9 10
12
(5) EVAPORATOR
21
Qevap
22
Chilled water
AC C
Low pressure flow
3
Q cw
(3) CONDENSER/ ABSORBER
(2) RECTIFIER
16
TE D
(8) Solar collector
Figure 1: Simplified scheme of the single stage ammonia-water absorption chiller.
The components included in the model are: a heat generator (1) co-working
with a rectifier (2), a preabsorber (6), a common vessel for condenser and absorber (3), a circulation pump (7), a subcooler (4) and an evaporator (5). The system operates at 3 pressure levels therefore 3 throttling valves are considered. 5
ACCEPTED MANUSCRIPT
105
High, intermediate and low pressure flows are marked in the Fig. 1, respectively. Ammonia-water solution is warmed up in the generator (1) by the heat rate (
RI PT
Q˙ gen , kW) coming from the solar collector field (8). Incoming heat produces
ammonia vapour, which is subsequently purified in the rectifier (2). The almostpure ammonia vapour flows to condenser (3) where it is cooled down by external 110
cooling water and by the rich solution coming from the absorber (3), producing
SC
waste heat (Q˙ cw , kW). The rich solution is passing through the condenser, the rectifier and the preabsorber before it finally reaches the generator. The aim of this long path is to gradually increase its temperature, reducing the tempera-
115
M AN U
ture difference in the generator. At the same time, it enhances the condensation process of ammonia vapour. The pressure of the condensed ammonia is reduced by the intermediate throttling valve. The subcooler (4) reduces the temperature of the liquid, which is subsequently throttled to a lower evaporator (5) pressure and then evaporated. Low temperature evaporation process is responsible for providing the cooling power (Q˙ evap , kW). After being warmed in subcooler, the 120
ammonia vapour enters the preabsorber (6) where it is absorbed by the weak
TE D
NH3 -H2 O solution flowing from the generator. Rich solution is warmed to saturation conditions in the common condenser/absorber vessel. The design model of the chiller is based on the application of mass and energy conservation equations in accordance with the 1st law of thermodynamics. It is possible to describe every component by the use of balances 1 and 2
AC C
EP
125
n X
m ˙ iin =
i=1
n X
m ˙ iout
(1)
i=1
n n X X (m ˙ i hi )in + Q˙ i = (m ˙ i hi )out i=1
(2)
i=1
Where subscripts in and out denote the inlet or outlet of a stream from the
component, while m ˙ i , kg/s and hi , kJ/kg, are the mass flow rate and the specific 130
enthalpy of given stream, respectively. Q˙ i , kW is the heat transfer rate within
6
ACCEPTED MANUSCRIPT
every system component. Its negative value means a heat loss, while a positive
RI PT
appears as a heat gain from an external source (in case of generator: from warm heat transfer fluid from the collector, evaporator: from chilled water). For the
135
pump, the energy balance also covers mechanical power utilization, namely: Pn Pn ˙ i hi )in + Npump = i=1 (m ˙ i hi )out i=1 (m
Primary sizing of the components is done on the base of the Peclet equation
SC
(3). Knowing the heat exchanger logarithmic temperature difference ∆Tlg , the
value of overall heat loss coefficient multiplied by the heat exchanger size may
M AN U
be defined (U A, W/K). Q˙ i = (U A)∆Tlg
140
(3)
The final sizing depends on assumptions about the type of the heat exchanger. All devices: evaporator, subcooler and condenser/absorber were assumed to be shell and tube heat exchangers. Detailed heat transfer models accounting for conduction and convection effects between external and internal
145
TE D
streams were applied. 2.2. Solar collector model
(4) [3]:
EP
The solar collector model follows the 2nd order Bliss Equation given in Eq.
ηsc = η0 − a1
AC C
where
∆Tm ∆Tm 2 − a2 G G
∆Tm = THTF − Tamb
(4)
(5)
∆Tm is defined as a difference between the average heat transfer fluid (HTF)
temperature and the ambient temperature. G, W/m2 stays for total solar radia-
150
tion on a tilted surface, which is a sum of the beam and scattered radiation. The equation includes performance constants: optical efficiency (η0 ), linear heat loss
coefficient (a1 , W/(m2 K)) and quadratic heat loss coefficient (a2 , W/(m2 K 2 )).
7
ACCEPTED MANUSCRIPT
The value of the heat rate produced by a solar collector is defined by equations
RI PT
(6) and (7). Q˙ u = ηsc · A · G
(6)
Q˙ u = m ˙ HTF (h17 − h18 )
SC
155
(7)
2.3. Modelling tools
The design and off-design models of the solar cooling cycle were prepared
solution of equations [13].
M AN U
within Engineering Equation Solver (EES)– a software enabling the numerical
Thermodynamic properties of ammonia-water solution were provided by an ex160
ternal routine available in EES. The thermodynamic data are based on the mixture equation of state described by Ibrahim O.M and Klein S.A. in [14]. Additionally, the properties of assumed heat transfer fluid for given temperature are recovered from the EES properties library as well.
165
TE D
The varying values of the incoming solar radiation and ambient temperature, needed for the evaluation of the solar collector performance, are obtained with the help of TRNSYS software and its Meteonorm libraries [15]. 2.4. Modelling assumptions
EP
• Chiller type and its design point parameters Since an ammonia-water chiller is characterized by an average 10-15% lower solar fraction than H2 O-LiBr chiller [16] and it requires higher driving temper-
AC C
170
ature to generate the ammonia vapours, it can be treated as a more demanding solar cooling system and as such less attractive to potential installers. In order to prove ammonia-water solar chillers competitive, some efforts should be made enabling the increase of solar fraction - which is the motivation behind
175
the present study. The only electric energy consumption in solar absorption chillers is due to operation of the solution pump. The cooling effect results from evaporation of 8
ACCEPTED MANUSCRIPT
pressurized ammonia under low temperature (around 0◦ C, 4 bar). The elec-
180
RI PT
trically driven compressor in vapour compression chillers is replaced by a heat driven absorption compressor set including a thermally driven desorber (generator) and a water-cooled absorber. The analysed system is assumed to be operating under moderate climate conditions with meteorological source data obtained from the weather station in Cracow, Poland.
185
SC
Several assumptions had to be made to enable the design analysis. They are listed in Table 1 and include: assumption of nominal driving heat rate Q˙ gen , temperature of desorbed ammonia coming out from generator Tgen , the
M AN U
temperature difference with respect to isothermal evaporation ∆Tglide , minimal temperature difference in the Condenser/Absorber heat exchanger ∆T13-6 , subcooling level in the regenerative heat exchanger ∆Tsub , concentration of rectified 190
ammonia xrect , difference of concentrations between rich and poor solutions ∆x and pressure levels (pcond , pint , pevap ).
Table 1: Design point assumptions.
Value
Q˙ gen
25.5 kW
Tgen = T5
120◦ C
∆Tglide = T11 − T10
3K
∆T13−6 = T13 − T6
0.2 K
AC C
EP
TE D
System parameter
∆Tsub = T8 − T9
10 K
xrect = x6
0.995
∆x = x1 − x14
0.15
pcond
15.5 bar
pint
11.5 bar
pevap
4 bar
The absorption cycle interaction with external flows must also be considered. Table 2 summarizes the design assumptions connected with these flows. Chosen
9
ACCEPTED MANUSCRIPT
Table 2: External flows design assumptions.
Component
External flow parameter
Evaporator
T21
7◦ C
• Solar collector type
∆THTF = T17 − T18
10 K
∆Tpp cw = T7 − T19
5K
∆Tcw = T20 − T19
10 K
M AN U
195
10 K
SC
∆Tapp = T18 − Tgen
Absorber-Condenser
Value
12◦ C
T22 Generator
RI PT
temperature limitations correspond to those present in the literature [17].
As presented in Table 1, the design generator temperature equals 120◦ C. Hence, a solar collector with stagnation temperature much higher than 140◦ C had to be chosen. The collector should ensure positive values of thermal ef-
200
TE D
ficiency even while pursuing higher outlet temperatures. For the purpose of the analysis, it was decided to consider an enhanced vacuum tube solar collector. The constants: optical efficiency (η0 ), linear heat loss coefficient (a1 , W/(m2 K)) and quadratic heat loss coefficient (a2 , W/(m2 K 2 )) were adapted
EP
from the manufacturer performance data and are collected in Table 3. Table 3: Solar collectors efficiency parameters [18]
Value
AC C
Parameter
Optical efficiency
η0 ,-
0.82 2
Linear heat loss coefficient
a1 , W/(m K)
1.62
Quadratic heat loss coefficient
a2 , W/(m2 K 2 )
0.0068
It was assumed that the collector would be a non-tracking surface with 205
a slope equal to location latitude (51◦ for Cracow). The collector is South 10
ACCEPTED MANUSCRIPT
oriented, consequently, the surface azimuth angle equals 0◦ . Since the analysis
RI PT
covers a geographical location of inferior radiation conditions and the possibility of morning frost even during the cooling season, it has been decided that a
thermal oil will work as a heat transfer fluid in order to exclude the danger of 210
working fluid freezing. Thermal Syltherm800 was chosen.
SC
3. Absorption chiller base model results - design and off-design
Under design operating conditions described in Table 2, the chiller reaches a COP of 0.488 which means that it may produce almost 12.5 kW of cooling
215
M AN U
power. At the same time, the pump consumes 99 W of electric energy (0.4% of the heat input) and 38 kW of low-temperature waste heat is generated. For systems integrated with energy sources of variable and intermittent nature, like solar, it is of high importance to perform an off-design simulation, since they usually operate in part-load conditions. The off-design simulation assumes that some relevant operating parameters are time- and weather-dependent: the 220
driving heat rate, the mass flow rate and the temperatures of the cooling water
TE D
and of the chilled water. It is expected that together with the change of external flow parameters, the performance of the solar chiller will be a matter of change. The chart in Fig. 2 shows the change of COP in the function of generator driving temperature. Generator temperature was the only variable. Temperatures of the cooling water and chilled water for this case were maintained at the nominal
EP
225
state conditions, T19 =35◦ C and T21 /T22 =12/7◦ C. It is evident that the COP of the absorption chiller becomes higher if the gener-
AC C
ator temperature is increased. However, if a cooling system is driven by the heat source strictly dependent on the ambient conditions and the surface of the solar
230
thermal collectors is fixed, it becomes relevant to consider a control strategy enabling efficient operation at lower temperature level. 3.1. Interdependence of collector efficiency and COP Solar absorption chillers are usually driven at a fixed temperature level in order to achieve a high COP. One should, however, take into consideration, 11
ACCEPTED MANUSCRIPT
235
that together with the change of external meteorological conditions (ambient
RI PT
temperature and incoming solar radiation), the efficiency of solar collector may suffer from trying to obtain the highest temperature possible. It is evident while looking again at the chart presented in Fig.2. The collector efficiency curve has
240
20◦ C and incoming solar radiation of 1000 W/m2 .
1
SC
1
0.8
0.6
M AN U
COP
0.8
0.4
0.2 0
145 150 155 generator driving temperature T17; °C
0.4
0.2 0
160
TE D
140
0.6
solar collector efficiency 𝜂sc
been prepared using equation (4) with the assumption of ambient temperature
COP
solar collector efficiency 𝜂sc
Figure 2: COP curve of analysed chiller and solar collector efficiency curve.
EP
It is shown that, although the COP of the chiller increases with the growth of driving temperature, if the temperature difference between the outlet stream from the solar collector and the ambient is higher, the efficiency of the collector
AC C
will be falling down. This phenomenon of inverse dependence between those two
245
performance coefficients speaks for operating the chiller at a variable driving temperature, adapting to outer conditions.
4. Idea of control procedure To investigate the phenomenon presented in subsection 3.1., a control strategy procedure is proposed. Authors state that it is not always a priority to 12
ACCEPTED MANUSCRIPT
250
maintain the generator temperature achieving the highest coefficient of perfor-
RI PT
mance (COP) under given conditions. If the solar radiation conditions are poor, it may occur that it is difficult to produce enough heat at the desired higher
temperature level. Under these conditions, it may be profitable to produce more
low-temperature heat, and accept a lower COP estimated from the curve, lead255
ing on the whole to a higher cooling power production from the solar collector.
SC
Some simplifications were introduced to manage the control routine. The chiller
is cooled down constantly by the inflowing cooling water temperature of 35◦ C, and the evaporator delivers chilled water at the level of 12/7◦ C, to maintain
260
M AN U
consistency with the exemplary COP recovered from off-design simulation. Following Fig. 2, COP depends mainly on the generator temperature. Since the pump mechanical energy consumption constitutes only 0.4% of the heat input, its contribution is not taken under consideration.
The heat transfer temperature differences are kept constant according to relations presented in (8) and (9). Lecuona et al. (2009) [10] mention that those ∆T usually ranges from 5 to 15◦ C.
TE D
265
∆THTF = T17 − T18 = 10K
(8)
∆Tgen = T18 − Tgen = 10K
(9)
EP
∆THTF , K is the temperature increase of the heat transfer fluid inside the solar collector, while the ∆Tgen , K is the temperature difference between the fluid stream coming back to the collector (T18 , ◦ C) and the generator temper-
AC C
ature (Tgen = T5 , ◦ C). An example control routine for finding optimal working
270
parameters of solar collector has also been described in [19], where the solar collector outlet temperature has been adapted to ensure maximum exergy efficiency of collector. It is ideally assumed that the useful heat gain from the collector equals the heat rate to generator assigned to solar contribution Q˙ u =Q˙ gen . In an iterative
275
way, the procedure searches for the generation temperature at which the cooling
13
ACCEPTED MANUSCRIPT
power (Q˙ evap ) produced by the solar chiller is the highest, which at the same
is presented in Eq. (10). Tgen = f (max : Q˙ evap = COP · Q˙ u )
RI PT
means the highest possible solar fraction. The objective function of this routine
(10)
To maintain consistency with the before-mentioned considerations, the rou-
tine looks for the desired temperature in the range between 140◦ C (the lowest
SC
280
temperature for which off-design simulation has been done) and 161◦ C. The value of driving temperature is searched for raising values of the cooling power
M AN U
output; once the cooling power output begins to decrease, the optimal solution is refined by chord interpolation. The collector mass flow rate enabling heating 285
the heat transfer fluid to the desired temperature is defined on the base of equation (7) and, consequently, it should be processed and transferred as a setup control signal to remote-controlled variable speed pump.
TE D
5. Example application of the model - the study case The effect of the routine was tested in simulations performed with a 5290
minutes time step for 3 representative days of the cooling season. The overall solar radiation and ambient temperature data were processed to generate 3 average days statistically representative of specific months of the year. The analysis
EP
covers May, June and July, as months belonging to cooling season. Meteorological data for Cracow were used. Since the data recovered from Meteonorm library are hourly, a cubic interpolation was applied to simulate the operation
AC C
295
according to the smaller time step. The variable driving temperature study case is compared with the reference case operation.
The reference case follows an assumed standard operation routine. In this rou-
300
tine, the solar chiller is driven by the heat output of the solar collector at constant temperature level. During every time step of operation, a constant value of the solar collector outlet temperature is maintained (161◦ C). This value was 14
ACCEPTED MANUSCRIPT
chosen for the simulation purpose, as the driving temperature assuring high
305
RI PT
COP value (up to 0.9 according to Fig. 2). The analysed study case, on the other hand, assumes that the solar collector outlet temperature is optimized during every time step, according to the dis-
cussed objective function in Eq. (10). The chilled water and cooling water temperature remain the same for the reference and study case, so that the re-
310
the unit surface area of the solar collector.
SC
sults are comparable. The cooling power produced in both cases is referred to
The results are presented in graphical form in Figs. 3, 4 and 5 (May, June,
M AN U
and July). The upper charts show the amount of cooling power that could be produced from 1 m2 of solar collector at a fixed driving temperature and the increment of cooling power production that constant adjusting of collector 315
outlet temperature may bring. The bottom charts present the daily profiles of the outlet temperature and collector energy efficiency for both cases: with optimal adjustment of the outlet temperature (temp.var.) and for fixed value of
AC C
EP
TE D
it (temp.f ixed).
15
ACCEPTED MANUSCRIPT
140 600 500
100
400 300
60
200
40
100
20 0
0
10
11
variable ∆ 𝑄evap/mtemperature 2 (𝑡𝑒𝑚𝑝. 𝑣𝑎𝑟. )
160 150 140
13
14
15
hour fixed temperature 𝑄evap/m 2 (𝑡𝑒𝑚𝑝. 𝑓𝑖𝑥𝑒𝑑)
16
solar radiation G 0.6 0.5 0.4 0.3
130 120 110 100
9
TE D
driving temperature T17, °C
170
12
M AN U
9
B
SC
unit cooling power
80
10
11
12
AC C
EP
hour fixed temperature T17 coll. efficiency (temp. fixed) 𝜂sc
Figure 3:
solar radiation G, W/m2
RI PT
120
0.2 0.1
solar collector efficiency 𝜂sc
, W/m 2
A
0
13
14
15
16
variable temperature T17 coll. efficiency (temp. var.) 𝜂sc
Results of the simulation for a representative day of May; A: distribution of unit
cooling power production and its increase at variable driving temperature, solar radiation distribution; B: distribution of driving temperature and collector efficiency for both study cases.
16
ACCEPTED MANUSCRIPT
140 600 500
100
400 300
60
200
40
100
20 0
0
10
11
variable ∆ 𝑄evap/mtemperature 2 (𝑡𝑒𝑚𝑝. 𝑣𝑎𝑟. )
160 150 140
13
14
15
hour fixed temperature 𝑄evap/m 2 (𝑡𝑒𝑚𝑝. 𝑓𝑖𝑥𝑒𝑑)
16
solar radiation G 0.6 0.5 0.4 0.3
130 120 110 100
9
TE D
driving temperature T17, °C
170
12
M AN U
9
B
SC
unit cooling power
80
10
11
12 hour
AC C
EP
fixed temperature T17 coll. efficiency (temp. fixed) 𝜂sc
Figure 4:
solar radiation G, W/m2
RI PT
120
0.2 0.1
solar collector efficiency 𝜂sc
, W/m 2
A
0
13
14
15
16
variable temperature T17 coll. efficiency (temp. var.) 𝜂sc
Results of the simulation for a representative day of June; A: distribution of unit
cooling power production and its increase at variable driving temperature, solar radiation distribution; B: distribution of driving temperature and collector efficiency for both study cases.
17
ACCEPTED MANUSCRIPT
140 600 120 500
100
400 300
60
200
100
20
0
10
11
variable ∆ 𝑄evap/mtemperature 2 (𝑡𝑒𝑚𝑝. 𝑣𝑎𝑟. )
160 150 140
13
14
15
hour fixed temperature 𝑄evap/m 2 (𝑡𝑒𝑚𝑝. 𝑓𝑖𝑥𝑒𝑑)
16
solar radiation G 0.6 0.5 0.4 0.3
130 120 110 100
9
TE D
driving temperature T17, °C
170
12
10
11
12
0.1 0
13
14
15
16
variable temperature T17 coll. efficiency (temp. var.) 𝜂sc
AC C
EP
hour fixed temperature T17 coll. efficiency (temp. fixed) 𝜂sc
0.2
solar collector efficiency 𝜂sc
0 9
B
SC
40
M AN U
unit cooling power
80
solar radiation G, W/m2
RI PT
, W/m 2
A
Figure 5:
Results of the simulation for a representative day of July; A: distribution of unit
cooling power production and its increase at variable driving temperature, solar radiation distribution; B: distribution of driving temperature and collector efficiency for both study cases.
Incorporating the given control strategy may result in an increase of the 320
cooling power production related to solar collector utilization by up to 34 W 18
ACCEPTED MANUSCRIPT
per 1 m2 of collector. For specific conditions, it is possible to more than double
RI PT
the cooling power yield, which is happening around 10 a.m. and 3 p.m. for every of the representative days. The simulations show that under poor radiation
conditions it is profitable to operate close to the minimal driving temperature 325
allowing for ammonia vapour generation (here: 140◦ C). Optimal adaptation of driving temperature enables to obtain a positive value of cooling power around
SC
9 a.m. and 4 p.m. in May, June and July. According to the charts, it is
convenient to operate solar collectors at the highest outlet temperature level only when high radiation conditions are possible (during noon hours). During this time, the chosen temperature is close to the reference fixed of 161◦ C. As
M AN U
330
expected, a lower collector outlet temperature always leads to higher solar collector efficiency, as is visible in the bottom charts of Fig. 3, 4 and 5. In order to broaden the comparative analysis, a similar simulation study was performed for the case of parabolic trough concentrating collectors. The effi335
ciency parameters were: optical efficiency η0 =0.68, linear heat loss coefficient a1 =0.04 W/(m2 K) and quadratic heat loss coefficient a2 =0.0015 W/(m2 K2 )
TE D
[18]. It occurs that, although this type of collector is only able to utilize beam solar radiation, its efficiency is much less dependent on the change of outlet temperature in the range between 140-161◦ C. Thereby, the iterative procedure 340
produces an outlet temperature close to the fixed one (161◦ C) during the ma-
EP
jority of simulation hours of analogic representative days. According to the summary of results: it occurs that it is more profitable to operate the parabolic trough collectors on a reduced collector temperature (140◦ C) only in the late
AC C
afternoon hours (5 p.m.) in May, June or July. During the rest of the days 345
(8 a.m.-4 p.m.) the procedure chooses outlet temperatures within the narrow range between 160 and 161◦ C. In order to discuss the effect of applying the control strategy, an additional
indicator was considered, that is, the collector exergy efficiency indicator: coefficient of quality (COQ). The exergy of thermal radiation (or actually the change of exergy occurring with thermal radiation transfer) has been investigated in terms of fundamental research by several Authors. The sun emits its 19
ACCEPTED MANUSCRIPT
radiation at an equivalent very high temperature of nearly 6000 K. The exergy
RI PT
of solar radiation is lost due to transmission through space, and scattering by the atmosphere, to be finally utilized at respectively low temperatures. Petela
(2003) [20] stated that in order to derive a correct exergy formula it is wise to
interpret the characteristic exergy equation h − hamb − Tamb · (s − samb ). For solar thermal radiation the enthalpies would be substituted by the brightness
SC
values of radiation propagating within the given solid angle and the entropy of the radiation brightness would be considered. The same, the entropy should not be calculated as for the heat, but in accordance with Planck's formulae. The
M AN U
author managed to obtain a formula of maximum conversion efficiency of direct 1 Tamb 4 solar radiation (1 − 34 Tamb Ts + 3 ( Ts ) ). These considerations have been further
developed by Chu et al. (2009) [21], who provided formulas for the exergy of solar radiation covering the beam and diffuse components. The authors used spectral analysis for the wavelength range from 0.28 to 4 µm. According to their findings, the maximum conversion efficiency (exergy-to-energy ratio) for total solar radiation is about 0.92. All this considered, for the only purpose of the
TE D
qualitative evaluation of a collector flow rate control strategy, it is possible to apply a simplified approach. The input exergy rate is thus taken as the overall incoming solar radiation over the gross solar collector aperture area. Hence, the exergy efficiency of a solar collector is calculated only indicatively with the help
EP
of coefficient of quality presented in equation (11) [22]: COQ =
E˙ sc G·A
(11)
AC C
The rate of physical exergy (E˙ sc ) gained by the material stream of the heat
transfer fluid is given by equation (12): E˙ sc = m(h ˙ 17 − h18 − Tamb · (s17 − s18 ))
(12)
Values of specific enthalpy and entropy of the Syltherm heat transfer fluid were obtained using the EES software built-in libraries discussed in the Modelling Tools section.
350
Taking into consideration the radiation transmission exergy loss, the denomina20
ACCEPTED MANUSCRIPT
tor in Equation (11) would be lower; the attenuation factor would be variable
RI PT
with time, depending on the solar declination and on the length of the transmission path across the atmosphere; it would also be affected by meteorological conditions across the transmission path (mainly, by the presence of water 355
vapour); however, the qualitative daily trend of the collector exergy efficiency indicator should be marginally affected. The data presented in Fig. 6 provide
SC
the comparison of the collector COQ for the case of fixed and variable outlet
temperature (in this last case, temperature is adjusted to the environmental conditions searching numerically for the maximum COP conditions); the COP profiles are also shown for both study cases.
AC C
EP
TE D
M AN U
360
21
ACCEPTED MANUSCRIPT
1
0.25
0.9
0.2
RI PT
0.8 0.7
0.15
COQ
COP
0.6 0.5 0.4
0.1
0.3 0.2
0.05
SC
0.1 0
0
9
10
11
12
13
14
15
hour 1
0.25
0.9 0.8 0.7
COP
0.6 0.5 0.4 0.3 0.2
0.1 0
9
10
11
M AN U
b
16
12
13
14
15
0.2
0.15
0.1
COQ
a
0.05
0
16
1 0.9 0.8 0.7
0.2
0.15
0.5
EP
COP
0.6
0.25
COQ
c
TE D
hour
0.4
0.1
0.3 0.2
0.05
AC C
0.1 0
9
Figure 6:
0
10
11
COP (temp. fixed) COQ (temp. fixed)
12
13
14
hour
15
16
COP (temp. var.) COQ (temp. var.)
Solar collector COQ and COP value distribution for fixed and variable outlet
temperature on representative day of May, June, and July
22
ACCEPTED MANUSCRIPT
It is visible that, although the heat from a collector with constantly adjusted
RI PT
outlet temperature is produced at lower temperature level, the COQ is always higher than that for the reference case. It is shown that, as the input exergy rate A · G is equal for both cases, the increase of COQ derives from the higher value 365
of the collector output exergy rate. It results from the method of controlling the outlet temperature, which is done by a dynamic change of mass flow rate.
SC
Lower outlet collector temperatures are obtained at higher mass flow rate of heat transfer fluid (e.g. 0.0087 kg/sm2 at 11 a.m. in June). The reference case produces high outlet temperatures, for which lower mass flow rate are calculated, respectively (e.g. 0.0048 kg/sm2 at 11 a.m. in June).
M AN U
370
As discussed in section 5 and is shown in the Fig. 6, the COP value suffers from a lower temperature level heat driving stream. However, considering solar thermal energy as a free input, it is profitable to consume more of it and, consequently, to achieve higher solar fraction. 375
For the purpose of results generalization, a sensitivity analysis was performed. While considering the solar collector efficiency and COP value sep-
TE D
arately, relative increments with respect to the design point were obtained. Namely: if the values of solar radiation, ambient temperature or outlet solar collector temperature were increased by 1%, the solar collector efficiency value 380
would be changed by 0.5% or 0.1% or -0.8% of the original value, respectively.
EP
On the other hand, if the outlet solar collector temperature is increased by 1%, the COP would rise about 2.5% of the original value. This preliminary evaluation was followed by the sensitivity analysis on the re-
AC C
sults of the control procedure. The analysis indicates how the changing weather 385
conditions affect the decision on the optimal solar collector outlet temperature resulting from application of the proposed procedure. The results of the sensitivity analysis are presented in Fig. 7. Chart a on the left hand side reveals the breakthrough value of solar radiation valid for given ambient temperature (on x-axis) when the procedure no longer chooses the lowest driving temperature
390
(140◦ C) as optimal. Increasing the solar radiation beyond this value, the optimal temperatures get close to the reference 161◦ C. This phenomenon is visible 23
ACCEPTED MANUSCRIPT
in the plot on the right hand side. Plot b treats separately 5 cases for 5 different
RI PT
ambient temperatures. It shows the dynamics of optimal temperature choice in the function of increasing solar radiation. It covers the events when it is no 395
longer profitable to operate on the lowest driving temperature (140◦ C) and the chosen temperature is reaching 161◦ C. 162
161
600
580
160
560
159
540 520
M AN U
solar radiation G, W/m2
620
SC
b
640
collector outlet temperature T17, °C
a
158
0
5
10
15
20
25
30
ambient temperature Tamb , °C
500
Tamb:
550
5°C
600
650
700
solar radiation G, W/m2 10°C
15°C
20°C
750
800
25°C
Figure 7: Sensitivity analysis results: chart a solar radiation values at given ambient temperatures when optimal solar collector outlet temperature is close to 161◦ C; chart b distribution
TE D
of optimal solar collector outlet temperature close to 161◦ C in the function of changing radiation conditions for given ambient temperature.
The sensitivity analysis confirms that both ambient temperature and solar
EP
radiation values have an impact on the choice of the optimal temperature at the outlet from the solar collector.
6. Conclusions
AC C 400
A simplified theoretical method of controlling the solar collector outlet tem-
perature (which represents the chillers driving temperature) was proposed for a solar absorption chiller. Its purpose is to increase the cooling power production and thereby to increase the solar fraction of the cycle. The idea relies on the
405
physical phenomenon of the inverse trend referring to the solar collector outlet temperature - between the collector efficiency and the COP of the absorption 24
ACCEPTED MANUSCRIPT
cycle. An iterative procedure was set in an EES simulation software environ-
RI PT
ment, with the purpose of finding an optimal collector outlet temperature that is, a value maximizing the cooling effect - at every hour of average days of May, 410
June, and July in Cracow, Poland.
The results presented in current study show that incorporating a control strat-
egy adapting solar collector outlet temperature could bring a high profit to a
SC
system operating under moderate climate conditions. The highest increase in
cooling power production takes place during morning and late afternoon hours 415
when the solar radiation is low. During these times of the day, the cooling effect
M AN U
can be more than doubled.
Since none of the previous studies considered controlling a solar ammonia-water chiller (additionally: working under moderate climate conditions), it is not possible to compare the results directly. However, the main conclusions of the 420
analysis presented by Lecuona et al. (2009) [10] are consistent with the findings presented here: a lower generator operating temperature increases the maximum value of solar-induced COP; additionally, the system is more sensitive to
power is low. 425
TE D
temperature change at sunrise and sunset hours when the radiation and cooling
A future advanced analysis should include a transient effect of the system warming up in the morning hours. Moreover, transient effects were neglected when
EP
assuming that changing the mass flow rate of the heat transfer fluid results in an immediate change to the desired temperature. Further enhancements of the method should include: a routine application of maximum system exergy efficiency, transient modelling of heat transfer and the investigation of the dynamic
AC C
430
response to a change of mass flow rate.
Acknowledgements The present work is part of a Joint Ph.D. project, resulting from a long-
standing cooperation between Silesian University of Technology and the Uni435
versity of Florence.
25
ACCEPTED MANUSCRIPT
This research was supported by the statutory research funds of the Faculty
RI PT
of Power and Environmental Engineering of the Silesian University of Technology and by the Internationalization Plan funds of Department of Industrial Engineering (DIEF) of the University of Florence.
Exergy rate, W
m ˙
Mass flow rate, kg/s
Q˙
Heat rate, W
A
Solar collector aperture area, m2
a1
Linear collector constant, W/(m2 K)
a2
Quadratic collector constant, W/(m2 K 2 )
COP
Coefficient of performance, -
SC
E˙
M AN U
445
Nomenclature
TE D
440
450
G
Total solar radiation, W/m2
h
Enthalpy, kJ/kg
p
EP
COQ Coefficient of quality, -
Entropy, J/(kgK)
AC C
s
Pressure, Pa
455
T
Temperature, ◦ C
UA
Overall heat loss coefficient multiplied by heat exchanger area, W/K
Greek Symbols ∆
Variation
η
Efficiency
26
ACCEPTED MANUSCRIPT
Subscripts
470
Optical
amb
Ambient
app
Approach
av
Average
cond
Condenser
cw
Cooling water
evap
Evaporator
ex
Exergy
gen
Generator/desorber
glide
Isothermal slip
HTF
Heat transfer fluid
in
Input to the system/component
int
Intermediate
lg
SC M AN U
Logarithmic
Medium
AC C
m
RI PT
0
TE D
465
Variable
EP
460
i
475
out
Output from the system/component
pp
Pinch point
rect
Rectifier
s
Sun
SC
Solar Collector 27
ACCEPTED MANUSCRIPT
sub
Subcooler
u
Useful gain
RI PT
480
References
[1] N. Pardo, K. Vatopoulos, A. Krook-Riekkola, J. Moya, A. Perez,
485
SC
Heat and cooling demand and market perspective, Tech. rep. (2012). doi:10.2790/56532. URL
https://pure.ltu.se/portal/files/42577403/ldna25381enn.
M AN U
pdf
[2] T. Otanicar, R. a. Taylor, P. E. Phelan, Prospects for solar cooling - An economic and environmental assessment, Solar Energy 86 (5) (2012) 1287– 490
1299. doi:10.1016/j.solener.2012.01.020.
URL http://dx.doi.org/10.1016/j.solener.2012.01.020 [3] W. Beckman, J. Duffie, Solar Engineering of Thermal Processes, John Wi-
TE D
ley and Sons, 1980.
[4] A. Shirazi, R. A. Taylor, S. D. White, G. L. Morrison, Transient simula495
tion and parametric study of solar-assisted heating and cooling absorption systems: An energetic, economic and environmental (3E) assessment, Re-
EP
newable Energy 86 (2016) 955–971. doi:10.1016/j.renene.2015.09.014. URL http://dx.doi.org/10.1016/j.renene.2015.09.014
AC C
[5] U. Eicker, D. Pietruschka, M. Haag, A. Schmitt, Systematic design and 500
analysis of solar thermal cooling systems in different climates, Renewable Energy 80 (2015) 827–836. doi:10.1016/j.renene.2015.02.019. URL
http://linkinghub.elsevier.com/retrieve/pii/
S0960148115001093
[6] U. Eicker, D. Pietruschka, Design and performance of solar powered ab505
sorption cooling systems in office buildings, Energy and Buildings 41 (1) (2009) 81–91. doi:10.1016/j.enbuild.2008.07.015. 28
ACCEPTED MANUSCRIPT
[7] U. Eicker, D. Pietruschka, R. Pesch, Heat rejection and primary energy
RI PT
efficiency of solar driven absorption cooling systems, International Journal of Refrigeration 35 (3) (2012) 729–738. doi:10.1016/j.ijrefrig.2012. 510
01.012.
URL http://dx.doi.org/10.1016/j.ijrefrig.2012.01.012
[8] A. Syed, M. Izquierdo, P. Rodr´ıguez, G. Maidment, J. Missenden,
SC
A. Lecuona, R. Tozer, A novel experimental investigation of a solar cooling
system in Madrid, International Journal of Refrigeration 28 (6) (2005) 859–871. doi:10.1016/j.ijrefrig.2005.01.007. URL
http://linkinghub.elsevier.com/retrieve/pii/
S0140700705000277
M AN U
515
[9] Z. Li, X. Ye, J. Liu, Optimal temperature of collector for solar double effect LiBr/H2O absorption cooling system in subtropical city based on 520
a year round meteorological data, Applied Thermal Engineering 69 (1-2) (2014) 19–28. doi:10.1016/j.applthermaleng.2014.04.039. URL
TE D
http://www.sciencedirect.com/science/article/pii/
S1359431114003032
[10] A. Lecuona, R. Ventas, M. Venegas, A. Zacar´ıas, R. Salgado, Optimum 525
hot water temperature for absorption solar cooling, Solar Energy 83 (2009)
EP
1806–1814. doi:10.1016/j.solener.2009.06.016. [11] J. Albers, A. Kuhn, S. Petersen, F. Ziegler, Control of absorption chillers
AC C
by insight: the characteristic equation, Test. [12] F. A. Al-Sulaiman, F. Hamdullahpur, I. Dincer, Performance assessment of
530
a novel system using parabolic trough solar collectors for combined cooling, heating, and power production, Renewable Energy 48 (2012) 161–172. doi: 10.1016/j.renene.2012.04.034.
URL http://dx.doi.org/10.1016/j.renene.2012.04.034
[13] EES-software, Product information (May 2016). 535
URL http://www.f-chart.com 29
ACCEPTED MANUSCRIPT
[14] O. Ibrahim, S. A. Klein, Thermodynamic properties of ammonia-water
RI PT
mixtures., ASHRAE Transactions 99 (0) (1993) 1495–1502. [15] TRNSYS-software, Meteonorm library available within the software package (May 2016). 540
URL http://www.trnsys.com
SC
[16] Z. F. Li, K. Sumathy, Technology development in the solar absorption air-
conditioning systems, Renewable and Sustainable Energy Reviews 4 (2000) 267–293. doi:10.1016/S1364-0321(99)00016-7.
545
M AN U
[17] D. Kong, J. Liu, L. Zhang, Z. Fang, G. Zheng, Theoretical and experimental analysis of a single stage ammonia-water absorption chiller performance, in: 2009 International Conference on Energy and Environment Technology, ICEET 2009, Vol. 1, 2009. doi:10.1109/ICEET.2009.121. [18] IEA, Solar collector data base available at (September 2016). URL http://www.iea-shc.org
[19] G. Manfrida, V. Gerard, Maximum Exergy Control of a Solar Thermal
TE D
550
Plant Equipped with Direct Steam Collectors, International Journal of Thermodynamics 11 (3) (2008) 143–149. [20] R. Petela, Exergy of undiluted thermal radiation, Solar Energy 74 (6)
555
EP
(2003) 469–488. doi:10.1016/S0038-092X(03)00226-3. [21] S. X. Chu, L. H. Liu, Analysis of terrestrial solar radiation exergy, Solar
AC C
Energy 83 (8) (2009) 1390–1404. doi:10.1016/j.solener.2009.03.011.
URL http://dx.doi.org/10.1016/j.solener.2009.03.011
[22] G. Manfrida, S. Kawambwa, Exergy control for a flat-plate Collector/Rankine Cycle Solar Power System, ASME Journal of Solar Energy
560
Engineering 113 (1991) 89–93. doi:10.1115/1.2929963.
30