1 Dynamic simulation and exergetic optimization of a Concentrating

2 Photovoltaic/ Thermal (CPVT) system.

3
4 I.K. Karathanassisa, b,*,1, E. Papanicolaoua, V. Belessiotisa and G.C. Bergelesb
5
6 a
Solar & other Energy Systems Laboratory, Institute of Nuclear and Radiological Sciences & Technology,
7 Energy and Safety, National Centre for Scientific Research DEMOKRITOS, Aghia Paraskevi, 15310 Athens,
8 Greece
9
10 b
Laboratory of Innovative Environmental Technologies, School of Mechanical Engineering, National Technical
11 University of Athens, Zografos Campus, 15710 Athens, Greece
12
13 *
Corresponding author: Ioannis.Karathanassis@city.ac.uk (I. K. Karathanassis).
14
15 1
Present address: School of Mathematics, Computer Science and Engineering, City, University of London,
16 Northampton Square, EC1V 0HB London, UK
17
18
19 Abstract. The development of a dynamic, theoretical model suitable for the prediction of
20 the long-term performance of a parabolic-trough Concentrating Photovoltaic/ Thermal CPVT
21 system is discussed in the present study. The formulation of the mathematical model and the
22 considered geometrical and operational parameters of the system, such as the characteristics
23 of the employed PV modules and active cooling system are described in detail. The effect of
24 heat capacity is taken into consideration in the thermal balances and thus the model is able to
25 capture the transient behavior of the system. Besides, the model is validated using available
26 experimental data of a manufactured prototype CPVT system. The daily performance of
27 system is predicted for different values of the cooling fluid flow rate and temperature under
28 various environmental conditions. At a second stage, an exergy analysis is conducted in order
29 to point out the effect of the characteristics of the main system sub-components on the
30 exergetic efficiency and exergy output of the CPVT system. It was established that the system
31 exergetic performance is primarily influenced by the optical quality of the parabolic trough
32 and the electrical efficiency of the PV module. Increasing these two factors to achievable
33 values, e.g. ηopt=0.75 and ηel=0.25, can yield an increase of the system exergetic efficiency
34 from 12% to 24%.
35
36
37 Keywords: CPVT system, combined heat and power, dynamic simulation, exergy analysis, parametric analysis
38
39
40 Nomenclature 55 L length, m
41 56 M mass, kg
42 A area 57 Nu Nusselt number
43 CR concentration ratio 58 Ns entropy generation
44 Cp specific heat, J/kgK 59 number
45 Dh hydraulic diameter, m 60 ṁ mass flow rate, kg/s
46 dt time step, s 61 P power, W
47 Ė x exergy rate, W 62 Qth thermal power, W
48 G solar radiation flux, 63 Rth thermal resistance, K/W
49 W/m2
64
Ṡ gen entropy generation rate
50 H height, m
65 Τ temperature, K
51 h heat transfer coefficient,
66 U0 thermal losses
52 W/m2K
67 coefficient, W/m2K
53 k thermal conductivity,
68 ts substrate thickness, m
54 W/mK

1
 69 W width, m 99 b beam
70
V̇ tot volumetric flow rate
100 CPVT Concentrating
101 Photovoltaic Thermal
71 102 d diffuse
72 103 dest destroyed
73 Greek symbols 104 el electrical
74 105 ext external
75 α absorptance 106 exp experimental
76 β cell temperature 107 gl glass
77 coefficient, %/K 108 hs heat sink
78 γ receiver intercept factor 109 f fluid
79 ε emmitance 110 i inlet
80 ηel electrical efficiency 111 ins insulation
81 ηth thermal efficiency 112 o outlet
82 ηtot total efficiency 113 irr irradiation
83 ηopt optical efficiency 114 max maximum
84 ηΙΙ exergetic efficiency 115 min minimum
85 θ incidence angle, o 116 PV photovoltaic
86 μ dynamic viscosity, Pa∙s 117 MPP maximum power point
87 ρ reflectance, density 118 opt optical
88 σ Stefan-Boltzmann 119 rad radiation
89 constant, W/m2K4 120 rec receiver
90 τ transmittance 121 ref reference
91 122 refl reflector
92 Subscripts/Abbreviations 123 s substrate
93 124 scat scattered
94 a aperture, ambient 125 TMY Typical Meteorological
95 ave average 126 Year
96 ch channel 127 th thermal
97 cond conductive 128 w wall
98 conv convective
129
130 1 Introduction
131
132 Analytical models have been widely used and proven adequate for the prediction of the
133 performance and the daily or annual yield of conventional PVT systems using sheet and tube
134 absorbers, similar to those encountered in flat-plate, solar thermal collectors [1-4]. In addition,
135 there is a number of studies in the literature that discuss the formulation of analytical models
136 representing CPVT systems. However, the majority of these studies do not take transient
137 effects into account.
138 The term Concentrating Photovoltaic/Thermal (CPVT) system refers to integrated
139 configurations that can simultaneously produce electrical and thermal energy, using the solar
140 irradiation. The system comprises three distinct constituents, namely the optical device to
141 concentrate solar irradiation, the solar-cell module and the active cooling system.
142 Concentrated solar irradiation is incident on the active surface of the solar-cell module at the
143 backside of which, a suitable cooling device (heat sink) is thermally bonded to extract excess
144 heat. An early study by O’Leary and Clements [5] illustrates the formulation of an analytical
145 model of a parabolic trough CPVT system. Transient effects were neglected, as the model was
146 intended for the determination of the optimal coolant flow-rate that maximizes the system net
147 electrical output. Helmers et al. [6] formulated a theoretical model to predict the influence of
148 operating parameters on the efficiency of a point-focus CPVT system. The investigation was
149 focused on the effect of the concentration ratio and cell operating temperature. Their results
150 showed that under high concentration (CR>300), the rate of thermal and electrical losses
151 decreases and a system overall efficiency of 75% can be achieved. Calise et al. [7] developed
152 an analytical steady-state model for a parabolic trough CPVT system with triple-junction cells

2
153 bonded to a receiver of triangular cross-section and a circular passage where the cooling fluid
154 flows. A parametric analysis regarding the variation of system energetic and exergetic
155 efficiency with the system length and the coolant flow rate, respectively, was conducted. It
156 was shown that increase of both parameters leads to increased efficiencies. Researchers from
157 the same group also integrated their developed model to a tri-generation system that uses an
158 absorption chiller to meet cooling loads [8]. The investigation illustrated the system capability
159 to cover the annual electricity, heating and cooling demands of a university hospital.
160 Buonomano et al. [9] used a one-dimensional algebraic model to predict the performance
161 of the solar-dish CPVT system introduced by Kribus [10]. Steady state conditions were
162 considered and the system of equations comprising the model was further simplified by
163 neglecting some of the material layers in the receiver. The authors provided predictions for
164 the system thermal and electrical efficiency and, in addition, illustrated the effect of the main
165 geometrical parameters and operating conditions on the system efficiency. A Fresnel lens
166 linear CPVT system that employs high efficiency triple-junction cells with thermal energy
167 storage was modeled by Kerzmann and Schaufer [11]. A rectangular duct was considered as
168 the cooling device for the PV module. The thermal balance equations were formulated for
169 each segment of the linear receiver assuming steady-state conditions and the system daily and
170 annual yield were predicted. Renno and Petito [12] evaluated the performance of a solar tri-
171 generation system that comprises a point-focus dish CPVT system with a remote duct
172 receiver, a storage tank and a absorption heat pump, in order to meet the electrical, thermal
173 and cooling loads respectively. A one-dimensional model, assuming steady-state operation of
174 all the devices was formulated, and the system yield for two site locations was predicted, by
175 also taking into account the number of solar cells and the concentration ratio as parameters.
176 Few studies have been found, where the effect of heat capacitance has been included in the
177 thermal balance equations and thus the models can predict the transient behavior of the CPVT
178 system as well. Helmers and Kramer [13] formulated a quasi-dynamic theoretical model that
179 can predict the thermal and electrical output of a CPVT system by making use of linear
180 regression based on constant coefficients. A transient model was developed by Coventry [14],
181 whereby the formulated set of energy balance differential equations was solved analytically
182 instead of being iteratively approximated. The model was used for the prediction of the
183 annual output of a system coupled to thermal storage. In addition, the effect of basic design
184 parameters, such as the insulation thickness and the conductivity of the thermal interface
185 material, on the annual efficiency was examined. Kosmadakis et al. [15] simulated the
186 performance of a linear trough CPVT system which also served as a prime mover for an
187 organic Rankine cycle in order to produce additional electricity.
188 The scope of this study is to illustrate the influence of the various components geometrical
189 layout, materials and parameters of operation on the CPVT system long-term energetic and
190 exergetic performance by means of a dynamic theoretical model. It comes as a follow up of a
191 previous study [16] discussing the design and experimental evaluation of a prototype CPVT
192 system. The present study focuses on the formulation of a 1-D theoretical model accompanied
193 by an optimization procedure, suitable for pinpointing the strongpoints and weaknesses of the
194 prototype system and assisting to the design and sizing of large-scale CPVT systems. The
195 model was employed to conduct a parametric analysis, where the system daily or monthly
196 performance was determined for different values of the geometrical and operational
197 parameters. The system receiver was divided in segments, so that the temperature distribution
198 at the various material layers along its length can be obtained and consequently the thermal
199 losses and the system electrical performance can be accurately predicted. The main factors
200 that influence the system performance from an exergetic point of view were identified and
201 suggestions have been made, in order to increase the CPVT system exergetic efficiency. It can
202 be therefore summarized that the presented methodology offers a novel and useful numerical

3
203 tool suitable for the sizing and optimization of large-scale parabolic-trough CPVT systems. It
204 is a fully versatile approach taking into account variable material properties and geometrical
205 parameters of the CPVT system sub-components, i.e. concentrator, PV module and cooling
206 system. Hence, the developed algorithm can enable engineers reduce the overall time needed
207 for CPVT system design and development, a crucial necessity for the relevant technology to
208 reach further maturity.
209
210 2 Formulation of the dynamic model
211
212 A novel parabolic-trough CPVT system, depicted in Fig. 1a, has been designed,
213 manufactured and experimentally evaluated by the authors, as discussed in detail in [16]. The
214 system receiver comprises an array of mono-crystalline silicon cells in thermal bond to a
215 suitable heat sink for extracting excess heat. The PV module and the heat sink are enclosed
216 within a heavily insulated aluminum container, as depicted in Fig. 1b. It must be noted that
217 two module designs have been considered for different variations of the system receiver that
218 have been evaluated, bearing solar-cells of widths equal to 40.0 mm (“narrow” cells) and 60.0
219 mm (“wide” cells), respectively. In addition, two different heat sink configurations employing
220 micro-channels of straight (fixed width or FW configuration) and stepwise-varying width
221 (variable width or VW configuration) have been implemented in different design-variations of
222 the system receiver, in order to highlight the effect of each device performance characteristics
223 on the system performance. Fig. 1c depicts the geometrical layout of the FW and VW
224 configurations. Referring to the VW design, it can be seen that the heat-sink length is
225 segregated in three sections and the number of channels is doubled at each section. The
226 specific design reduces the pressure drop penalty compared to straight-channel plate-fin heat
227 sinks and increases the flow uniformity on the substrate surface. Despite the fact of reduced
228 number of fins compared to the FW layout, the thermal resistance exhibits only a minor
229 increase due to the different prevailing heat transfer conditions (mixed instead of forced
230 convection) and the increased heat transfer rate due to distinct flow features, such as the
231 presence of longitudinal vortices. Further information with regard to the geometrical layout
232 and the performance characteristics of the employed devices can be found in [17, 18].
233

4
234
235

236
237
238
239
240
241
242
243
244
245
246
247
248
249
250 Fig. 1 (a) Photograph of the prototype CPVT system, (b) “exploded” view of the system receiver and (c)
251 geometrical layout of the FW and VW heat-sink configurations employed in the system receiver.

5
252
253 The simulation of the CPVT system transient performance was based on energy balances
254 performed on every layer of the system receiver. The model was formulated under the main
255 assumptions of negligible gravitational terms and negligible losses due to end effects. The
256 latter assumption is legitimate, as end losses occurring on the CPVT receiver for high
257 incidence angles can be avoided either by incorporating a two-axes tracking system or by
258 extending the reflector beyond the receiver active length. As was confirmed during the
259 experimental evaluation of the prototype CPVT system [16], manufacturing imperfections
260 and partial distortion in the shape of the parabolic trough incorporated in the system had a
261 significant effect on the irradiation distribution on the PV module. The effect of non-uniform
262 irradiation distribution on the module surface was taken into account through a constant factor
263 Funi, whose value is properly set so that the predictions are in agreement with the experimental
264 measurements. Seven material layers were considered and the resulting network of thermal
265 resistances that represents the receiver is depicted in Fig. 2. For each node i of the thermal
266 network, i.e. for each material of the receiver, the formulated thermal-balance equation takes
267 the general form:
268
dT i
M iCi =∑ Qin−∑ Qout
269 dt j k (1)
270
271 where M and C are the material mass and specific heat, respectively. Transient effects are
272 taken into account through the term on the left-hand side of Eq. (1), which corresponds to the
273 heat capacitance of the material. Heat flows Q are correlated to the temperatures of
274 consecutive nodes through the relation:
275
T i −T i+1
Q=
276
R th, i−i+1 (2)
277
278 where Rth,i-i+1 is the thermal resistance between consecutive nodes. Each node of the thermal
279 system is placed at the mid-thickness of the material layer and thus the thermal resistance
280 between nodes is treated as half the sum of the thermal resistances of consecutive layers:
281
Ri + Ri+1
Ri−i+1=
282 2 (3)

6
283
284 Fig. 2 Schematic layout of the CPVT system and thermal network used for the formulation of the energy
285 balances.
286
287 By applying Eqs. (1)-(2) on each node of the thermal network, the following set of equations
288 is derived:
289
290  Glass cover of the PV module
291
dT gl T pv −T gl
M gl C gl =Q gl −Qloss +
292
dt R gl− PV (4)
293
294 where Qgl is the heat absorbed by the glass cover, which is equal to:
295

296
Qgl =agl Gb Aa ηopt (5)
297
298 with agl being the glass absorptivity. Qloss is the heat dissipated to ambient through external
299 convection and radiation:
300
T gl −T a
Q loss=
301
R loss (6)
302
303 As depicted in the thermal network of Fig. 2, the overall external thermal resistance Rloss
304 results by the connection in parallel of the thermal resistances due to convection and
305 radiation:
306
1 1 1
= +
307
R loss Rconv , ext R rad (7)
308

7
309 The heat transfer coefficients required for the calculation of Rconv,ext and Rrad, respectively, are
310 calculated as follows:
311
Nu MC⋅k air
hconv , ext =
312 L (8)
313
σε gl (T 4gl −T a4 )
hrad , ext=
314
T gl −T a (9)
315
316 where the values of the Nusselt number NuMC used in Eq. (8) correspond to external air flow
317 under mixed convection conditions and were calculated using the following relation:
318 n n n
319
NuMC =Nu FC + NuNC (10)
320
321 where the value of n=3 has been proposed in the literature [19]. The Nusselt number values
322 for pure forced (NuFC) and natural convection (NuNC) conditions were calculated using
323 correlations available in [20, 21].
324
325  PV laminate
326
dT pv T gl −T PV T PV −T tape
M PV C PV =Qirr −Pel + −
327
dt R gl−PV R PV−tape (11)
328
329 where Qirr is the radiation flux incident on the surface of the solar cells:
330
Qirr =( 1−a gl )( τ gl a cell ) eff Gb A a ηopt
331 (12)
332
333 and Pel is the electrical power produced by the module:
334
335
Pel=ηel Qirr (13)
336
337 The degradation in the cell efficiency due to elevated operating temperature was calculated
338 using the following relation [22]:
339

 cell   cell ,ref 1   ref Tcell  Tref   (14)

341 where the values of the temperature coefficient β were determined by the experimental
342 evaluation of the PV module described by the authors in [16].
343
344  Adhesive tape (bonding the cells to the module substrate)
345
dT tape T PV −T tape T tape −T s1 ,
M tape C tape = −
346
dt R PV−tape Rtape−s 1 (15)
347
348  PV module substrate S1

8
349
dT s1 T tape −T s1 T s 1 −T re sin
M s1 C s 1 = −
350
dt R tape−s1 R s 1−re sin (16)
351
352  Adhesive resin (bonding the PV module to the heat sink)
353
dT re sin T s 1−T re sin T re sin−T s 2
M re sin C re sin = −
354
dt R s 1−re sin R re sin−s2 (17)
355
356  Heat sink substrate S2
357
dT s 2 T re sin −T s 2 T s2 −T f
M s 2C s 2 = −
dt R re sin−s2 R conv + R cond
358 2 (18)
359
360  Cooling fluid
361
dT f T s 2−T f
M f Cf = − ṁ C f ( T f ,max−T f , in )
dt Rcond
Rconv +
362 2 (19)
363
364 The conductive thermal resistances of the solid materials appearing in the thermal balances
365 are defined Rm=tm/(km∙Am) where t, k, A are the thickness, thermal conductivity and heat
366 transfer area of each material layer m. As the nodes of the thermal network are placed in the
367 mid-thickness of each solid layer, the conductive resistance between the substrate and the
368 fluid in Eq. (18) is divided by 2, since it corresponds to half the substrate thickness. The
369 convective thermal resistance referring to the cooling-fluid layer, Eq. (18), is equal to
370 Rconv=1/(h∙A). The overall heat transfer coefficient h was determined through the use of proper
371 Nusselt correlations for rectangular channels with a constant heat-flux boundary condition, as
372 explained by the authors in [23], since Nu=h·Dh/k, with Dh being the channel hydraulic
373 diameter. A special note must be made for the calculation of the convective thermal resistance
374 Rconv, as it is dependent on the flow conditions and the geometry of the heat sink. Especially in
375 the case of the VW heat-sink configuration, the heat transfer coefficient, required for the
376 determination of the convective thermal resistance, can vary significantly between
377 consecutive sections. Besides, the heat-sink wall temperature distribution has a non-
378 monotonical distribution, as was shown by the authors in [17]. The heat capacitance term
379 appearing on the left-hand side of each layer was approximated using finite differences:
380
n n−1
dT i T i −T i
=
381 dt Δt (20)
382
383 where the subscript i corresponds to each node of the thermal resistance network, i.e to the
384 various material layers that comprise the receiver and to the cooling fluid.
385 In order to more accurately predict the temperature distribution of the various material
386 layers and consequently the receiver thermal and electrical losses, the receiver of the system
387 was discretized along its longitudinal axis in a uniform manner and the system of equations
388 (4), (11), (15)-(19) was iteratively solved for each segment (Fig. 3). All material layers were
389 segregated in 100 cells, as the use of a finer segmentation only produced negligible

9
390 differences in the obtained results. Each segment corresponding to a solid material was
391 considered to have a uniform temperature Ti. On the contrary, a linear temperature distribution
392 was considered for the cooling-fluid layer, since the segment inlet and outlet temperatures are
393 required, as well, to calculate the negative term on the right hand side of Eq. (19). The fluid
394 outlet temperature of each segment imposed as inlet temperature for the next segment, i.e.,
395 T fj−1 j
, o =T f ,i. Global quantities corresponding to the entire receiver length, e.g. the thermal and
396 electrical output, were calculated for each time-step by integrating the elemental output
397 produced by each segment.
398

399
400 Fig. 3 Segmentation of the material layers comprising the receiver, allowing for variations in the flow direction
401 to be included in the thermal-network model.
402
403 The variation of the cooling fluid thermophysical properties with temperature was
404 introduced in the model using the correlations for thermal conductivity kf, density ρf and
405 viscosity μf of water available in [24]. Eqs. (4), (11), (15)-(19) constitute a system of seven
406 equations with seven unknowns, namely the temperature Ti at each layer of the receiver. The
407 system of algebraic equations was iteratively solved using the Gauss-Seidel method. Finally,
408 the cooling fluid pressure drop required for the determination of the parasitic pumping power
409 was calculated using theoretical correlations for rectangular ducts presented by the authors in
410 [23] and which are available in [25]. The fluid average temperature in the receiver, at which
411 the fluid properties for the pressure drop calculation were evaluated, was known beforehand
412 by the solution of the thermal-network system.
413
414
415 2.1 Validation of the theoretical model
416
417 As a first step, the theoretical results produced by the transient model were compared
418 against experimental data available from the evaluation of the prototype CPVT system
419 performed by the authors in [16]. In order to ensure that the model can accurately predict the
420 transient behavior of the system, the measured system response time, i.e. the time period
421 required for the system to reach steady state once exposed to constant irradiation of high
422 intensity, was compared to that produced by the model for the same environmental
423 conditions. Besides, the predictions produced by the model in regard to the system electrical
424 and thermal efficiency were compared against experimental data. Two receiver configurations
425 were taken into consideration having different heat-sink and PV-module designs. The input
426 parameters required by the model are summarized in Table 1.
427

10
428 Table 1 Input values required by the theoretical model for the validation cases.
Model Input Symbol Values
Environmental data Gb, Vair,Ta, θ ΤΜΥ file
System aperture area Aa 1.0m2
Receiver active length Lrec 0.5m
Receiver active width (Narrow/Wide cells) Wrec 0.04m/0.06m
Number of heat sink sections (FW/VW) Nsec 1/3
Channel width (FW/VW) Wch 0.56mm/5.24-2.12-0.56mm
Wall thickness (FW/VW) Ww 0.69mm/1.00mm
Channel height (FW/VW) Hch 3.50mm/10.60mm
Heat-sink substrate thickness (FW/VW) ts 2.6mm/1.7mm
Resin thickness tresin 152.4μm
Resin thermal conductivity kresin 0.43 W/mK
Absorptance of front glass agl 0.05
Emmitance of front glass εgl 0.88
Transmittance of front glass and EVA τeff 0.92/0.90
Concentrator optical efficiency ηopt 0.57
PV module reference efficiency ηpv,ref 0.138
PV module temperature coefficient (wide/narrow cells) β 0.00461/0.00603
Irradiation uniformity factor (narrow cells) Funi 0.85
Cooling fluid volumetric flow rates (for the two
validation cases considered) V̇ tot 30/40 mL/s
Cooling fluid inlet temperature Tf,in 285K-315K
Simulation time step dt 10s
Simulation total time time 300-400s
Residual set for convergence Res 10-4
429
430 Fig. 4 depicts the evolution of the fluid (Tf), heat sink (Ths) and PV-module (TPV) maximum
431 temperatures with time, as predicted by the theoretical model. The respective outlet fluid
432 temperature evolution as measured during the system experimental evaluation is also
433 presented for comparison. The system is exposed to concentrated sunlight at t=0. As can be
434 seen the model can accurately predict the system thermal-response time and the fluid
435 temperature rise trend in both receiver variations considered. It is interesting to notice that
436 both configurations obtain steady-state operation in a relatively short time period
437 approximately equal to 150s.
438

11
439
440 Fig. 4 Thermal response of the CPVT system: (a) FW-wide cells configuration, (b) VW-narrow cells
441 configuration.
442
443 The maximum FW heat-sink temperature (Fig 4a) is higher than the fluid outlet
444 temperature by approximately 1.0 K. The respective temperature difference in regard to the
445 VW configuration is approximately equal to 6.0 K (Fig. 4b). It must be pointed out that the
446 maximum VW heat-sink temperature occurs at the first section of the device, as was proven
447 by the authors in [16]. Nevertheless, the temperature difference between PV module
448 maximum and the fluid outlet temperatures is considerably higher equal to approximately 9.0
449 K and 14.0 K for the FW and VW devices respectively. It can be therefore deduced that the
450 conductive resistance posed by the solid materials of the receiver has a far more considerable
451 effect on the receiver overall thermal resistance compare to the thermal resistance of the
452 cooling devices.
453 A further validation study was conducted to verify the accuracy of the model under steady-
454 state conditions as well. The steady-state efficiencies of the two system-receiver
455 configurations considered above, which were measured according to the procedure discussed
456 in [16], were compared against the theoretically predicted ones for the same environmental
457 and operating conditions. Fig. 5 presents the calculated and measured system thermal and
458 electrical efficiencies for two receiver configurations under a wide range of operating
459 conditions. It is evident that the system steady-state operation can be accurately predicted by
460 the model and the small influence of the operating temperature on the system efficiency is
461 clearly captured. It has therefore been proven that the model can accurately predict both the
462 system transient and steady-state operation and hence it can be used to predict the long-term
463 performance of the system.
464

12
465
466
V̇
Fig. 5 Steady-state efficiency of the CPVT system: (a) FW-wide cells configuration ( tot =30 mL/s), (b) VW-
467
V̇
narrow cells configuration ( tot =40 mL/s).
468
469 Finally, it is interesting to illustrate that the expected form of the temperature distribution at
470 the cooling-devices substrate can be produced by the analytical model. The bottom wall and
471 fluid temperature distributions along the axial (X) direction for the two cooling configurations
472 employed are presented in Fig. 6 for steady-state conditions. In reference to the actual system
473 of length and aperture area equal to 0.5m and 1.0 m 2, respectively, (Fig. 6a), the stepwise
474 temperature reduction regarding the VW configuration is predicted with the point of
475 maximum wall temperature located at the end of the first section. Due to the low number of
476 fins in the initial section of the VW configuration compared to the subsequent sections, the
477 thermal resistance locally in that section is higher and hence a peak temperature is reached.
478 This behavior has been verified by the authors both experimentally and through CFD in [17].
479 Regardless of the cooling device employed, the wall temperature at the heat-sink outlet is
480 slightly higher than the fluid outlet temperature.

481
482 Fig. 6 Longitudinal temperature distribution under steady-state conditions (Tf,i=298 K, Ta=283 K, uwind=2 m/s,
483 Gb=850 W/m2): (a) 0.5 m long (Aa=1.0 m2) system, (b) 5.0 m long 10.0 m2 system.
484
485 The longitudinal temperature distribution of a similar system with greater length and
486 respective aperture area equal to 10m2 is depicted in Fig. 6b. Once again, the temperature
487 distributions over the entire FW heat-sink length and for the most part of the VW heat-sink

13
488 length closely follow the respective distribution in the fluid, indicating the small thermal
489 resistance of the devices. In regard to the VW cooling configuration, it is evident that
490 although the wall temperature distribution still exhibits a stepwise trend, its maximum value is
491 located at the heat sink outlet. The steep increase in the distribution gradient discernible in
492 Fig. 6b at approximately Z=1.3 m occurs because in that location the flow reaches full
493 development and thus the correlations applied for the calculation of the heat transfer
494 coefficient switch from those referring to developing flow to those referring to fully-
495 developed flow, respectively.
496
497 3 Prediction of the system long-term performance
498
499 Apart from evaluating the system performance on an annual basis, three time periods in
500 different seasons of the year, namely 1st-4th of March, 21st-24th of June and 1st-4th of December,
501 have been selected in order to illustrate the effect of variable environmental conditions on the
502 system efficiency and output. Fig. 7 presents the prevailing environmental conditions for the
503 time periods in question. The climatological data (wind velocity uwind, ambient temperature
504 Ta, insolation Gb) were provided by a Meteonorm [26] TMY (Typical Meteorological Year)
505 file for Athens, Greece.
506

507
508

509
510 Fig. 7 Climatological data for the evaluation periods considered. Direct solar radiation and wind velocity: (a) 1-4
511 March, (b) 21-24 June, (c) 1-4 December. (d) Ambient temperature for the three time periods in mention.

14
512
513 A CPVT system with the manufacturing and operating characteristics of the manufactured
514 prototype evaluated by the authors in [16], yet with larger aperture area equal to 10m2, so that
515 the effect of the thermal losses can be elucidated, has been selected as a “base-case system”.
516 The specific system, the characteristics of which are summarized in Table 2, will be
517 considered as a reference in the following results, in order to highlight the effect of the
518 various parameters examined on the energetic and exergetic system output and efficiency.
519
520 Table 2 Technical specifications and operating parameters of the base-case system.
Model Input Values
System aperture area Aa [m2] 10.0m2
Receiver active length Lrec [m] 5.0
Receive active width Wrec [m] 0.06
Heat sink design: FW (Wch-Ww-Hch) [mm] 0.56-0.69-3.5
Resin thickness tresin [μm] 152.4
Resin thermal conductivity kresin [W/mK] 0.43
Absorptance of front glass agl 0.05
Emmitance of front glass εgl 0.88
Transmittance of front glass and EVA τeff [-] 0.92
Concentrator optical efficiency ηopt [-] 0.57
PV module reference efficiency ηpv,ref [-] 0.138
PV module temperature coefficient β [-] 0.00461
V̇
Cooling fluid volumetric flow rate tot [mL/s] 30.0
Cooling fluid inlet temperature Tf,in [K] 298.0
521
522 As an initial step, it must be examined whether the performance of the base-case system
523 varies considerably depending on the incorporated cooling device. For this reason, the
524 influence of the cooling-device design on the system efficiency is presented in Fig. 8a. It is
525 evident that the effect of the heat-sink configurations on the efficiency of the base-case
526 system is negligible. The system electrical efficiency, for the case where the FW cooling
527 device is employed, is higher by up to 1.2%-points compared to the respective for the VW
528 device, whereas the thermal efficiency remains approximately equal regardless of the
529 incorporated device. According to the formulation of the theoretical model, the overall heat
530 losses from the system receiver are calculated considering the difference between the average
531 PV glass, which, in essence, is designated by the heat-sink thermal resistance, and ambient
532 temperatures, refer to Eq. (6). It has been established through a comparative experimental
533 investigation performed in [18] that the thermal resistance values for the two heat-sink
534 configurations based on the average wall temperature do not differ considerably and thus the
535 similar overall system performance is justified. Consequently, it is adequate to analyze the
536 performance of the base-case system considering only the FW heat sink as cooling device. It
537 is essential to point out that the effect of temperature non-uniformity on the electrical
538 efficiency could not be predicted by the present model as there are no performance data
539 available from the manufacturer of the actual PV modules with regard to operation at non-
540 uniform temperature.
541

15
542
543 Fig. 8 Performance characteristics of the base-case system in March: (a) efficiency and (b) power production.
544
545 The power output of the base-case system is depicted in Fig. 8b. As can be seen, the major
546 percentage of the irradiation incident on the receiver is exploited either as electrical or thermal
547 power, with the thermal losses representing less than 3% of the thermal output. For the
548 environmental conditions corresponding to the time period in March, the base-case system
549 can produce up to 3.5 kW of thermal power and 0.5 kW of net electrical power.
550 It was demonstrated during the experimental evaluation of the prototype CPVT system that
551 the influence of the water volumetric flow rate on the system efficiency could not be
552 elucidated, as the difference in the system performance was in the order of the experimental
553 uncertainty [16]. The dynamic model has been therefore utilized to illustrate the specific
554 effect on the efficiency of the base-case system, which has a ten-fold larger aperture area than
555 the prototype system. The time period in June was considered for the simulations so that the
556 environmental conditions are similar to the actual ones prevailing during the experimental
557 evaluation.
558 Fig. 9 reveals that the value of the water volumetric flow rate has only a minor effect on
559 the system energetic efficiency for both the cases of the FW (Fig. 9a) and VW (Fig. 9b)
560 configurations. An increase of the flow rate from 20.0 to 40.0 mL/s leads to an increase of the
561 electrical and thermal efficiencies by approximately 6% and less than 2%, respectively. The
562 thermal losses from the system receiver are of small magnitude due to its confined overall
563 dimensions and the exceptional performance of the cooling devices. Additionally, the
564 conductive thermal resistance posed by the solid materials has a more significant contribution
565 on the receiver overall thermal resistance compared to the heat-sink convective resistance. It
566 must also be mentioned that even for a system with an overall length of 5.0m, such as the
567 base-case system, the effect of the parasitic pumping power on the system efficiency is
568 negligible; however this cannot be considered a general remark, as linear CPVT systems of
569 larger scale can have overall lengths of several meters.
570

16
571
572 Fig. 9 Effect of the water volumetric flow rate on the system energetic efficiency. The results refer to the base-
573 case system in June employing: (a) the FW and (b) VW devices, respectively.
574
575 Moreover, it is of interest to examine the effect of the water inlet temperature on the
576 system efficiency, as the CPVT system is most likely intended to be coupled with heat
577 storage, e.g. in a domestic hot-water system, and thus to be fed with water of elevated
578 temperature from the storage tank during its daily operation. It is evident from Fig. 10,
579 referring to the base-case system, that water entering the system with an elevated temperature
580 has a considerable detrimental effect on both the system thermal and total efficiency. In fact,
581 the decrease is slightly more substantial in the case of total efficiency (black line) compared to
582 the thermal efficiency (red line), as the increased receiver temperature leads to increased
583 thermal losses and, in addition, to reduced electrical production due to the elevated PV-
584 module temperature. For example, for increase of the water inlet temperature from 298.0 K to
585 338.0 K and referring to the environmental conditions corresponding to December (Fig. 10b)
586 the total efficiency decreases by approximately 10%, while the thermal efficiency decreases
587 by approximately 8%. The relative deterioration in the system efficiency caused by the
588 increased water inlet temperature is also influenced by the environmental conditions, as the
589 effect is stronger at time periods with low or fluctuating solar irradiation values, e.g. as is
590 evident for the curves corresponding to the 4th of March (Fig. 10a) and the 4th of December
591 (Fig. 10b).

592
593 Fig. 10 Effect of the water inlet temperature on the efficiency of the base-case system: (a) March, (b) December.

17
594
595 4 System optimization
596
597 The objective of the following paragraph is to designate the value range of operating
598 parameters, for which the system attains the most effective performance from a second-law of
599 thermodynamics point of view. The latter allows for a straightforward comparison between
600 thermal and electrical output. In addition, the influence of the performance characteristics of
601 key sub-components on the system exergetic performance will also be elucidated. A
602 parametric study is conducted to determine the factors that can lead to increase of the system
603 net exergetic efficiency on a daily and annual basis having as reference the exergetic
604 efficiency of the base-case system (see Table 2).
605
606 4.1 Exergy analysis
607
608 In a conventional energy analysis, it is assumed that the heat gain of the cooling fluid can
609 be completely converted to useful work. However, this assumption is invalid as there is an
610 upper limit to the conversion of heat to useful work, known as the Carnot efficiency and thus
611 exergy, i.e. the maximum useful work that can be produced until a thermodynamic system
612 reaches thermal equilibrium with a reference environment is defined as:
613

614
Ė x th=Qth 1−
( ) Tc
Th
(21)
615
616 where Qth is the available thermal energy, Th is the temperature at which the heat is available
617 and Tc is the dead-state temperature where the useful work that can be produced is equal to
618 zero. The general exergy balance of a thermodynamic system with inbound ( Ė xin ) and
619 outbound exergy flows ( Ė x out ) can be written as:
620
∑ Ė x in−∑ Ė x out=∑ Ė x dest
621 i i i (22)
622

623 where
Ė x
dest is the destroyed exergy within the system. Exergy is destroyed when a process
624 involves a temperature change or fluid friction within a duct. The term is also referred to as
625 system irreversibility
Ė x dest = İ . The exergy destruction is proportional to the entropy increase
626 of the system [27]:
627
1
Ṡ gen =
Tc
( Ė x in− Ė x out )
628 (23)
629
630 In the case of solar applications, the inbound stream of exergy or exergy fuel is the solar
631 irradiation, which according to Petela [28] is:
632

( ( ))
4
4 Ta 1 Ta
Ė x sun=Gb A a 1− +
3 T sun 3 T sun
633 (24)
634

18
635 where Tsun is the sun temperature equal to 4350 K [29]. Eq. (24) provides the maximum work
636 that can be produced from an isotropic blackbody radiation at T sun [30]. The useful exergy
637 streams that comprise the outbound exergy stream are the extracted heat and electrical power:
638

639
Ė x th=Qth 1−
( Ta
T f ,o ) (25)
640

641
Ė x el=Pel (26)
642
643 As can be directly inferred from Eq. (26), the exergy flow associated with photovoltaic electricity is
644 equal to the produced electrical power [29, 31, 32]. The destroyed exergy rate or system irreversibility
645 comprises four terms due to the optical losses, the temperature difference between the receiver and the
646 sun temperature, the thermal losses and the coolant pressure drop respectively:
647

( ( ))
4
4 Ta 1 Ta
Ė x opt , loss=G b A a ( 1−ηopt )( 1−τα eff ) 1− +
3 T sun 3 T sun
648 (27)
649

650
Ė x ΔΤ , CPVT−sun=Gb A a T a
( 1
−
1
T pv T sun ) (28)
651

652
( )
Ė x th, loss=Q th ,loss 1−
Ta
T pv
(29)
653

654
Ė x pump=P pump (30)
655
656 The Petela factor was used for the calculation of the exergy flow associated with optical
657 losses, as Eq. (27) refers to conversion of solar irradiation to useful work. Furthermore, the
658 term referring to the temperature difference between the receiver and the sun surface, Eq.
659 (28), exists, as the heat available at the temperature of the sun surface, which is considered as
660 the “hot reservoir”, could be ideally converted to useful work without having to transition to a
661 lower temperature level Tpv, a process accompanied by exergy loss [33]. By combining Eqs.
662 (24)-(30) and Eq. (22) the exergy flow balance for the CPVT system takes the following
663 form:
664

( ( )) ( )
4
4 Ta 1 Ta Ta
Gb A a 1− + =Qth 1− +P el−
3 T sun 3 T sun T f ,o

( ( )) ( ) ( )
4
4 Ta 1 Ta 1 Ta Ta
−Gb A a ( 1−ηopt )( 1−τα eff ) 1− + −Gb A refl T a − −Qth ,loss 1− −P pump
3 T sun 3 T sun T pv T sun T gl
665
666
667
İ (31)
668

19
669 where İ is the destroyed exergy or system irreversibility. Finally, the exergetic efficiency of
670 the integrated system is defined as the useful, net exergy flows to the sun input exergy:
Ė xth + Ė x el− Ė x pump
η II =
671
Ė x sun (32)
672
673 An additional quantity which has been used for the optimization of solar applications [34] is
674 the entropy generation number, which quantifies the system irreversibility and is defined as:
675
T a Ṡ gen
N s=
676
A a Gb (33)
677

678 where the entropy generation rate Ṡ gen can be calculated by Eq. (23). It is obvious that the
679 system optimal operation is achieved when the entropy generation number is minimized.
680
681 4.2 Effect of the system parameters on overall performance
682
683 The net exergetic efficiency is reckoned as a suitable quantity to highlight the significance
684 of the operating conditions and the various key-components characteristics on the system
685 overall performance, since it allows the straightforward treatment of the produced thermal and
686 electrical power, which are appropriately reduced to exergy streams of equal “value”.
687 Fig. 11 depicts the variation in the system exergetic efficiency caused by the increase of
688 the water volumetric flow rate. As was shown in the previous section (see Fig. 9), the
689 energetic efficiency of the system is relatively independent of the volumetric flow rate value.
690 However the water flow rate appears to have a considerable effect in regard to the system
691 exergetic efficiency, and, in fact, increase of the volumetric flow rate leads to decrease of the
692 overall exergetic efficiency. Although a higher volumetric flow rate tends to decrease the PV
693 module operating temperature and thus increase its efficiency, it also leads to reduced fluid
694 temperature at the outlet of the system Tf,o. According to Eq. (25), the exergy stream
695 associated with thermal power, apart from the heat output, is also depended on the
696 temperature at which the heat is delivered and hence a lower fluid outlet temperature
697 decreases exergetic efficiency. It is interesting to notice that the system exergetic efficiency is
698 higher in the winter period due to the lower ambient temperature values, which allow for a
699 greater temperature difference between the fluid at the system outlet and the system dead-
700 state. Decrease of the water volumetric flow rate from 40.0 mL/s to 20.0 mL/s can lead to an
701 increase of the system exergetic efficiency up to approximately 20% for a day with high
702 insolation and low ambient temperature.
703

20
704
705 Fig. 11 Effect of water volumetric flow rate on the system exergetic efficiency in regard to the base-case system:
706 (a) June, (b) December
707
708 Likewise, as was shown in Fig. 10, increasing the temperature of the water entering the
709 system has an adverse effect on the system energy efficiency. On the contrary, increased
710 water inlet temperature enhances the system exergetic performance regardless of the
711 environmental conditions, as illustrated in Fig. 12. The enhanced exergetic performance
712 despite the increased system operating temperature constitutes proof that, from a second-law
713 analysis point of view, it is beneficial to provide heat at elevated temperature despite the fact
714 that the latter is associated with increased thermal losses and decreases of the system electric
715 efficiency as well. Nevertheless, this conclusion is not of general validity and can only be
716 applied to systems having performance characteristics similar to the developed prototype
717 CPVT system, where the thermal output considerably exceeds the electrical output. As shown
718 in Fig. 12b, increasing the fluid inlet temperature from 298.0 K to 338.0 K causes an increase
719 of approximately 25% on the system exergetic efficiency that can be achieved in December.
720 In general, the results produced by the dynamic model referring to the base-case system
721 have established that the exergetic efficiency achieved is relatively low and lies in the range
722 9.0-15.5% depending on the operating and environmental conditions. It was proven during the
723 system experimental evaluation [16] that the main output is in the form of thermal power.
724 However, as has been already discussed the exergy stream associated with the thermal output
725 is dependent on the temperature level of the delivered heat. Since the intrinsic orientation of
726 the application is the production of relatively low-temperature heat, in the range of 333-363
727 K, the useful thermal exergy stream is considerably decreased. On the contrary the electrical
728 exergy stream is directly associated with the produced power (see Eq. 26) and hence it is
729 essential to increase the system electrical output in order to achieve high performance from an
730 exergetic point of view.
731

21
732
733 Fig. 12 Effect of fluid inlet temperature on the system exergetic efficiency in regard to the base-case system: (a)
734 March, (b) December.
735
736 It was established during the system experimental evaluation that the optical losses have
737 the most significant contribution on the reduction of the overall efficiency. Fig. 13 elucidates
738 the influence that a higher optical quality of the parabolic trough would have on the exergetic
739 efficiency. Higher system optical quality leads to augmented irradiation on the system
740 receiver and influences both the useful thermal and electrical exergy streams in a beneficial
741 manner. The system can reach exergetic efficiencies up to 18% and 21% in the summer (Fig.
742 13a) and winter (Fig. 13b) periods, respectively, for an optical efficiency of 85%. These
743 values are significantly larger compared to the 12% achieved by a system with the optical
744 quality of the manufactured prototype (see Fig. 12b). It is easily perceivable that increased
745 optical quality is of vital importance for the efficient performance of a CPVT system.
746

747
748 Fig. 13 Effect of optical efficiency on the system exergetic efficiency in regard to the base-case system: (a) June,
749 (b) December.
750
751 The relatively low efficiency of the solar-cell modules (ηel=13.8%) has been identified as a
752 weak point of the prototype system [16]. Fig. 14 presents the influence of the module
753 electrical efficiency on the system exergetic efficiency. Increasing the portion of irradiation
754 converted directly to electricity by the solar cells will certainly reduce the excess heat
755 available for extraction and this effect will also be propagated on the respective exergy
756 stream. It can be observed that a 5%-points absolute increase of the electrical efficiency leads

22
757 to an increase of the exergetic efficiency approximately equal to 2%, regardless of the
758 environmental conditions. The graphs of Fig. 14 include an additional curve that corresponds
759 to a system with high, yet certainly achievable, optical and (reference) electrical efficiencies,
760 equal to 75% and 25%, respectively. As can be seen, the system is able to reach exergetic
761 efficiencies up to 24% for low values of the ambient temperature, i.e. in March (Fig. 14a) and
762 December (Fig. 14c). By comparing Figs. 13b and 14c, it can be deduced that, for a day with
763 high insolation in the winter, incorporating high-efficiency cells to a precisely manufactured
764 CPVT system increases the exergy efficiency from 17.5% to 23%. The general conclusion
765 can be drawn that achieving high manufacturing and materials quality in reference to the key
766 components of a CPVT system, such as the PV modules and the optical device, can cause an
767 approximately two-fold increase of its exergetic efficiency (see Fig. 12 for comparison to the
768 base-case system).
769

770
771

772
773 Fig. 14 Effect of the PV module efficiency on the system exergetic efficiency in regard to the base-case system:
774 (a) March, (b) June and (c) December.
775
776 The monthly distribution of the base-case system exergy output is shown in Fig. 15a and
777 allows the specification of the time periods were the system achieves its most effective
778 performance. An initial observation is that the electrical exergy exceeds the respective thermal
779 throughout the year and obtains maximum values in the summer months, 2.5 times higher
780 compared to the respective thermal exergy (in July and August). It is interesting to notice that
781 the peak electrical exergy is delivered by the system in July, whereas the maximum thermal

23
782 exergy is obtained in May. This variation in regard to the two useful exergy streams occurs on
783 grounds that the electrical exergy is primarily depended on the irradiation intensity, whereas
784 thermal exergy is significantly influenced by the ratio of the fluid outlet to the ambient
785 temperature and hence the combination of these two factors in the spring and autumn tend to
786 enhance the thermal exergy produced by the system and render it comparable to that produced
787 during the summer months.
788 Fig. 15b illustrates the monthly exergy output of the “high-quality” system for a water
789 volumetric flow rate of 20.0 mL/s. The annual produced thermal and electrical exergy by the
790 “high-quality” system are approximately 1.9 and 2.6 times higher in comparison to the
791 respective of the base-case system. In the case of the “high-quality” system the peak electrical
792 exergy efficiency is still achieved in July, however, in contrast to the base case system, the
793 respective thermal exergy is achieved in June. In general, it can be deduced that the thermal
794 exergy output of the “high-quality” system is slightly less sensitive to the ratio of the outlet
795 fluid to ambient temperature and for this reason it achieves higher values during the summer
796 months, unlike the base-case system. Moreover, the electricity to thermal exergy ratio is
797 larger in the “high-quality” system due to the increased efficiency of the solar-cell modules.
798

799
800

801
802 Fig. 15 Monthly exergy output of the CPVT system: (a) base-case system (b) “high-quality” system (ηopt=0.75,
803 ηel=0.25) with water volumetric flow rate
V̇ tot =20.0 mL/s.
804
805 The high-quality system is also comparatively evaluated against the base-case system in
806 terms of destroyed exergy rate and entropy generation, so that the main causes of system

24
807 irreversibility can be identified. Fig. 16 illustrates the main exergy destruction mechanisms
808 for the two system variations in question under variable environmental conditions. It is clearly
809 demonstrated that the main cause of system irreversibility are the optical losses, as
810 approximately half of the available solar irradiation exergy is destroyed in the optical device.
811 On the contrary, the exergy destroyed due to the system thermal losses is two orders of
812 magnitude lower, indicating thus the high system thermal performance. The destroyed exergy
813 rate closely follows the trend of the solar irradiation intensity and the system output (see Figs.
814 7-8b) and consequently it is maximized in time periods with high insolation. As also revealed
815 by Fig. 16, increasing the system optical efficiency reduces the respective exergy destruction
816 rate by approximately 53% regardless of the environmental conditions. However, as depicted
817 by the curves in red color of Fig. 16 the destroyed exergy stream that corresponds to the
818 thermal losses in fact increased in the high quality system. This variation is justified as the
819 solar irradiation incident on the receiver and therefore the additional heat available for
820 extraction are producing higher temperature gradients in the receiver materials and increased
821 thermal losses. Apparently, the benefit due to the decreased exergy destruction in the optical
822 device exceeds by far the negligible increase in the exergy destruction stream referring to the
823 thermal losses.
824

825
826

827
828 Fig. 16 Destroyed exergy rate referring to the base-case and high-quality systems: (a) March, (b) June, (c)
829 December.
830

25
831 The system entropy generation rate is presented in Fig. 17 in the form of either absolute
832 values or dimensionless entropy generation number Ns. In all cases the entropy generation rate
833 is lower for the “high-quality” system, where the exergy destruction due to optical losses and
834 poor electrical performance is considerably reduced. However, it can be seen that the entropy
835 generation rate closely follows the trend of the solar irradiation intensity and consequently of
836 the system output (see Figs. 7 and 8b). The entropy generation number Ns (Eq. 33) appears as
837 a more appropriate quantity for the clarification of the system irreversibilities, as it is
838 independent of the solar irradiation intensity and obtains comparable values regardless of the
839 environmental conditions. As can be observed, the entropy generation number for the base-
840 case system lies in the range 0.80-0.82 for all the time periods considered. In regard to the
841 “high-quality” system, the same quantity is reduced to values in the range 0.69-0.72. It is also
842 interesting to notice that the entropy generation number, which in general exhibits a flatter
843 profile during the day, obtains its minimum daily value when the value of the absolute
844 entropy generation rate is maximized, i.e. when the system output is maximized. The entropy
845 generation number decrease is steeper for the high-quality system. It becomes therefore
846 evident that the minimization of the entropy generation number indicates the time period of
847 the day where the system obtains its most effective operation and coincides with the
848 maximization of the exergetic efficiency ηΙΙ. Hence, it is recommended as a suitable index for
849 CPVT applications.
850

851
852

853
854 Fig. 17 Entropy generation rate for the base-case and high-quality systems: (a) March, (b) June and (c)
855 December.

26
856
857
858 5 Conclusions
859
860 The prediction of the CPVT system long-term energetic and exergetic performance was
861 performed in this study using a dynamic theoretical model. The simulations clearly
862 demonstrated the applicability of the manufactured cooling devices to large-scale CPVT
863 systems, as the thermal losses were found to be of small magnitude in reference to systems
864 oriented towards the domestic sector. Additionally, the parasitic pumping power had a
865 negligible effect on the system efficiency for the heat-sink geometrical parameters considered.
866 The analysis of the system exergetic performance revealed that, for the performance
867 characteristics of the prototype system, namely substantially higher thermal than electrical
868 output, it is beneficial to operate the system using low flow rate and elevated fluid inlet
869 temperature, in order to ensure that heat is delivered at a high temperature level and thus the
870 exergetic efficiency is increased. The basic constituents characteristics identified to have a
871 significant effect on the system exergetic performance are the solar-cell module efficiency
872 and the optical quality of the parabolic trough, with the optical losses being clearly established
873 as the primary source of exergy destruction. Increasing the aforementioned system quality
874 characteristics to higher yet achievable values, e.g. ηopt=0.75 and ηel=0.25, can lead to an
875 increase of the exergetic efficiency from aprroximately 12% to 24%. An interesting general
876 remark that can be made is that the cooling fluid flow rate and inlet temperature can be
877 utilized as free parameters to adjust the system thermal and electrical output, as the latter two
878 quantities are of interdependent nature, based on the respective thermal and electrical loads
879 required by a specific application. Overall, it has been clearly demonstrated that the developed
880 numerical model is capable of pointing out the crucial design and operating parameters
881 affecting the energy output, as well as energetic and exergetic efficiencies of both small and
882 large-scale CPVT systems.
883
884
885 References
886
887 [1] A. Hazi, G. Hazi, R. Grigore, V. Sorin, Opportunity to use PVT systems for water heating in industry, Appl.
888 Therm. Eng. (2013).
889
890 [2] N. Amrizal, D. Chemisana, J.I. Rosell, Hybrid photovoltaic-thermal solar collectors dynamic modeling,
891 Appl. Energy. 101 (2013) 797–807.
892
893 [3] H.A. Zondag, D.W. De Vries, W.G.J. Van Helden, R.J.C. Van Zolingen, a. a. Van Steenhoven, The thermal
894 and electrical yield of a PV-thermal collector, Sol. Energy. 72 (2002) 113–128.
895
896 [4] T.T. Chow, Performance analysis of photovoltaic-thermal collector by explicit dynamic model, Sol. Energy.
897 75 (2003) 143–152.
898
899 [5] M.J. O'Leary, L.D. Clements, Thermal-electric performance analysis for actively cooled, concentrating
900 photovoltaic systems, Sol. Energy 25 (1980) 401-406.
901

27
902 [6] H. Helmers, A. Bett, J. Parisi, C. Agert, Modeling of concentrating photovoltaic and thermal systems, Prog.
903 Photovolt.: Res. Appl. (2012).
904
905 [7] F. Calise, A. Palombo, L. Vanoli, A finite-volume model of a parabolic trough photovoltaic/thermal
906 collector: Energetic and exergetic analyses, Energy 46 (2012) 283–294.
907
908 [8] A. Buonomano, F. Calise, G. Ferruzzi, L. Vanoli, A novel renewable polygeneration system for hospital
909 buildings: design, simulation and thermo-economic optimization, Appl. Therm. Eng. (2014).
910
911 [9] A. Buonomano, F. Calise, M. Dentice d' Accadia, L. Vanoli, A novel solar trigeneration system based
912 onconcentratingphotovoltaic/thermal collectors. Part 1: Design and simulation model, Energy 61 (2013) 59-71.
913
914 [10] A. Kribus, D. Kaftori, G. Mittelman, A. Hirshfeld, Y. Flitsanov, A. Dayan, A miniature concentrating
915 photovoltaic and thermal system, Energy Convers. Manag. 47 (2006).
916
917 [11] T. Kerzmann, L. Schaefer, System simulation of a linear concentrating photovoltaic system with an active
918 cooling system, Renew. Energy. 41 (2012) 254–261.
919
920 [12] C. Renno, F. Petito, Design and modeling of a concentrating photovoltaic thermal (CPV/T) system for a
921 domestic application, Energy Build. 62 (2013) 392–402.
922
923 [13] H. Helmers, K. Kramer, Multi-linear performance model for hybrid (C)PVT solar collectors, Sol. Energy 92
924 (2013) 313–322.
925
926 [14] J.S. Coventry, A solar concentrating photovoltaic/thermal collector, PhD thesis, Australian National
927 University, Canberra, 2004.
928
929 [15] G. Kosmadakis, D. Manolakos, G. Papadakis, Simulation and economic analysis of a CPV/thermal system
930 coupled with an organic Rankine cycle for increased power generation, Sol. Energy 85 (2011) 308–324.
931
932 [16] I.K. Karathanassis, E. Papanicolaou, V. Belessiotis, G.C. Bergeles, Design and experimental evaluation of a
933 parabolic-trough Concentrating Photovoltaic/Thermal (CPVT) system. Ren. Energy 101 (2018) 467-483.
934
935 [17] I.K. Karathanassis, E. Papanicolaou, V. Belessiotis, G.C. Bergeles, Experimental and numerical evaluation
936 of an elongated plate-fin heat sink with three sections of stepwise varying channel width, Int. J. Heat Mass
937 Transfer 84 (2015) 16-34
938
939 [18] I.K. Karathanassis, Development and optimization of a concentrating photovoltaic/thermal cogeneration
940 system, PhD Thesis, National Technical Univeristy of Athens, Greece, 2015
941
942 [19] T.S. Chen, B.F. Armaly, Mixed convection in external flows, in: S. Kakac, R.K. Shah, W. Aung (Eds.),
943 Handbook of Single Phase Convective Heat Transfer, Wiley & sons, New York, 1987.

28
944 [20] R.H. Pletcher, External Flow Forced Convection, in: S. Kakac, R.K. Shah, W. Aung (Eds.), Handbook of
945 Single Phase Convective Heat Transfer, Wiley & sons, New York, 1987.

946 [21] Y. Jaluria, Basics of Natural Convection, in: S. Kakac, R.K. Shah, W. Aung (Eds.), Handbook of Single
947 Phase Convective Heat Transfer, Wiley & sons, New York, 1987.

948 [22] E. Skoplaki, J.A. Palyvos, On the temperature dependence of photovoltaic module electrical performance:
949 A review of efficiency/power correlations, Sol. Energy. 83 (2009) 614–624.

950 [23] I.K. Karathanassis, E. Papanicolaou, V. Belessiotis, G.C. Bergeles, Multi-objective design optimization of a
951 micro heat sink for Concentrating Photovoltaic/Thermal (CPVT) systems using a genetic algorithm, Appl.
952 Therm. Eng. 59 (2013) 733-744.
953
954 [24] B. Gebhart, Y. Jaluria, R.L. Mahajan, B. Sammakia Buoyancy-induced Flows and Transport, 1 st ed.,
955 Hemisphere Publishing Co., New York, 1988.
956
957 [25] R.D. Blevins, Applied Fluid Dynamics Handbook, Van Nostrand Reinhold Company, New York, 1984.
958
959 [26] Meteotest, Meteonorm Version 6.1, Bern, Switzerland,2008
960
961 [27] A. Bejan, Entropy Generation Minimization, 1st ed., CRC Press, New York, 1996.
962
963 [28] R. Petela, Exergy of heat radiation, J. Heat Transfer 86 (1964) 187-192.
964
965 [29] F. Calise, A. Palombo, L. Vanoli, A finite-volume model of a parabolic trough photovoltaic/thermal
966 collector: Energetic and exergetic analyses, Energy 46 (2012) 283–294.
967
968 [30] E. Saloux, A. Teyssedou, M. Sorin, Analysis of photovoltaic (PV) and photovoltaic/thermal (PV/T) systems
969 using the exergy method, Energy Build. 67 (2013) 275–285.
970
971 [31] S. Agrawal, G.N. Tiwari, Energy and exergy analysis of hybrid micro-channel photovoltaic thermal module,
972 Sol. Energy. 85 (2011) 356–370.
973
974 [32] A. Tiwari, S. Dubey, G.S. Sandhu, M.S. Sodha, S.I. Anwar, Exergy analysis of integrated photovoltaic
975 thermal solar water heater under constant flow rate and constant collection temperature modes, Appl. Energy. 86
976 (2009) 2592–2597.
977
978 [33] A. Bejan, Extraction of exergy from solar collectors under time-varying conditions, Int. J. Heat Fluid Flow.
979 3 (1982) 67–72.
980
981 [34] E. Torres-Reyes, J.J. Navarrete-Gonzalez, B.A. Ibarra-Salazar, Thermodynamic method for designing
982 dryers operated by flat-plate solar collectors, Renew. Energy 26 (2002) 649–660.
983
984

29
985

30