Renewable Energy 74 (2015) 139e147

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

A detailed nonuniform thermal model of a parabolic trough solar

receiver with two halves and two inactive ends
Ze-Dong Cheng, Ya-Ling He*, Yu Qiu
Key Laboratory of Thermo-Fluid Science and Engineering of MOE, School of Energy and Power Engineering, Xi’an Jiaotong University, Xi’an, Shaanxi 710049,
China

a r t i c l e i n f o a b s t r a c t

Article history: In this paper a detailed one dimensional nonuniform thermal model of a parabolic trough solar collector/
Received 11 December 2013 receiver is presented. The entire receiver is divided into two linear halves and two inactive ends for the
Accepted 31 July 2014 nonuniform solar radiation, heat transfers and fluid dynamics. Different solar radiation and heat transfer
Available online
modes can be taken into consideration for these four different regions respectively. This enables the
study of different design parameters, material properties, operating conditions, fluid flow and heat
Keywords:
transfer performance for the corresponding regions or the whole receiver. Then the nonuniform model
Solar energy
and the corresponding uniform thermal model are validated with known performance of an existing
Parabolic trough solar collector
Nonuniform solar flux
parabolic trough solar collector/receiver. For applications, the uniform thermal model can be used to
Heat transfer analysis quickly compute the integral heat transfer performance of the whole PTC system while the nonuniform
Nonuniform thermal model thermal model can be used to analyze the local nonuniform solar radiation and heat transfer perfor-
mance characteristics and nonuniform heat transfer enhancements or optimizations. Later, it could also
be effectively used with an intelligent optimization, such as the genetic algorithm or the particle swarm
optimization, to quickly evaluate and optimize the characteristics and performance of PTCs under series
of nonuniform conditions in detail.
© 2014 Elsevier Ltd. All rights reserved.

1. Introduction Since the 1970s, numerous thermal models have been proposed
for the heat transfer analysis of this process, as it is very important
The parabolic trough solar collector is currently one of the most for quickly calculating thermal losses or collector efficiencies, sizing
mature and field tested solar thermal technologies for large-scale the solar field, evaluating the effects of collector degradation and
exploitations of solar energy [1e3]. It uses mirrored surfaces of a heat transfer fluid flow rate control strategies on overall solar field
linear parabolic reflector to focus solar radiation onto an evacuated performance [8]. The most important ones are the study of Dudley
tubular receiver placed along the focal line of the parabola. The et al. [9], Forristall [10], Padilla et al. [11] and Kalogirou [12]. In the
receiver mainly includes an inner absorber tube surrounded by an 1990s, Dudley et al. [9] developed a significant one dimensional
outer glass cover and supported brackets [1,2]. The concentrated steady-state analytical model for an LS-2 SEGS parabolic trough
solar radiation reaching the absorber tube heats the heat transfer solar collector after many experiments were performed on the
fluid that flows through it, thus transforming the radiant energy AZTRAK rotating test platform at Sandia National Laboratories. The
from the sun into useful thermal energy [2e7]. In the solar field, results showed good agreements between the theoretical and
parabolic trough solar collectors are always built in modules to experimental collector efficiencies or thermal losses under the
operate at up to temperatures of 400
C, and synthetic oil is known tested conditions. Based on this, Forristall [10] presented a
commonly used as the heat transfer fluid. The whole solar to more detailed one-dimensional or two-dimensional heat transfer
thermal conversion process is coupled with large concentrated model for short and long receivers implemented in the engineering
solar radiations, conjugated heat transfers and fluid dynamics, equation solver, to determine the parabolic trough solar collector/
including all modes of heat transfer and temperature-dependent receiver thermal performance with various geometric parameters,
properties. fluid flows, material properties and operating conditions. In recent
years, Padilla et al. [11] proposed a one dimensional numerical heat
transfer analysis of a parabolic trough solar collector/receiver by
* Corresponding author. Tel.: þ86 (0)29 82665930; fax: þ86 (0)29 82665445. applying the mass and energy balance in several segments of the
E-mail addresses: yalinghe@mail.xjtu.edu.cn, hylepe@gmail.com (Y. Qiu).

http://dx.doi.org/10.1016/j.renene.2014.07.060
0960-1481/© 2014 Elsevier Ltd. All rights reserved.
140 Z.-D. Cheng et al. / Renewable Energy 74 (2015) 139e147

steel absorber and glass envelope, in which the thermal interaction distributions and temperature-dependent properties) and these
between neighboring surfaces for thermal radiation losses was special considerations made by other researchers (i.e., the
implemented by a comprehensive radiative analysis. Kalogirou [12] nonuniform performance improvements on one half of the
also developed a detailed thermal model of a parabolic trough solar absorber/cover) mentioned above. The entire receiver is divided
collector/receiver using the engineering equation solver. The ther- into two linear halves and two inactive ends, and different solar
mal analysis of the collector receiver takes into consideration all radiation and heat transfer modes can be taken into consideration
modes of heat transfer. Up to now, all these thermal models have for these four different regions respectively. The model is then
been thoroughly tested, experimentally validated and well applied validated with known performance of an existing parabolic trough
for general parabolic trough solar collector systems. Usually, most solar collector, and some further studies for nonuniform heat
of them assumed that the solar energy flux density, wall temper- transfer enhancements or optimizations based on the proposed
ature or physical properties were uniform for the whole circum- model are also discussed.
ferential direction at a cross-section of the receiver.
In fact, the solar flux on the collector absorber side opposite to
2. Nonuniform thermal model
the parabolic trough collector receives large amounts of concen-
trated solar radiation but there is the opposite for the other side of
A detailed schematic diagram of the parabolic trough evacuated
the absorber [13e22], and it is no concentrated solar radiations on
tubular receiver is presented in Fig. 1. It mainly includes an inner
the two inactive ends, i.e., the part includes bellows and glass-to-
stainless steel absorber tube with a solar-selective absorbing
metal joints at either end of the receiver. Because of this, the fluid
coating on its outer surface, an outer glass cover with anti-reflective
inside the tube is heated asymmetrically and thus the temperature
(AR) coatings on its both surfaces, an annular space, getters, bel-
distributions and temperature-dependent properties are nonuni-
lows and glass-to-metal joints (i.e., the inactive ends). The
form. Some significant studies have been published recently to
absorbing coating combines a high absorptance for solar radiation
investigate the nonuniform heat transfers and fluid dynamics in the
and a low thermal emittance to reduce thermal reradiation from
parabolic trough solar receivers by numerical models (three-
the outer absorber surface [10], while the AR coatings are used to
dimensional nonuniform computational fluid dynamics (CFD)
reduce Fresnel reflective losses at the glass surfaces [1,2]. The glass
models) [4,15,19e37], but it is revealed that the numerical methods
cover and the annular vacuum space are used to significantly
need great computation time. Especially, the time-consuming
reduce convection heat losses and protect the coating from
computational process is a critical question for an intelligent opti-
oxidation. The vacuum is initially maintained at about 0.0001 Torr
mization on the PTC performance. For example, it may take several
determined by the Knudsen gas conduction range [5,21], but it may
weeks or months to make a completed intelligent optimization
be degraded with time. The getters are placed in the annular vac-
based on these complex three-dimensional nonuniform CFD
uum space to absorb residual gases that permeate into the annulus
models with the genetic algorithm (GA) or the particle swarm
over time and indicate the status of vacuum, while the vacuum-
optimization (PSO). Alternatively, a new simplified and quick
tight enclosure is sealed by glass-to-metal joints at both ends
thermal model that can take into account these nonuniform con-
with bellows that accommodate for thermal expansion difference
ditions to some extent may also be another feasible method to
between steel and glass materials [1,27]. The space outside the
apply.
bellows at either end provides a place to attach the support
Furthermore, to develop an energy-efficient receiver for solar
brackets.
parabolic trough concentrators, various special considerations on
The nonuniform heat transfer model of the parabolic trough
either side of the receiver were proposed. Reddy et al. [38e43]
solar receiver is based on the energy balance between the incoming
proposed various porous inclusions inserted in one side of the in-
solar radiation (i.e., the solar energy input on the collector opening
ner absorber to enhance heat transfer in the receiver. Grena [44]
minus the optical losses), the heat transfer fluid (HTF), the absorber
presented a new receiver tube half-covered with an IR-reflecting
tube, the glass cover and the surroundings for both the two linear
layer in the non-irradiated half of the glass cover to reduce the
halves and the two inactive ends. Figs. 2 and 3 show the heat
thermal emission. As a preliminary study, this was evaluated by
transfer in a cross-section of the receiver and the thermal resistance
means of optical and thermal simulations with a simplified two
model used in the heat transfer analysis. The incoming solar radi-
halves model derived from that for a linear Fresnel solar collector
ations at the two linear halves of the receiver (as side 1 and side 2
[45], while the reliability and accuracy for the parabolic trough
shown in Fig. 2) and the two inactive ends are nonuniform. Side 2
solar collector need to be further validated and investigated. Al-
toward the reflector receives much higher solar flux than the side 1
Ansary and Zeitoun [46] numerically investigated the conduction
because of the radiation reflected by the parabolic mirror and the
and convection heat losses from a proposed receiver of a half-
two inactive ends receive no concentrated solar radiations at all.
insulated air-filled annulus with a heat-resistant thermal insu-
The incident solar radiation at the side 1 is directly absorbed by the
lation material. Cheng et al. [47] presented three-dimensional nu-
upper half absorber surface and the upper half glass cover which
merical studies on the turbulent flow and coupled heat transfer
enhancement in a novel unilateral multi-longitudinal vortexes
enhanced parabolic trough solar receiver, where longitudinal vor- Getters Annular space Glass cover with AR Evacuation nozzle
tex generators are only located on the side of the absorber tube
with the concentrated solar radiation. Considering the complex
modeling process and the numerical computing cost, mentioned
above, a quick analytical model also needs to be further developed Side 1
HTF
for these special cases.
Thus, a detailed one dimensional nonuniform thermal model is Side 2

proposed and used in this paper, to quickly and accurately estimate

the performance characteristics of the parabolic trough solar re-
ceivers for both the nonuniform conditions occurring in the para- Glass-to-metal joints Absorber Selective coating Bellows
bolic trough solar collector/receiver system (i.e., the nonuniform
solar flux and the corresponding nonuniform temperature Fig. 1. Schematic of the physical model of a PTC tubular receiver.
 Z.-D. Cheng et al. / Renewable Energy 74 (2015) 139e147 141

qa1, sol thermal exchange between the two halves is due to the HTF [45].
Sky - s q go
gs, r The inactive ends of the receiver are modeled by substituting the
gi glass cover with a thin-wall steel shell with corresponding prop-
Air - a qag, r erties and setting all the absorbed solar energy sources to zero [45],
ao qg1, sol while the heat transfer models are almost the same as the two
qga, c
ai halves.
qag, c The energy balance equations for both the two linear halves and
Side 1 f the two inactive ends are determined by considering that the en-
qaf, c Side 2 ergy is conserved at each surface of the receiver cross-section, as
the surfaces named ai, ao, gi and go shown in Fig. 2. They can be
listed as follows:

qg2, sol qa, cond

qab, end qaf;c ¼ qa;cond (1)

qa2, sol qg, cond

qak;sol  qa;cond
¼ qag;c þ qag;r þ qab;end (2)

Fig. 2. Heat transfer model in a cross-section of a parabolic trough solar receiver.

qag ¼ qag;c þ qag;r ¼ qg;cond þ s$qgk;sol ¼ qgi (3)

can be set as a uniform surface heat source (qa1,sol) and a uniform

volumetric heat source (qg1,sol) respectively in the thermal model qgo ¼ qg;cond þ ð1 þ sÞ$qgk;sol ¼ qga;c þ qgs;r (4)
later, while the incoming concentrated solar energy reaching the
side 2 is absorbed as qa2,sol and qg2,sol. Most of the radiant energy where the subscript k in qak,sol or qgk,sol is equal to 1 for the side 1 or
absorbed by either side of the absorber tube is conducted through 2 for the side 2 of the receiver shown in Fig. 2. The coefficient s is
the steel absorber tube (qa,cond) and eventually transformed into used to calculate the part of the solar energy through the inner
thermal energy of the HTF by convection (qaf,c). surface of the glass cover (qgi) since the solar absorption in the glass
The remaining energy is transferred back to the glass cover by cover wall is a volumetric phenomenon. It can be derived from a
thermal radiation (qag,r) and convection (qag,c) and lost through the heat conduction differential equation in a thin-wall cylindrical shell
support brackets by conduction (qab,end). The heat loss from the with an internal heat source, given as [45]:
absorber tube (qag ¼ qag,c þ qag,r) reaching the glass cover passes !
through the glass wall by conduction (qg,cond) and along with the R2gi 1
radiant energy absorbed by either half of the glass cover (qgi at the s¼     (5)
R2go  R2gi 2 ln Rgo Rgi
inner surface or qgo at the outer surface) is lost to the ambient air by
convection (qga,c) and to the sky by thermal radiation (qgs,r). The All terms in the energy balance equations can be classified into
convection heat transfer mechanism in the annular space depends the solar energy absorption of the solid tubes (i.e., the absorber and
on the annulus pressure of the residual gases, which may either be glass tubes), useful heat gain or net thermal loss of the parabolic
molecular conduction (free-molecular convection) or natural con- trough solar collector/receiver (i.e., the useful heat gain of the HTF
vection [9,10]. The convection heat transfer between the glass cover under the on-sun condition, or the net thermal loss from the HTF
and the atmosphere is also either forced or natural, depending on under the off-sun condition or at the inactive ends), the conduction
whether there is wind around the glass cover. Note that only the heat transfer in the solid tubes, the heat exchange in the annulus
solar energy absorbed by either side of the absorber tube and the and the thermal loss of the entire receiver. These heat transfer in-
glass tube are indicated by subscripts 1 and 2, as only these terms teractions are analyzed in detail below. These depend on the
are calculated by different expressions for the two halves. The heat parabolic trough solar collector/receiver (PTC) optical parameters
exchange between the two halves is omitted and this simplification (the PTC geometric parameters, the solar irradiation and the optical
is not expected to introduce severe inaccuracies for small solar loss), material properties, operating parameters (the temperature
collectors usually used. This is because more accurate simulations and the mass flow of the HTF, and the annulus pressure of the re-
show that the heat conduction and thermal radiation between the sidual gases) and ambient conditions (the external atmospheric
two halves are negligible or very limited, while almost all the temperature, the sky temperature and the wind condition).

Fig. 3. Thermal resistance network in a cross-section of a parabolic trough solar receiver.

142 Z.-D. Cheng et al. / Renewable Energy 74 (2015) 139e147

2.1. Solar energy absorption by the solid tubes 8

> 4:36; Ref  2300;
>
>  
<
The solar energy absorbed in the glass cover or at the absorber
ðf =8Þ Ref  1000 Prf  0:11
Nuaf ;c ¼ Prf
tube is mainly determined by the solar irradiation, solar angle, >
>   ; Ref > 2300:
>
: 1 þ 12:7ðf =8Þ0:5 Pr 2=3  1 Prw
optical properties and imperfections of the trough mirrors and
f
receiver components [10,11]. For different solar angles during the
day or year, the factor named incident angle modifier (Kq) can be (12)
used to account for all optical and geometric losses due to a
with the friction factor f calculated by the Filonenko equation [52]:
nonzero incident angle [12], while the imperfections in either the
reflector (or mirror) or the shape of the concentrator are accounted h   i2
by the intercept factor (gif) which is the ratio of the energy inter- f ¼ 1:82 ln Ref  1:64 (13)
cepted by the receiver to the energy reflected by the focusing devise
[11]. Thus, the energy absorbed by the absorber tube and the glass The HTF is synthetic oil (e.g., Syltherm 800 and Therminol VP-1
cover at side 1 without concentrated solar radiation is given by usually used in the PTC system, both have temperature-dependent
[40e42]: properties), which along with other HTF types can be found in
Refs. [10,29,32,53]. Thus, here kf, Ref and Prf are evaluated according
 
qg1;sol ¼ Dao Ig ag (6) to the mean HTF temperature while Prw is calculated according to
the inner absorber wall temperature. The correction term (Prf/
  Prw)0.11 is used to adjust the fluid property variations between the
qa1;sol ¼ Dao Ig tg aa (7)
inner absorber wall temperature and the bulk fluid temperature.
And the energy absorbed at side 2 toward the reflector with
concentrated solar radiation is given by: 2.3. Conduction heat transfer in the solid tubes
 
qg2;sol ¼ gif rr Kq ½ða  Dao ÞIb ag (8) Conduction heat transfer through the steel absorber wall and
  the glass cover wall are determined by the Fourier's law of con-
qa2;sol ¼ gif rr Kq ½ða  Dao ÞIb tg aa (9) duction through a thin-wall cylindrical shell respectively, can be
given by [49,50]:
where Ib and Ig are the beam radiation and the global radiation. rr, h  . i
ag and tg and aa are the reflectivity of the reflector, the absorptance qa;cond ¼ Sg $p ka Ta lnðDao =Dai Þ ðTao  Tai Þ (14)
and the transmittance of the glass cover and the absorptance of the
absorber tube, respectively. All the optical properties are assumed h  .   i 
to be independent of temperature and along with the incident qg;cond ¼ p kg Tg ln Dgo Dgi Tgi  Tgo (15)
angle modifier and the intercept factor are obtained from a com-
bination of sources in Refs. [9e12]. Under the on-sun condition, the temperature of the outer
absorber surface (Tao) is larger than that of the inner surface (Tai),
but it is reversed under the off-sun condition or at the inactive ends.
2.2. Useful heat gain or net thermal loss of the PTC
Thus Sg is also used to indicate the heat flux direction for the heat
transfer from higher temperature to lower temperature as
The useful heat gain or net thermal loss of the PTC is obtained or
described above. The thermal conductivity in these terms depends
caused by the convective heat transfer between the inner absorber
on the solid tube material type and the temperature. It changes
wall and the HTF, which can be calculated by the Newton's law of
slightly for Pyrex glass during the temperature range of PTC [54],
cooling as follows:
thus the thermal conductivity of the glass cover kg ðTg Þ is assumed
  as a constant of 1.04 W m1 K1. But the thermal conductivity of the
qaf ;c ¼ Sg $pRai haf;c Tai  Tf (10) stainless steel absorber ka ðTa Þ changes approximately as a linear
function of the temperature. The correlations for different material
where Sg is a sign to indicate the heat flux direction for the heat types (e.g., 304L, 316L, 321H, copper, etc) can be determined by
transfer from higher temperature to lower temperature. It is linearly fitting data from metal handbooks [12]. Here the correla-
equal to 1 for HTF cooling (under the on-sun condition) or 1 for tion from Ref. [45] for steel tubes is adopted:
HTF heating (under the off-sun condition or at the inactive ends)
 
in all subsequent equations, and the convection heat transfer ka Ta ¼ 14:106 þ 0:0138Ta ; with Ta ¼ ðTai þ Tao Þ=2 (16)
coefficient haf,c, based on the hydraulic diameter of Dh, is given
by: In the simulation of inactive ends, the value of the thermal
. conductivity of the glass cover must be substituted with the ther-
haf;c ¼ Nuaf ;c kf Dh (11) mal conductivity of the substitutive thin-wall steel shell, with the
same expression as ka ðTa Þ but based on the average temperature of
where Dh is usually equal to the inner diameter of the absorber tube the steel shell [45].
(Dai), but if there is a solid plug inserted in the tube center as used on
the AZTRAK rotating test platform at Sandia National Laboratories 2.4. Heat transfer in the annular space
(SNL), Dh is then calculated by Dai  Dplug [48]. The Nusselt number
Nuaf,c depends on the type of flow through the absorber tube which In the annular space between the absorber tube and the glass
is determined by the Reynolds number Ref (also based on the hy- cover, heat transfer occurs by thermal radiation and convection
draulic diameter of Dh). For laminar flow with Ref  2300, the heat transfer. As was mentioned before, the convection heat
Nusselt number is a constant equal to 4.36 under constant heat flux. transfer mechanism in the annular space depends on the presence
For full developed transitional and turbulent flow with Ref > 2300, of residual gases in this space or the annulus pressure of the re-
the correlation developed by Gnielinski [49e51] is used. Therefore: sidual gases. It may either be molecular conduction (or free-
 Z.-D. Cheng et al. / Renewable Energy 74 (2015) 139e147 143

molecular convection) at low pressures (0.013 Pa) or natural with

convection at higher pressures (>0.013 Pa) [55]. The threshold level
of vacuum is about 0.0001 Torr (0.013 Pa) which is determined by 2  C0 9ggas  5
B0 ¼ $   (22)
the Knudsen gas conduction range to make heat losses by con- C0 2 ggas þ 1
duction and convection through the residual gases insignificant. As
the vacuum may be degraded over time, the natural convection also  
needs to be modeled to evaluate the PTC performance under these Tag þ 273:15
lgas ¼ 2:331  1020 (23)
conditions. The thermal radiation heat transfer occurs due to the Pgas d2gas
temperature difference between the outer absorber surface and the
inner glass cover surface and it is simplified by assuming that the where kgas, ggas, Pgas, dgas, lgas and C0 are the thermal conductivity of
glass cover is opaque to infrared radiation with gray and diffuse the annulus gas at standard temperature and pressure
surfaces [10]. (W m1 K1), the ratio of specific heats of the annulus gas, the
annulus gas pressure (mmHg), the molecular diameter of annulus
2.4.1. Radiation heat transfer gas (cm), the mean-free-path between collisions of a molecule (cm)
The thermal radiation heat transfer between the outer absorber and the accommodation coefficient respectively, which can be
wall and the inner cover surface is estimated with the following referred to vacuum in annulus condition in Ref. [12] and evaluated
equation [12,50]: at the average temperature Tag ¼ ðTao þ Tgi Þ=2.
h  4 i When the vacuum is lost over time with ambient air (pressure
qag;r ¼ pRao s3 ag ðTao þ 273:15Þ4  Tgi þ 273:15 (17) >0.013 Pa), the heat transfer is natural convection dominated. The
effective heat transfer coefficient can be substituted by [10,58]:
with  0:25
2:425kag Raag;Dao Prag
" !#1 hag;c ¼ Fag  0:25 (24)
1 Dao 1  3 gi 2pRao 0:861 þ Prag
3 ag ¼ þ (18)
3 ao Dgi 3 gi
with
where 3 ao and 3 gi are the emissivity of the selective absorbing h   0:6 i1:25
coating and the inner glass surface respectively. The glass emis- Fag ¼ 1 þ Dao Dgi (25)
sivity has a constant value of 3 gi ¼ 0.90 [11,54], while the coating
emissivity is usually a nearly linear function of wall temperature where kag, Prag and Raag;Dao are the thermal conductivity of annulus
and coincides with the selective coating types (e.g., black chrome, gas, the Prandtl number and the Rayleigh number for gas properties
cermet) [10]. Taking Luz cermet used on the AZTRAK rotating test evaluated at Tag ¼ ðTao þ Tgi Þ=2, respectively.
platform at SNL as an example, it can be given by [10,12]:

3 ao ¼ 0:000327ðTao þ 273:15Þ  0:065971 (19) 2.5. Thermal loss of the entire receiver
It should be noted that several assumptions were made in the
above equations, i.e., opaque glass cover to infrared radiation, The thermal losses and the thermal efficiency are the most
long concentric isothermal cylinders with gray and diffuse sur- important performance characteristics of the PTC system. The
faces and non-participating gas for thermal radiation in the thermal efficiency of the PTC system is defined as the ratio of the
annulus. All units of the temperatures are in degrees Celsius to be heat transferred to the HTF to the solar radiation absorbed by the
consistent with other equations. For the inactive ends, the value absorber tube calculated previously. The thermal loss of the entire
of the thermal emissivity of the glass must be substituted with parabolic trough solar receiver mainly includes the thermal radia-
the emissivity of the thin-wall steel shell as 0.0434 þ 0.1062 ln Tgi tion heat loss to the sky, the convection heat loss to the atmosphere
[45]. and the conduction heat loss to support brackets, which are
examined separately below.
2.4.2. Convection heat transfer
The convection heat transfer in the annulus can be calculated by
2.5.1. Thermal radiation heat loss to the sky
a unified formula with different effective convection heat transfer
Compared to the large sky, the whole glass cover can be
coefficients (hag,c) for the molecular conduction (or free-molecular
assumed to be a small convex gray object in a large blackbody. Since
convection) and the natural convection [9], as follows:
it is common practice to model the sky as a blackbody and to use an
  effective sky temperature (Ts) to compensate for the difference
qag;c ¼ pRao hag;c Tao  Tgi (20)
[10,12], especially for non-clear conditions. Thus the net radiation
where the effective convection heat transfer coefficient can be transfer, caused by the temperature difference between the outer
examined as presented below, based on the theory of conduction glass cover surface and the sky, can be given by [48e50]:
heat transfer and kinetic theory [9]. h 4 i
When the annulus is under vacuum (pressure 0.013 Pa), the qgs;r ¼ pRgo s3 go Tgo þ 273:15  ðTs þ 273:15Þ4 (26)
intermolecular collisions seldom occur, thus the heat transfer is
free-molecular convection dominated. The effective heat transfer where 3 go is the emissivity of the outer glass cover surface with a
coefficient for the annular space is given by [9e12,56,57]: constant value of 0.90 [11,54]. Ts depends on the meteorological
conditions and it can be approximated by a simple relation from
k Swinbank [59] as Ts ¼ 0.0553(Tgo þ 273.15)1.5  273.15 in the
hag;c ¼    gas    (21)
ðDao =2Þln Dgi Dao þ B0 lgas Dao Dgi þ 1 absence of meteorological data [11]. For inactive ends, the value of
144 Z.-D. Cheng et al. / Renewable Energy 74 (2015) 139e147

the thermal emissivity of the glass must be substituted with the the wind speed and can be calculated by the same expressions for
emissivity of the thin-wall steel shell as 0.0434 þ 0.1062 ln Tgo [45]. no wind and wind cases mentioned above [10].

3. Model validation and discussion

2.5.2. Convection heat loss to the atmosphere
The convection heat transfer from the glass cover to the atmo-
To solve the energy balance of the receiver, the expressions
sphere may be either natural or forced, depending on whether
(5)e(31) for the corresponding conditions are substituted in
there is wind around the glass cover. Both can be determined from
equations (1)e(4), forming a nonlinear systems with the interme-
Newton's law of cooling [10,12]:
diate temperatures (i.e., Tai, Tao, Tgi and Tgo) as unknown. The
  nonlinear algebraic equations were written in Fortran 6.5 and
qga;c ¼ pRgo hga;c Tgo  Tair (27)
solved iteratively within a few seconds if the HTF temperature and
 temperature-dependent properties, the mass flow rate, the energy
hga;c ¼ Nuga;Dgo kair Dgo (28)
absorbed at the absorber surface and in the glass cover on both
sides are specified under given operating conditions. From the
where the convection heat transfer coefficient hga,c or the Nusselt temperatures obtained, the fluid flow and heat transfer perfor-
number Nuga;Dgo also depends on whether the convection heat mance characteristics can be calculated accordingly.
transfer is natural (no wind) or forced (with wind). For heat con- In order to validate the proposed nonuniform thermal model, it
vection with no wind, the correlation developed by Churchill and was compared with known experimental performance data of the
Chi [11,60] is recommended for the long horizontal cylinder (glass LS-2 collector tested at the AZTRAK rotating platform at the SNL [9],
cover) under natural convection, as follows [11,12]: with the same collector geometrical configuration, operating con-
8 92 ditions, material and optical properties summarized in Table 1 or in
< 1=6
0:387Raga;Dgo = Refs. [9] and [10]. The Sandia test was performed for both on-sun
Nuga;c ¼ 0:60 þ h   i
9=16 8=27 ; (29) and off-sun condition and different scenarios for the annulus
:
1 þ 0:559 Prga (vacuum and air), using Syltherm 800 as HTF. A 2-inch diameter
solid plug used as the flow restriction device was centered inside
where Prga and RagaDgo are the Prandtl number and Rayleigh the absorber tube, due to limitations in the experimental set up. In
number for the ambient air evaluated at Tga ¼ ðTgo þ Tair Þ=2, addition, to corroborate the assumption that heat exchanges be-
respectively. tween the two halves are negligible or very limited, the nonuniform
For heat convection with wind, the Nusselt number is estimated thermal model are also compared with a corresponding uniform
with Zhukauskas' correlation for a horizontal cylinder (glass cover) thermal model of uniform solar energy around the whole PTC
in cross flow [11,12,60]: cross-section. The uniform thermal model is performed by setting
the values of heat sources at the absorber surface and in the glass
n0   0:25
Nuga;c ¼ Cmn Rem 0
air;D
Prair Prair Prgo (30) cover as the averages of that in the two halves of the nonuniform
go
thermal model.
where Reair;Dgo and Prair are evaluated according to the air tem- Fig. 4 shows a comparison of the known experimental efficiency
perature while Prgo is calculated according to the outer glass wall and heat loss with the values determined from the nonuniform or
temperature. The constants Cmn, m0 and n0 suggested for this uniform thermal model developed, when vacuum exists in the
equation are given by [10,48,60]. annulus. Similar results for air in the annulus are presented in Fig. 5.
From these figures, it can be seen that the agreement between the
experimental results and those obtained by the nonuniform or
2.5.3. Conduction heat loss to support brackets uniform thermal model is very acceptable for all cases. The model is
The solar receiver is placed at the collector focal line by support validated for the whole operating temperature range of PTC sys-
brackets that run from the collector structure to the absorber tube. tems and as seen the deviation increases slightly with increasing
There is a support bracket at each end of every receiver element, operating temperature. One of the possible reasons is the depen-
consisting of two rectangular connection tabs and two square dence of the optical properties on the temperature which was
metallic tubes [10,11]. The two connection tabs form the smallest ignored [12]. Another is from the assumption that there is no
cross-section area near the base of the support bracket where thermal interaction between the neighboring surfaces, which
conduction heat transfer occurs from the absorber tube, and the eventually underestimates the radiation losses at high absorber
square tubes are the parts of the support brackets that are exposed temperatures slightly. A comprehensive radiative analysis with the
to convection heat transfer to the environment. To calculate the temperature-dependent optical properties can be further imple-
bracket losses, the support bracket is treated as an infinite fin with mented for radiation heat transfer in the annulus and from the glass
base temperature (Tbase ¼ Tgo  10) at the absorber end where the cover to the sky or surroundings [11].
bracket is attached [10,11]. The bracket heat loss for each half of the In addition, the difference between the collector efficiency and
solar receiver is estimated with the following equation [48]. heat loss calculated by the nonuniform and uniform thermal model
 0:5
qab;end ¼ hba;c Pb kb Ab;cs ðTbase  Tair Þ=ð2Lr Þ (31) Table 1
Main parameters, conditions and optical properties of the parabolic trough solar
where Lr is the length of the solar receiver, Pb is the perimeter collector [9e12].
around the two square tubes that run from the absorber attach- PTC parameters and conditions Optical
ment bracket to the collector structure, Ab,cs is the cross-sectional properties
area of the two connection tabs connecting the tubes to the a 5m da 0.003 m rr 0.930
absorber bracket. kb is the conduction coefficient for the support L 7.8 m Dgi 0.110 m tg 0.950
bracket with a constant equal to 48.0 W m1 K1 (plain carbon steel fr 1.84 m Dgo 0.115 m ag 0.023
at 600 K). hba,c is the average convection coefficient of the bracket Dai 0.064 m dg 0.0025 m 3g 0.900
Dao 0.070 m HTF Syltherm 800 aa 0.960
and assumed as uniform over the length of the fin. It depends on
 Z.-D. Cheng et al. / Renewable Energy 74 (2015) 139e147 145

80 results agree well with the experimental data, proving that the
nonuniform or uniform thermal model used in the present study is
75 feasible and reliable. For applications, the uniform thermal model
can be used to quickly compute the integral heat transfer perfor-
mance of the whole PTC system while the nonuniform thermal
70
model can be used to analyze the local nonuniform solar radiation
and heat transfer performance characteristics and nonuniform heat
Efficiency / %

65 transfer enhancements. This enables the study of different design

parameters, material properties, operating conditions, fluid flow
60 and heat transfer performance for the two nonuniform regions or
Air condition (Dudley et al.) the whole receiver.
55 Experimental data (Dudley et al.) As compared to the time-consuming three-dimensional CFD
Uniform thermal model PTC models, this quick nonuniform thermal model provides a
50 Nonuniform thermal model research idea/strategy/tool or certain reference significance for
Vacuum conditon (Dudley et al.) evaluating the characteristics or determining the limitations of
45 nonuniform conditions in PTCs. It could also be effectively used
0 50 100 150 200 250 300 350 400 with an intelligent optimization (e.g., GA or PSO) to evaluate and
Average Temperature Above Ambient /K optimize the characteristics and performance of PTCs under series
of nonuniform conditions in detail within a few hours compared to
Fig. 4. Comparison of collector efficiency calculated from the proposed model with the complex modeling process of the intelligent optimization with
experimental data of cermet coating under on-sun case [9]. three-dimensional nonuniform computational fluid dynamics
models which need great computations time of several weeks or
is very small under on-sun condition whereas there is almost no months. Before that, a comprehensive radiative analysis with the
difference under off-sun condition. It corroborates the assumption temperature-dependent optical properties for radiation heat
that heat exchanges between the two halves can be ignored for transfer in the annulus and from the glass cover to the sky or sur-
small solar collectors available under the conditions studied [45]. rounding, and the validation for other conditions or corresponding
Other conditions or corresponding correlations could be further correlations need to be studied further.
studied later. For applications, the uniform thermal model can be
used to quickly compute the integral heat transfer performance of Acknowledgments
the whole PTC system while the nonuniform thermal model can be
used to analyze the nonuniform solar radiation and heat transfer The study is supported by the National Natural Science Foun-
performance characteristics and nonuniform heat transfer en- dation of China (No. 51306149), the China Postdoctoral Science
hancements, considering different design parameters, material Foundation (No. 2014M550489), and the Research Project of Chi-
properties, operating conditions, fluid flow and heat transfer per- nese Ministry of Education (Nos. 20120201130006, 113055A).
formance for the corresponding regions or the whole receiver.
Nomenclature
4. Conclusions
A area (m2)
This study presented a detailed one dimensional nonuniform or a aperture width (m)
uniform thermal model of a parabolic trough solar collector system. D diameter (m)
This work was carried out by dividing the entire receiver into two Dh hydraulic diameter (m)
linear halves and two inactive ends with different solar radiation Dplug diameter of the solid plug (m)
and heat transfer modes. The model is then validated with known dg outer diameter of glass cover (m)
performance of an existing PTC. It is revealed that the numerical f friction factor
fr focal length (m)
180 h convective heat transfer coefficient (W m2 K1)
Lr receiver length (m)
160 Air condition (Dudley et al.)
Experimental data (Dudley et al.) Nu Nusselt number
140 Uniform thermal model k thermal conductivity (W m1 K1)
Nonuniform thermal model Kq incident angle modifier
-2

120
Thermal Loss / Wm

Vacuum conditon (Dudley et al.) Ib the beam radiation (W m2)

100 Ig the global radiation (W m2)
Pr Prandtl number
80 R radius (m)
Ra Rayleigh number
60 Re Reynolds number of HTF
40 q heat flux per unit length (W m1)
qv volume flow rate (L min1)
20 T temperature (
C)
Vw wind speed (m s1)
0
0 50 100 150 200 250 300 350 400
Greek symbols
Average Temperature Above Ambient /K
a absorptivity of the absorber
Fig. 5. Comparison of thermal losses calculated from the proposed model with ag absorptivity of the glass cover
experimental data of cermet coating under off-sun case [9]. d thickness (m)
146 Z.-D. Cheng et al. / Renewable Energy 74 (2015) 139e147

3 emissivity [14] Jeter SM. Calculation of the concentrated flux density distribution in parabolic
trough collectors by a semifinite formulation. Sol Energy 1986;37:335e45.
gif intercept factor
[15] Eck M, Steinmann WD. Modeling and design of direct solar steam generating
hc collector efficiency (%) collector fields. J Sol Energy Eng 2005;127:371e80.
rr reflectivity of the reflector [16] Lüpfert E. Parabolic trough receiver testing e thermal loss tests. Report of
s StefaneBoltzmann constant (5.67E10-8 W m2 K4) DLR; 2006.
[17] Lüpfert E. Parabolic trough optical performance analysis techniques. Report of
tg transmissivity of the glass cover DLR; 2006.
€nder M, Lüpfert E, Pistor P. Infrared temperature measurements on solar
[18] Pfa
trough absorber tubes. Sol Energy 2007;81:629e35.
Subscripts [19] Cheng ZD, He YL, Xiao J, Tao YB, Xu RJ. Three-dimensional numerical study of
a absorber or atmosphere heat transfer characteristics in the receiver tube of parabolic trough solar
collector. Int Commun Heat Mass Transf 2010;37(7):782e7.
ab absorber-bracket [20] Lu JF, Ding J, Yang JP. Heat transfer performance of the receiver pipe under
ai inner wall of the absorber unilateral concentrated solar radiation. Sol Energy 2010;84:1879e87.
ao outer wall of the absorber [21] He YL, Xiao J, Cheng ZD, Tao YB. A MCRT and FVM coupled simulation method
for energy conversion process in parabolic trough solar collector. Renew
af absorber-HTF Energy 2011;36:976e85.
ag absorber-glass [22] Cheng ZD, He YL, Cui FQ. A new modelling method and unified code with
a1 side 1 of the absorber MCRT for concentrating solar collectors and its applications. Appl Energy
2013;101:686e98.
a2 side 2 of the absorber [23] Cheng ZD, He YL, Xiao J, Tao YB, Xu RJ. Numerical study of heat transfer
b bracket characteristics in the solar receiver tube combined with MCRT method. In:
ba bracket-air 2009 Inaugural USeEUeChina thermophysics conference e renewable en-
ergy; May 28e30, 2009. Beijing, China.
c convection [24] Eck M, Feldhoff JF, Uhlig R. Thermal modelling and simulation of parabolic
cond conduction trough receiver tubes. In: Proceedings of the ASME 4th international confer-
cs cross section ence on energy sustainability 2010, Phoenix, Arizona, USA; May 17e22, 2010.
p. 1e8.
f fluid or HTF
[25] Wang FQ, Shuai Y, Yuan Y, Yang G, Tan HP. Thermal stress analysis of eccentric
g glass cover tube receiver using concentrated solar radiation. Sol Energy 2010;84:1809e15.
ga glass-air [26] Aldali Y, Muneer T, Henderson D. Solar absorber tube analysis: thermal
gas residual gas in annulus simulation using CFD. Int J Low-Carbon Technol 2011:1e6.
[27] Roesle M, Coskun V, Steinfeld A. Numerical analysis of heat loss from a
gi inner wall of the glass cover parabolic trough absorber tube with active vacuum system. J Sol Energy Eng
go outer wall of the glass cover 2011;133:031015-1e031015-5.
[28] Mun ~ oz J, Ab
anades A. Analysis of internal helically finned tubes for parabolic
gs glass-sky
trough design by CFD tools. Appl Energy 2011;88:4139e49.
g1 side 1 of the glass cover [29] Cheng ZD, He YL, Cui FQ, Xu RJ, Tao YB. Numerical simulation of a parabolic
g2 side 2 of the glass cover trough solar collector with nonuniform solar flux conditions by coupling FVM
r radiation and MCRT method. Sol Energy 2012;86:1770e84.
[30] Wang FQ, Shuai Y, Yuan Y, Liu B. Effects of material selection on the thermal
s sky or steel stresses of tube receiver under concentrated solar irradiation. Mater Des
sol solar 2012;33:284e91.
st smooth tube [31] Wang P, Liu DY, Xu C. Numerical study of heat transfer enhancement in the
receiver tube of direct steam generation with parabolic trough by inserting
w wall metal foams. Appl Energy 2013;102:449e60.
[32] Roldan MI, Valenzuela L, Zarza E. Thermal analysis of solar receiver pipes with
superheated steam. Appl Energy 2013;103:73e84.
References [33] Hachicha AA, Rodríguez I, Capdevila R, Oliva A. Heat transfer analysis and
numerical simulation of a parabolic trough solar collector. Appl Energy
[1] Price H, Lüpfert E, Kearney D, Zarza E, Cohen G, Gee R, et al. Advances in 2013;111:581e92.
parabolic trough solar power technology. J Sol Energy Eng 2002;124:109e25. [34] Mwesigye A, Bello-Ochende T, Meyer JP. Numerical investigation of entropy
[2] Kalogirou SA. Solar thermal collectors and applications. Prog Energy Combust generation in a parabolic trough receiver at different concentration ratios.
Sci 2004;30(3):231e95. Energy 2013;53:114e27.
[3] Grena R. Optical simulation of a parabolic solar trough collector. Int J Sustain [35] Wang K, He YL, Cheng ZD. A design method and numerical study for a new
Energy 2010;29(1):19e36. type parabolic trough solar collector with uniform solar flux distribution. Sci
[4] Wirz M, Roesle M, Steinfeld A. Three-dimensional optical and thermal nu- China Technol Sci 2014;57(3):531e40.
merical model of solar tubular receivers in parabolic trough concentrators. [36] Wu ZY, Li SD, Yuan GF, Lei DQ, Wang ZF. Three-dimensional numerical study
J Sol Energy Eng 2012;134:041012-1e21-9. of heat transfer characteristics of parabolic trough receiver. Appl Energy
[5] Fernandez-García A, Zarza E, Valenzuela L, Pe rez M. Parabolic-trough solar 2014;113:902e11.
collectors and their applications. Renew Sustain Energy Rev 2010;14: [37] Cheng ZD, He YL, Wang K, Du BC, Cui FQ. A detailed parameter study on the
1695e721. comprehensive characteristics and performance of a parabolic trough solar
[6] Yang B, Zhao J, Yao WB, Zhu Q, Qu H. Feasibility and potential of parabolic collector system. Appl Therm Eng 2014;63:278e89.
trough solar thermal power plants in Tibet of China. In: Power and Energy [38] Patil G, Reddy KS. Thermal analysis of energy efficient receiver for solar
Engineering Conference (APPEEC) 2010 Asia-Pacific, Chengdu, China; March parabolic trough collector system. In: Global conference on renewable energy
28e31, 2010. p. 1e4. approaches for desert regions, Jordan; 2006. p. 361e9.
[7] Cheng ZD, He YL, Cui FQ, Du BC, Zheng ZJ, Xu Y. Comparative and sensitive [39] Satyanarayana GV, Ravi Kumar K, Reddy KS. Numerical study of porous
analysis for parabolic trough solar collectors with a detailed Monte Carlo ray- enhanced receiver for solar parabolic trough collector. In: 3rd International
tracing optical model. Appl Energy 2014;115:559e72. conference on solar radiation and day lighting, New Delhi, vol. 2; 2007.
[8] Patnode AM. Simulation and performance evaluation of parabolic trough solar p. 46e55.
power plants. Master thesis. University of Wisconsin-Madison: College of [40] Reddy KS, Satyanarayana GV. Numerical study of porous finned receiver for
Engineering; 2006. solar parabolic trough concentrator. Eng Appl Comput Fluid Mech 2008;2(2):
[9] Dudley V, Kolb G, Sloan M, Kearney D. SEGS LS2 solar collector e test results. 172e84.
Report of Sandia National Laboratories, SANDIA 94-1884, USA; 1994. [41] Reddy KS, Ravi Kumar K, Satyanarayana GV. Numerical investigation of energy
[10] Forristall R. Heat transfer analysis and modeling of a parabolic trough solar efficient receiver for solar parabolic trough concentrator. J Heat Transf Eng
receiver implemented in engineering equation solver. Technical Report, NREL/ 2008;29(11):961e72.
TP-550-34169; 2003. [42] Ravi Kumar K, Reddy KS. Thermal analysis of solar parabolic trough with
[11] Padilla RV, Demirkaya G, Goswami DY, Stefanakos E, Rahman MM. Heat porous disc receiver. Appl Energy 2009;86:1804e12.
transfer analysis of parabolic trough solar receiver. Appl Energy 2011;88: [43] Ravi Kumar K, Reddy KS. Effect of porous disc receiver configurations on
5097e110. performance of solar parabolic trough concentrator. Heat Mass Transf
[12] Kalogirou SA. A detailed thermal model of a parabolic trough collector 2012;48:555e71.
receiver. Energy 2012;48(1):298e306. [44] Grena R. Efficiency gain of a solar trough collector due to an IR-reflective film
[13] Jeter SM. The distribution of concentrated solar radiation in paraboloidal on the non-irradiated part of the receiver. Int J Green Energy 2011;8(7):
collectors. J Sol Energy Eng 1986;108:219e25. 715e33.
 Z.-D. Cheng et al. / Renewable Energy 74 (2015) 139e147 147

[45] Grena R, Tarquini P. Solar linear Fresnel collector using molten nitrates as heat [53] Odeh SD, Morrison GL, Behnia M. Modelling of parabolic trough direct steam
transfer fluid. Energy 2011;36:1048e56. generation solar collectors. Sol Energy 1998;62(6):395e406.
[46] Al-Ansary H, Zeitoun O. Numerical study of conduction and convection heat [54] Touloukian YS, DeWitt DP, editors. Radiative properties, nonmetalic solids,
losses from a half-insulated air-filled annulus of the receiver of a parabolic thermophysical properties of matter, vol. 8. New York: Plenum Publishing; 1972.
trough collector. Sol Energy 2011;85:3036e45. [55] KJC Operating Company. Final report on HCE heat transfer analysis code.
[47] Cheng ZD, He YL, Cui FQ. Numerical study of heat transfer enhancement by SANDIA Contract No. AB-0227. Albuquerque, NM: Sandia National Labora-
unilateral longitudinal vortex generators inside parabolic trough solar re- tories; 1993.
ceivers. Int J Heat Mass Transf 2012;55:5631e41. [56] Dushman S. Scientific foundations of vacuum technique. 2nd ed. New York:
[48] Incropera F, DeWitt D. Fundamentals of heat and mass transfer. 3rd ed. New Wiley; 1962.
York: John Wiley and Sons; 1990. [57] Ratzel A, Hickox C, Gartling D. Techniques for reducing thermal conduction
[49] Cengel YA. Heat transfer and mass transfer: a practical approach. 3rd ed. and natural convection heat losses in annular receiver geometries. J Heat
McGraw Hill Book Company; 2006. Transf 1979;101(1):108e13.
[50] Tao WQ. Numerical heat Transfer. 2nd ed. Xi'an: Xi'an Jiaotong University [58] Bejan A. Convection heat transfer. 2nd ed. New York: John Wiley & Sons;
Press; 2001. 1995.
[51] Gnielnski V. New equations for heat and mass transfer in turbulent pipe and [59] Swinbank W. Long-wave radiation from clear skies. Quar J R Meteorol Soc
channel flow. Int Chem Eng 1976;16(2):359e63. 1963;89(381):339e48.
[52] Incropera FP, DeWitt DP. Introduction to heat transfer. 3rd ed. New York: John [60] Kakat S, Shah RK, Aung W. Handbook of single-phase convective heat transfer.
Wiley & Sons; 1996. New York: John Wiley & Sons; 1987.