Energy 193 (2020) 116775

Contents lists available at ScienceDirect

Energy
journal homepage: www.elsevier.com/locate/energy

Combined optics and heat transfer numerical model of a solar conical

receiver with built-in helical pipe
Yanping Zhang a, b, Hu Xiao a, Chongzhe Zou a, *, Quentin Falcoz b, c, Pierre Neveu b, c
a
School of Energy and Power Engineering, Huazhong University of Science and Technology, China
b
China-EU Institute for Clean and Renewable Energy, Huazhong University of Science and Technology, China
c
Processes Materials and Solar Energy Laboratory, PROMES-CNRS, Font-Romeu Odeillo, France

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

Article history: The performance of a solar conical receiver greatly influences the concentrated solar power (CSP) system.
Received 22 July 2019 It is worthy to devote efforts on performance optimization for a solar conical receiver with a built-in
Received in revised form helical pipe. A comprehensive investigation on the whole solar-heat conversion process is significant.
9 December 2019
Using a combined optics and heat transfer model, this study simulated the influence of the conical angle,
Accepted 14 December 2019
Available online 16 December 2019
loop number, and insulation thickness on cavity receiver performances. The combination consists of two
parts: an optical simulation using the optics software TracePro 7.3.4 and a heat-transfer simulation using
the CFD software ANSYS 17.0. Model validation is performed by comparing the simulation results to the
Keywords:
Solar conical receiver
published experiment data. Good agreement is achieved. The results indicate that the optimal value of
Geometric parameters conical angle is 5
, the overall efficiency in this case is 63.6%. From 4 to 15, both thermal efficiency and
Comprehensive model overall efficiency increase as the loop number increases; however, optical efficiency shows a different
Heat losses trend. The optimal value of insulation thickness ranging from 25 mm to 200 mm is 175 mm, achieving
Overall efficiency 70.4% overall efficiency. This study provides an effective method of performance optimization for a
Performance optimization conical cavity receiver.
© 2019 Elsevier Ltd. All rights reserved.

1. Introduction critical geometric parameters for achieving maximum overall effi-

ciency in solar-heat conversion.
Solar energy is one solution to the current energy crisis. One of A number of techniques have been proposed in the recent
the most promising applications of solar energy is concentrated literature on cavity receiver investigations, such as theoretical
solar power (CSP), which is widely used in power generation sys- calculations, CFD simulations, and experimental investigations.
tems. Solar-heat conversion plays an important role in CSP tech- These investigations have involved differently shaped receivers,
nology, which is highly dependent on achieving high-temperature such as hemisphere, cuboid, cylinder and other irregularly shaped
working fluid. Cavity receivers with built-in helical pipes have receivers. These studies have led to a better understanding of the
become popular in recent years due to their high efficiency and low influence mechanism of relevant parameters on heat losses. In the
cost. work of S. Paitoosurikarn and K. Lovegrove [1], this issue is
There are broad interests and literatures base for the perfor- addressed using CFD simulation, and the results revealed that the
mance optimization of cavity receivers. Some have simulated the calculated heat losses shows a nonlinear dependence on the incli-
optics and heat transfer processes separately. Yet it is rare to see nation angle. Reddy and Kumar et al. [2] proposed numerical
studies on conical cavity receivers using a combined optics and heat models of a hemisphere cavity for researching heat losses and
transfer model, especially for investigating the influence of the indicated that maximum convection loss occurs at a 0
incline. Ngo
conical angle, loop number, and insulation thickness on receiver et al. [3]investigated a plate fins cavity model with numerical
performance. Thus, it remains to be elucidated in more detail of the simulation, showing that natural convection heat loss and radiation
method, which is able to determine the optimum values of three heat loss decreased by 20% and 5%, respectively. Shirvan et al. [4]
numerically studied the effects of a porous media receiver on
convection heat loss and surface radiation heat loss, which pro-
* Corresponding author. posed that the total heat-transfer rate increases with an increase of
E-mail address: zcz@hust.edu.cn (C. Zou). the Rayleigh number, Darcy number, inclination angle and wall

https://doi.org/10.1016/j.energy.2019.116775
0360-5442/© 2019 Elsevier Ltd. All rights reserved.
2 Y. Zhang et al. / Energy 193 (2020) 116775

surface emissivity. Wu et al. [5] researched a cylindrical cavity the inner radius by 2 mm.
receiver in both of 3-D numerical simulation and experimental Although researchers have carried out investigations with
investigation, revealing that tilt angle and aperture position influ- comprehensive models, in the computational model of Cui et al. the
ence natural convection heat loss. Loni et al. [6] constructed a cavity geometric parameters of the cavity are invariant. While Qiu et al.
receiver model in square prismatic tubular, revealing that methanol varied the tube diameter and operation conditions, they neglected
and R11 provide the maximum and minimum thermal efficiencies other critical geometric parameters. In order to undertake an in-
within accounted range of turbine inlet temperature. Prezenak tegrated investigation, it is necessary to construct a comprehensive
et al. [7] constructed a flat dish receiver model with pipes wiggling simulation using a combined optics and heat transfer model. It is
across it. The CFD and MCRT simulation results indicated that a likewise necessary to identify a performance optimization method
distance of 76 cm and a fluid flow velocity of 0.6 m/s are optimal for conical cavity receivers.
value for maximizing the heat-transfer. Overall conversion efficiency is a significant indicator for the
More recently, there have been several attempts at geometrical system performance assessment. In this study, a comprehensive
influences on the performance of cavity receivers. Solutions in model is constructed with combined optics-heat transfer process,
literature have pursued a variety of methods, such as C.-A. Asseli- using basic knowledge of optics, heat transfer and fluid mechanics.
neau et al. [8] used a two-stage optimization process to maximize After solving the model and processing the results, this study ob-
the heat rate output of a four-parameter axisymmetric cavity tained the overall efficiency of this solar-heat system. With this
receiver, and the results showed that a significant range of geo- performance indicator, it is able to achieve the optimum parame-
metrics show similar efficiencies while varying wall flux. Tu et al. ters set. Three critical parameters, conical angle, loops number and
[9]investigated the influence of depth on the thermal performance insulation thickness, are investigated for studying the influence
of a steam receiver in relation to a three-dimensional (3D) nu- mechanism on the overall efficiency. Thus, the optimization
merical simulation, finding that 2 m is the optimum cavity depth method for conical cavity receiver is proposed.
for this irregularly shaped receiver. Prakash et al. [10] compared The contributions of this work are presented as follows, the heat
three different shapes (cubical, spherical and hemispherical ge- flux distribution inside the conical cavity is simulated with the
ometries) with each other in equal heat-transfer area by numerical assistance of the MCRT method, using optics software TracePro©
analysis on natural convection, the results observed that a hemi- 7.3.4. The non-uniform heat flux distribution is subsequently used
spherical open cavity has the greatest natural convection loss. Gil as the boundary condition to simulate the heat transfer process of
et al. [11] regarded aperture diameter to maximize thermal effi- the conical receiver. To create a combined optics and heat transfer
ciency of a cylindrical receiver by using a finite differences model, model, a novel procedure that enables a precise simulation and
as well as receiver absorptivity, which revealed that the optimal reduces the time cost is applied. To this end, the 3D flow fields of
aperture height depends on the minimum focal distance. Hussain the conical receiver are solved using ANSYS© 17.0, and the heat
et al. [12] performed an experiment to research the influence of the losses and the efficiencies are obtained from the calculation results.
aperture ratio (AR ¼ d/D) and aperture position (AP ¼ H/D) on heat Ultimately, this study reveals the influence mechanism of the
losses of a cylindrical cavity receiver, using a Fresnel lens focusing conical angle, loop number, and insulation thickness on perfor-
to heat a Stirling engine, exploring that the optimum values were mance and proposes a performance optimization method for
equal to 0.5 and 0.53, respectively. Using optical simulation, Daabo conical cavity receivers that will be useful in CSP technology
et al. [13] presented a comparison among a cubical cavity, spherical research and development.
cavity, semi-spherical cavity and a cylindrical cavity, no one special
shape has obvious advantage in optical efficiency. In the pre- 2. Conical receiver modeling
liminary stage of the present study, Zou et al. [14] optimized three
critical geometric parameters (aperture diameter, cavity inner The modeling of the conical cavity receiver used in this study in
diameter, and cavity length) by presenting a theoretical model of a this section consists in three steps: (1) structural design and
cylindrical cavity receiver with a built-in helical pipe. Li et al. [15] physical model; (2) optical simulation to obtain the heat flux dis-
proposed an experimental investigation of a conical receiver in a tribution using TracePro©7.3.4; and (3) the combination of the
double-reflecting CSP system, finding that the optimal conical optical-thermal model and CFD numerical investigation using
angle of 100
leads to minimum heat loss. In Li et al.‘s study, ANSYS© 17.0.
however, the effect of the conical angle on heat flux distribution
was calculated with an identical optical efficiency assumption, 2.1. Structural design and physical model
which is not sufficiently representative of actual conditions.
In the aforementioned studies, the optics and heat transfer The conical receiver, which absorbs reflected solar irradiation
process models were separately constructed. More precise re- and transfers the energy to the working fluid (air) via helical pipe, is
searches that involves the construction of a comprehensive simu- a key component of the CSP system. Fig. 1 (a) and (b) present
lation model (including both the optics process and heat transfer schematics of the conical cavity receiver configuration. The receiver
process) are necessary. Some simplified assumptions, such as that includes an insulation layer and consists of an aperture in front, an
of the uniform boundary condition, reduced the accuracy of the enclosed round bottom in the back, and a helical pipe inside. The
simulation results. A range of techniques have been proposed in the receiver is installed at the focal point of the parabolic dish in order
literature to improve on this, Cui et al. [16] combined the MCRT to ensure that the focused sunspot is inside the cavity [18]. To
(Monte-Carlo Ray Tracing) method and the SIMPLE method, protect against rain and snow, the outer cover of the cavity is made
developing a model coupling the inhomogeneous radiation flux of a 304 stainless steel sheet. The helical pipe is also made of 304
distribution and 3D numerical simulation and revealing that the stainless steel. Although the stainless steel pipe has lower thermal
temperature distributions of an inner receiver surface are signifi- conductivity than a copper pipe, excellent manufacture ensures a
cantly influenced by inhomogeneous radiation flux. Qiu et al. [17] well-formed helical pipe without any folded regions or cracks. The
also used the MCRT to obtain the radiation flux distribution of the focused point is at the aperture of the cavity, as shown in Fig. 1 (b).
tube’s external wall in order to build a comprehensive simulation The maximum diameter (dcav) of the cavity is 460 mm. The
model coupling the optical and heat transfer processes. The results diameter (dap) of the aperture is 200 mm. These sizes are consid-
indicated that thermal efficiency was improved 8% by decreasing ered to the actual receiver installing in the experiment system
 Y. Zhang et al. / Energy 193 (2020) 116775 3

(a) Cross-section (b) External view

Fig. 1. Schematics of the conical receiver.

([14,19]). The conical angle (b) and the length (L) of the cavity, as optical simulation in this study. Adopting the Monte Carlo ray
well as the thickness (s) of the thermal insulation layer, vary. For tracing method, users of TracePro can trace random rays emitted
this model, the loop number of the helical pipe (n) is determined by from a light source, as well as simulate the position and direction of
the cavity length. The diameter (dt) of the helical tube is 42 mm and the random rays and the reflection, refraction, and scattering pro-
the wall thickness (dt) is 3 mm, according to the size of the actual cesses that occur during the propagation. Each ray is traced and
helical pipe installing in the experiment system. Considering counted until its energy decreases to a negligible level or the ray
manufacturing errors and thermal stress deformation, a 2 mm gap escapes the system [24].
is between each pair of loops. The energy of each ray reflected by the parabolic dish is deter-
The experimental investigation of Prakash et al. [20] showed mined as Eq. (1):
that a decrease in the curvature leads to a sharp increase in
manufacturing stress and residual stress, which increases pR2 I0 r
manufacturing difficulty. To avoid an overly small curvature radius, E0 ¼ (1)
N
the upper limit of the conical angle is set to 20
in this study. The
conical angle varies from 0
to 20
, the loop number is 8, and the where R is the light source radius and r is the dish reflectivity. The
insulation layer is a constant 75 mm. When the loop number is reflection model is based on the bidirectional reflection distribu-
smaller than 3, the heating surface is insufficient for the working tion function (BRDF). When the number of a ray reflection time in
fluid in the pipe [20]. Therefore, the lower limit of the loop number the cavity is m, the ray energy is calculated as Eq. (2):
is set to 4. The loop number varies from 4 to 15 with a conical angle
of 5
and insulation layer at a constant 75 mm. The lower limit of Em ¼ E0 ð1  aÞm m ¼ 0; 1; 2; 3; … mmax (2)
insulation thickness is set to 25 mm, primarily to prevent excess
conduction loss, and the upper limit is set to avoid excess weight where a is the pipe absorptivity and also the cover absorptivity,
and volume of the cavity [21]. The insulation layer varies from mmax is the maximum number of reflections time. The energy on a
25 mm to 200 mm with conical angles in increments of 5
and the certain area of the receiver is calculated as Eq. (3).
constant loop number of 13. In an extreme situation, if the conical
angle is so large that cause an extrusion of pipe space in the cavity, X
mmax

it would affect the loops number. To avoid this situation, the value
Qn ¼ a Em Nm;n (3)
m¼0
range is set as Table 1 stated below, to assure that these three pa-
rameters are independent in this study. It is also because the pipe where Nm,n is the number of rays reaching area n and the number of
diameter is fixed, and increasing the loops number brings an in- reflections is m. The flux density in area n is calculated as Eq. (4).
crease of cavity height, so no corresponding effects between conical
angle and loops number. Qn
In ¼ (4)
Sn
2.2. Optical simulation
where Sn is the surface of the corresponding part.
The commercial software TracePro 7.3.4, which has been proven The parabolic dish model is established based on an experi-
reliable for optical analysis in published papers ([22,23]), is used for mental platform designed for preliminary research, as illustrated in
Fig. 2. A virtual disc is identified as the light source plane emitting
sun beams with a radius of 2.5 m. The parabolic dish diameter is
Table 1
Value range of geometric parameters.
5.06 m, while the focal length is 3.2 m. It should be noticed that
light source plane is a little smaller than the dish. Because the
Geometric parameter Value range
simulated parabolic dish is set according to the real dish of exper-
Conical angle 0
e20
iment system in 1:1 scale. In reality the rim of the dish has no silver
Loop number 4e15 plated reflective coating. Thus, this rim, around 3 cm, could not be
Insulation thickness 25 mme200 mm
considered as reflection area. The effective reflection area in the
4 Y. Zhang et al. / Energy 193 (2020) 116775

of parabolic dish is set to 0.9. The inner wall of the receiver has a
reflectance of 0.15 and an absorptivity of 0.85. The light flux
threshold is 0.001. It is a ratio of remaining energy carried by light
flux to initial energy entering the cavity carried by light flux,
meaning that below this threshold, light rays are neglected [24].

2.3. Numerical simulation based on combined optics and the heat-

transfer model

A flow chart illustrating the combined model simulation is

(a) Virtual light source and parabolic dish (b) Actual dish shown in Fig. 3. The steps of the modeling process are listed below.
The heat flux distributions of all the cases are obtained via the
Fig. 2. Parabolic dish model based on actual dish. optical simulation. This includes various conical angles, loop
numbers, and insulation thicknesses. For instance, a non-uniform
distribution is shown on the surface of the helical pipe of a cavity
simulation is as same large as the light source plane actually, which receiver with 8 loops, as illustrated in Fig. 4. The heat flux on each
is 5 m diameter. loop is different from those of the others. The deep loop generally
Following the MCRT method, discrete ray propagation, rather obtained less heat flux than the outer loop. For the accuracy of this
than continuous ray prorogation, is applied for simulating actual simulation, uniform heat flux is not assumed. The non-uniform
conditions. Discrete rays are randomly sampled based on proba- heat flux distributions are used as the boundary conditions for
bility density. On one hand, increasing the number of tracing rays numerical simulation to integrate the heat transfer process of the
helps enhance the accuracy of the simulation results; on the other conical receiver.
hand, an excessive number of tracing rays leads to unnecessarily A novel method is proposed in this step for a precise simulation
high time costs. Comparing to the reference by S. Zhu et al. [25] and a reduction of time costs. Since the unexposed side that is
using 119401 rays, this model had a much complicated structure, so attached to the insulation layer obtains no rays, this side is not
the rays was initially selected 10 times as much, which was 106. considered. However, the exposed side presented very different
Then, this study simulated the optics model from 106 rays to 2, 106 heat flux distributions, which is very non-uniform. Even one loop of
rays, the results revealed that heat flux distribution on every sur- the exposed side is also non-uniform, as shown in Fig. 5 (a). The
faces represented tiny differences between 106 rays and 2, 106 area on a loop exposed to solar beams absorbed most of the rays; it
rays. Therefore, the number of tracing rays in this optical simulation is thus considered to be the absorbing ray area. In the view of cross
is set to 106. section shown in Fig. 5 (b), the absorbing area occupies a quarter of
To obtain the heat flux distribution inside the conical receiver, the section ring. This quarter absorbing ray area is then used as the
the following assumptions are made during the optical simulation: boundary condition for heat transfer simulation. The other three
(1) The parabolic dish is an ideal paraboloid, and (2) The reflection quarters absorbed almost no rays; they are thus considered to be
mode of rays is diffuse reflection inside the cavity, since the surface no-ray areas. Normally, it is only the absorbing ray area that absorbs
of helical pipe would be rough after high temperature oxidation all the solar energy on the helical tube and heats the working fluid.
and dust accumulation. In this way, the optics and heat transfer processes of the cavity
The simulation procedure is performed with following sets. The receiver are combined.
direct normal irradiance (DNI) is set to 1000 W/m2. The reflectivity A comparison is proposed to illustrate this novel method could

Fig. 3. Flow chart of combined model simulation.

 Y. Zhang et al. / Energy 193 (2020) 116775 5

Table 2
The comparison of time-costs and the calculation results.

t1 (minutes) t2 (minutes) hov;1 hov;2

Case 1 178 299 63.61% 63.62%
Case 2 196 345 66.24% 66.21%
Case 3 215 371 69.53% 69.55%

(a) (b)

Fig. 6. (a) Computational grid for numerical analysis of the conical receiver; (b) cubical
enclosure around the cavity.

Fig. 4. Heat flux distribution on the surface of a helical pipe with 8 loops.

atmosphere with a temperature equal to that of the ambient air.

The zone outside the receiver was a cubical enclosure that had been
reduce the time-cost without any damage on the calculation pre- expanded to a relatively large size. To save time, fine meshes were
cision. As shown in Table 2, time-costs and simulation results both used inside the receiver, while outside the receiver coarse meshes
of with using novel method and without using novel method were were used. This allows for a good compromise between time costs
listed. Three cases were randomly selected in the present study to and accuracy. The mesh generation was built using ICEM. A
be the representation. pressure-based solver was applied. It assumes perfect gas behavior.
Case 1 is the cavity with 5
conical angle, 8 loops and 75 mm All temperature-dependent air propertiesdconductivity, viscosity,
insulation thickness, Case 2 is the cavity with 7.5
conical angle, 9 and heat capacitydwere approximated with polynomials with 4

loops and 100 mm insulation thickness, Case 3 is the cavity with 10
and a deviation less than 0.1%. For pressure-velocity coupling, the
conical angle, 11 loops and 150 mm insulation thickness. t1 is the SIMPLE scheme was used. Pressure discretization employed the
time cost with the novel method, t2 is the time cost without the PRESTO! scheme. The momentum and energy equation dis-
method, hov;1 is the calculation result of overall efficiency with the cretization were the second order upwind type.
novel method, hov;2 is the calculation result of overall efficiency The outer domain was treated as a pressure outlet boundary
without the method. condition. The pipe outside the cavity was encased in insulation.
It could be seen from Table 2 that the novel method reduced the The ambient temperature was fixed at 300 K. The inlet temperature
time cost apparently, almost the half of the cost without using it. of the working fluid air is 623.15 K. The pressure is 0.45 MPa. The
The calculation results of overall efficiencies presented that no mass flow rate is 0.03 kg/s. These values are derived from previous
obvious differences were existed between two different boundary studies [26]. The emissivity values for the pipe and the cavity’s
condition setting methods. Meanwhile, this method would not only inner wall are set to 0.85. The ambient wall was set as the pressure
be limited to this system, but also could be applied to a similar inlet. A convergence criterion of 103 was considered for the re-
cavity receiver with built-in helical pipe. siduals of continuity. A Boussniesq model was applied. Turbulence
The computational grids of the conical receiver are schemati- was calculated using the k-e model. The S-2-S model accounted for
cally represented in Fig. 6. The receiver is surrounded by an infinite the radiation exchange. As for the grids checking, the determinant

(a) (b)
Fig. 5. Heat flux distribution on one loop of the helical pipe.
6 Y. Zhang et al. / Energy 193 (2020) 116775

results of ICEM presented that the quality value of meshes is 3.2. Influence of conical angle
ranging from 0.533612 to 1, which is qualified for the numerical
simulation. This section analyzes the influence of the conical angle on
thermal performance. Conical angles ranging from 0
to 20
were
considered. In this part, the loop number was 8 and the insulation
2.4. Calculation of efficiencies layer was 75 mm. The parameters that were not related to the
variation in the conical angle were fixed. Heat losses and effi-
As defined by Li et al. [27], the optical efficiency hopt is calculated ciencies were solved in the simulation with the assistance of CFD
as Eq. (5). software FLUENT®. The tilt angle of the cavity is 45
, which is equal
to the cases in our previous studies [14], since the same value of tilt
Qabs angle would be strict for the comparison.
hopt ¼ (5)
Qtot As can be seen from Fig. 7, the influence of increasing the conical
angle can be observed. Five conical anglesd0
, 5
, 10
, 15
, and
where Qabs is absorbed energy in the cavity and Qtot is the total solar 20
dwere selected for demonstration purposes. As the conical
energy casting on the mirror. angle increased from 0
to 20
, high-temperature areas, as well as
The thermal efficiency hth is determined by Eq. (6). the surface of the stagnant zone areas, became closed to the upper
side of the helical pipe surface. In the 5
case, the average cavity air
Qabs  Qcd  Qcv  Qrad Qabs  Qloss
hth ¼ ¼ (6) temperature was more than 1100 K, whereas it was 1000 K for the
Qabs Qabs 20
case. High-temperature air appeared near the helical pipe in
each case; this is because the helical pipe absorbed most of the
where Qcd is conduction losses, Qcv is convection losses, Qrad is
solar irradiation. A comparison of the five images in Fig. 7 in terms
radiation losses, Qloss is total heat losses which is also the sum of
of temperature contours shows that the blue area slightly increases
conduction losses, convection losses and radiation losses.
with the conical angle. It also shows that cold-temperature air
Overall efficiency hov is determined by Eq. (7).
penetrates the inner zone of the conical cavity more and more
hov ¼ hopt :hth (7) deeply.
Fig. 8 shows the results of this influence on three types of heat
losses. In the examined angle range, total heat losses had a ten-
dency to decrease and then increase. Convection losses had the
same tendency as total heat losses, decreasing and then increasing.
3. Results and discussion The minimum of convection losses was 1684 W. Radiation losses
gradually increased from 2370 W to 2773 W, while conduction
3.1. Model validation losses gradually decreased from 530 W to 349 W. The conical angle
thus affected the thermal performance, as shown in the varying
Overall efficiencies and output power in different operation levels of convection losses.
conditions are able to describe the optical-thermal processes in a According to the Nusselt number definition, the Nusselt number
cavity. These processes have been used by some researchers [28,29] increases with the Grashof number, and the Grashof number is
as a means of verification. Undertaking an experimental investi- proportional to the third power of the characteristic length. The
gation for the thermal analysis of a conical receiver, using research increase in the conical angle influenced the characteristic length. In
methods different from that of the present study, was proposed by this cubic function, convection losses first decreased and then
Chu et al. [30]. increased. Due to the low thermal conductivity of the insulation
After applying the same operation conditions, the numerical layer, the conduction losses were relatively less than the other
results were compared to published data obtained with the types of heat losses. With the increase in the conical angle, the
experimental investigation. This validation was focused on overall surface of the conduction heat transfer decreased, leading to
efficiencies and the output power of the receiver. Table 3 compares reduced conduction losses.
the results of Chu et al. [30] and the present study for two cavities The thermal efficiency, optical efficiency, and overall efficiency,
that are the same size across 8 different operation conditions. Good are illustrated in Fig. 9. As the conical angle increased, the optical
agreement with the experiment results was observed, with an efficiency hopt decreased from 87.6% to 86.6%. On the other hand,
average deviation of about 8.08%. The values of the numerical both the thermal efficiency hth and overall efficiency hov first
model’s results were larger than the experiment results in all cases, increased and then decreased with the increase in the conical
but the differential values were small. Thus, it can be concluded angle. It should be noted that the maximum overall efficiency and
that this study offers credible numerical simulation results. thermal efficiency, which were 63.6% and 72.7%, respectively,
occurred with the 5
conical angle.
The slow decrease in optical efficiency was related to decreased
Table 3 absorption. As the conical angle increased, the cavity bottom sur-
Comparison of overall efficiencies and output powers.
face decreased, which led to a decrease in the surface available for
Overall efficiency Overall efficiency Output power of Output power of absorption. Eq. (6) shows that thermal efficiency is related to total
of Chu et al. [28] of present study Chu et al. [28] present study heat losses. The absorption remained nearly unchanged, the vari-
(%) (%) (kW) (kW)
ation tendency of thermal efficiency was opposite that of total heat
40.74 45.52 3.90 4.36 losses. As Eq. (7) indicates, overall efficiency is the production of
43.04 49.91 3.01 3.53
optical efficiency and thermal efficiency; therefore, optical effi-
53.16 55.08 3.96 4.11
47.23 51.16 2.45 2.65 ciency, like thermal efficiency, has a tendency to first increase and
55.72 59.94 1.27 1.39 then decrease. This analysis enables the discovery of the optimal
47.07 50.60 2.45 2.63 value of overall efficiency in the examined range, as well as a
52.06 57.21 2.09 2.26 method of optimization for the thermal performance of the conical
38.30 43.55 2.95 3.36
cavity receiver.
 Y. Zhang et al. / Energy 193 (2020) 116775 7

Fig. 7. Temperature contours of a conical receiver with 8 loops and 75 mm insulation thickness at five different conical angles.

3.3. Influence of loop number

This section analyzes the influence of loop number on thermal

performance. Loop numbers ranging from 4 to 15 were considered.
In this part, the conical angle was 5
and the insulation layer was
75 mm. In Fig. 10, the influence of increasing the loop number can
be seen. Five examplesdloop numbers of 5, 7, 10, 13, and 15dare
shown. As the loop number increased from 5 to 15, high-
temperature areas, as well as the surface of the stagnant zone
area, decreased. In the 5-loop case, the air temperature in the
stagnant zone was found to be around 1200 K, whereas it was
found to decrease to 1050 K for the 15-loop case. The maximum
temperature of cavity air is over 1280 K, which is found to be in the
stagnant zone of 5 loop case. The temperature contours of the five
examples. The blue areas inside the cavity are small and show no
obvious differences. It can thus be concluded that cold-temperature
air shallowly penetrates the inner zone of the conical cavity (see
Fig. 8. Influence of conical angle on heat losses. Fig. 11).
In the examined range, total heat losses had a tendency to first
decrease and then remain invariant. The loop number can affect
thermal performance within a certain range; however, increasing
the loop number beyond the range does not help to reduce total
heat losses. Convection losses decreased very slightly from 1945 W
to 1566 W. Radiation losses decreased comparatively sharply from
3377 W to 2304 W. Conduction losses gradually increased from
441 W to 607 W.
Radiation losses are predominant when considering varying the
loop number, since the Stefane Boltzmann law includes the fourth
power of temperature, and the temperature of a helical pipe surface
can be as high as approximately 1250 K. The law also explains the
sharp decrease in radiation losses when the loop number increases,
indicating that the temperature in the cavity apparently decreases
with the increase in the loop number.
And normally, when increasing the bumber of loop, T should
decrease. The increase in the loop number also increases the sur-
face available for conduction heat transfer, which ultimately leads
to increased conduction losses.
As illustrated in Fig. 12, as the loop number increased, the optical
efficiency remained nearly unchanged, increasing only from 87.1%
to 87.5%. The thermal efficiency and overall efficiency first
Fig. 9. Influence of conical angle on efficiencies. increased and then remained invariant. The maximum overall
8 Y. Zhang et al. / Energy 193 (2020) 116775

Fig. 10. Temperature contours of the conical receiver with 5
conical angle and 75 mm insulation thickness using five different loop numbers.

Fig. 11. Influence of loop number on three types of heat losses.

Fig. 12. Influence of loop number on efficiencies.

efficiency and thermal efficiency, which were 64.8% and 74.0%,
respectively, occurred with the 13-loop case.
The increase in optical efficiency is due to the increased ab- established in the previous sections. In Fig. 13, the influence of
sorption surface, and it is also the reason why a deep cavity is increasing the thickness of the insulation layer can be observed.
beneficial for preventing ray escape. Considering Eq. (6), the ab- Five thickness levelsd25 mm, 50 mm, 75 mm, 125 mm, and
sorption heat remains unchanged and the total heat losses 175 mmdwere chosen for demonstration purposes. In the 25 mm
decreased; the thermal efficiency increased with the increased of thickness case, the air temperature around the cavity’s aperture
loop number. Overall efficiency had a similar tendency, which can was found to be around 350 K, whereas it was found to increase to
be attributed to the relation between overall efficiency and thermal 550 K for the 125 mm case. As the thickness increased from 75 mm
efficiency shown in Eq. (7). It can be concluded that there are limits to 200 mm, high-temperature areas inside the cavity, as well as the
to increasing the loop number; too many loops increase production surface of the stagnant zone area, showed no obvious changes.
costs but do not contribute to enhanced thermal performance, and Fig. 13 shows the temperature contours of the five cases.
increases the pressure drop of HTF in the helical pipe. Unlike the trends associated with varying the loop number, total
heat losses kept decreasing as the insulation thickness increased,
shown in Fig. 14. This suggests that insulation thickness can
3.4. Influence of insulation thickness continually affect total heat losses. Convection losses very obvi-
ously decreased from 1972 W to 835 W. Similarly, conduction losses
This section analyzes the influence of the insulation layer on decreased from 1213 W to 340 W. Radiation losses decreased
thermal performance. Insulation layers ranging from 25 mm to relatively slowly from 2535 W to 1783 W. The conduction losses in
200 mm in thickness were considered. The lower limit 25 mm is set the 25 mm case were triple the losses of the 125 mm case, which
to prevent excess conduction loss, and the upper limit is set to avoid indicates that 25 mm thickness is insufficient.
excess weight and volume of the cavity. In this part, the conical As the insulation thickness increases, convection losses and
angle was 5
and the loop number was 13, which is the optimum conduction losses are dcreased, and the decreases are relatively
 Y. Zhang et al. / Energy 193 (2020) 116775 9

Fig. 13. Temperature contours of the conical receiver with 5
conical angle and 13 loops at five different insulation thicknesses.

obvious comparing to the decrease of radiation loess. This is

because the influence of insulation thickness on the thermal
resistence coefficient is very strong. Increasing the thickness leads
to an increase in the thermal resistence coefficient of the cavity,
which results in a decrease in convection losses and conduction
losses. Radiation losses related to the temperature of helical pipe
surface decreased slightly with increasing thickness.
As illustrated in Fig. 15, as the insulation thickness increased
from 25 mm to 150 mm, the optical efficiency decreased slowly,
varying from 87.6% to 86.8%. However, the optical efficiency sharply
decreased to less than 85.2% when the thickness increased to more
than 175 mm. The thermal efficiency kept increasing from 66.8% to
82.3%. The overall efficiency varied in ways similar to the optical
efficiency, which first increased and then decreased. Maximum
overall efficiency, which was 70.4%, occurred at 175 mm thickness.
The increase in insulation thickness led to an increase of blocked
sun beam, ultimately decreasing the obtained solar flux of the
receiver. When the thickness increased within a certain range, the
blocking effect does not appear, but when the thickness increased
beyond this range, the amount of blocked sun beam sharply Fig. 15. Influence of loop number on efficiencies.

increased. Thus, the optical efficiency sharply decreased. As shown

in Fig. 16, the yellow area and the blue area on the aperture surface
Thermal efficiency had a tendency opposite to that of heat losses
represented the heat flux of blocked sun beams.
when the thickness was varied. The overall efficiency is the result of
thermal efficiency multiplied by optical efficiency, since the ten-
dency was to at first increase and then decrease. The results suggest
an optimal insulation thickness value in the range of
25 mme200 mm. According to the graph, optimum is 175 mm,
achieving a 70.4% overall efficiency.

4. Conclusion

In this study, a combined optics and heat transfer model was

constructed. The model simulation took into consideration the in-
fluence of geometric parameters on the performance of a conical
cavity receiver. Model validation was performed by comparing the
simulation results and the experiment results from other studies,
and good agreement was achieved. This study also proposed a
performance optimization method for a conical cavity receiver. A
summary of key findings follows below:

(1) When the conical angle increased from 0
to 20
, the optical
Fig. 14. Ilustrates the results of this influence on three types of heat losses. efficiency hopt decreased from 87.6% to 86.6%. On the other
10 Y. Zhang et al. / Energy 193 (2020) 116775

(a) 175 mm (b) 200 mm

Fig. 16. The insulation thickness influences the optical efficiency.

hand, both the thermal efficiency hth and overall efficiency loss in cavity-type solar receivers. Sol 2002-Austrailian New Zeal Sol Energy
Soc 2002;133:6. https://doi.org/10.1115/1.4003582.
hov first increased and then decreased with the increase in
[2] Reddy KS, Sendhil Kumar N. Combined laminar natural convection and surface
the conical angle. The maximum overall efficiency occurred radiation heat transfer in a modified cavity receiver of solar parabolic dish. Int
with a 5
conical angle. J Therm Sci 2008;47:1647e57. https://doi.org/10.1016/
(2) When the loop number was increased from 4 to 15, there was j.ijthermalsci.2007.12.001.
[3] Ngo LC, Bello-Ochende T, Meyer JP. Numerical modelling and optimisation of
only a slight decrease in optical efficiency from 87.5% to natural convection heat loss suppression in a solar cavity receiver with plate
87.1%. The increases in thermal efficiency and overall effi- fins. Renew Energy 2015;74:95e105. https://doi.org/10.1016/
ciency became extremely small after 8 loops. It can be j.renene.2014.07.047.
[4] Shirvan KM, Mamourian M, Mirzakhanlari S, Ellahi R, Vafai K. International
concluded that there are limits to increasing the loop num- Journal of Heat and Mass Transfer Numerical investigation and sensitivity
ber. Too many loops would be an inefficient use of resources analysis of effective parameters on combined heat transfer performance in a
and does not contribute to enhanced performance. porous solar cavity receiver by response surface methodology. Int J Heat Mass
Transf 2017;105:811e25. https://doi.org/10.1016/
(3) As with the increase in the conical angle, the increase in j.ijheatmasstransfer.2016.10.008.
insulation thickness from 25 mm to 200 mm caused thermal [5] Wu SY, Xiao L, Cao Y, Li YR. Convection heat loss from cavity receiver in
efficiency hth and overall efficiency to first increase and then parabolic dish solar thermal power system: a review. Sol Energy 2010;84:
1342e55. https://doi.org/10.1016/j.solener.2010.04.008.
decrease. However, increased insulation thickness led to a
[6] Loni R, Askari Asli-Ardeh E, Ghobadian B, Kasaeian AB, Bellos E. Thermal
consistent decrease in optical efficiency. The results indicate performance comparison between Al2O3/oil and SiO2/oil nanofluids in cy-
that 175 mm is the optimal value for insulation thickness, lindrical cavity receiver based on experimental study. Renew Energy
2018;129:652e65. https://doi.org/10.1016/j.renene.2018.06.029.
which represents a compromise between the decrease in
[7] Przenzak E, Szubel M, Filipowicz M. The numerical model of the high tem-
optical efficiency and the increase in thermal efficiency. perature receiver for concentrated solar radiation. Energy Convers Manag
2016;125:97e106. https://doi.org/10.1016/j.enconman.2016.07.036.
This study proposed a new methodology for optimizing the [8] Asselineau CA, Zapata J, Pye J. Geometrical shape optimization of a cavity
receiver using coupled radiative and hydrodynamic modeling. Energy Proce-
performance of conical receivers; it did not consider a specific dia 2015;69:279e88. https://doi.org/10.1016/j.egypro.2015.03.032.
conical angle, loop number, or insulation thickness. The findings [9] Tu N, Wei J, Fang J. Numerical investigation on uniformity of heat flux for
could be applied to various situations. If certain environment pa- semi-gray surfaces inside a solar cavity receiver. Sol Energy 2015;112:
128e43. https://doi.org/10.1016/j.solener.2014.11.028.
rameters are changed, the optics and heat transfer behavior of a [10] Prakash M, Kedare SB, Nayak JK. Investigations on heat losses from a solar
receiver will shift and a new optimum will be found. However, the cavity receiver. Sol Energy 2009;83:157e70. https://doi.org/10.1016/
performance optimization method remains valid and applicable to j.solener.2008.07.011.
[11] Gil R, Monne
C, Bernal N, Mun ~ oz M, Moreno F. Thermal model of a dish
different cases. stirling cavity-receiver. Energies 2015;8:1042e57. https://doi.org/10.3390/
Future work will include manufacturing and testing the opti- en8021042.
mized conical receiver to verify the results of this study. Receivers [12] Hussain T, Islam MD, Kubo I, Watanabe T. Study of heat transfer through a
cavity receiver for a solar powered advanced Stirling engine generator. Appl
will be installed within the solar-dish platform for experimental
Therm Eng 2016;104:751e7. https://doi.org/10.1016/
investigation. For a complete CSP technology study, an analysis of j.applthermaleng.2016.05.108.
some operation parameters must also be completed. [13] Daabo AM, Mahmoud S, Al-Dadah RK. The effect of receiver geometry on the
optical performance of a small-scale solar cavity receiver for parabolic dish
applications. Energy 2016;114:513e25. https://doi.org/10.1016/
Acknowledgement j.energy.2016.08.025.
[14] Zou C, Zhang Y, Falcoz Q, Neveu P, Zhang C, Shu W, et al. Design and opti-
mization of a high-temperature cavity receiver for a solar energy cascade
This research is financially supported by Major Technology utilization system. Renew Energy 2017;103:478e89. https://doi.org/10.1016/
Innovation of Hubei Province (2019AAA017), National Youth j.renene.2016.11.044.
[15] Li X, Dai YJ, Wang RZ. Performance investigation on solar thermal conversion
Foundation of China (51906080) and China Postdoctoral Science
of a conical cavity receiver employing a beam-down solar tower concentrator.
Foundation (2019M662620). Sol Energy 2015;114:134e51. https://doi.org/10.1016/j.solener.2015.01.033.
[16] Cui F-Q, He Y-L, Cheng Z-D, Li Y-S. Modeling of the dish receiver with the
effect of inhomogeneous radiation flux distribution. Heat Transf Eng 2014;35:
References 780e90. https://doi.org/10.1080/08832323.2013.838098.
[17] Qiu K, Yan L, Ni M, Wang C, Xiao G, Luo Z, et al. Simulation and experimental
[1] Paitoonsurikarn S, Lovegrove K. Numerical investigation of natural convection study of an air tube-cavity solar receiver. Energy Convers Manag 2015;103:
 Y. Zhang et al. / Energy 193 (2020) 116775 11

847e58. https://doi.org/10.1016/j.enconman.2015.07.013. [24] Jacobsen D. Making custom optical properties in TracePro. 2017.
[18] Azzouzi D, Boumeddane B, Abene A. Experimental and analytical thermal [25] Si-Quan Z, Xin-Feng L, Liu D, Qing-Song M. A numerical study on optical and
analysis of cylindrical cavity receiver for solar dish. Renew Energy 2017;106: thermodynamic characteristics of a spherical cavity receiver. Appl Therm Eng
111e21. https://doi.org/10.1016/j.renene.2016.12.102. 2019;149:11e21. https://doi.org/10.1016/j.applthermaleng.2018.10.030.
[19] Zou C, Zhang Y, Feng H, Falcoz Q, Neveu P, Gao W, et al. Effects of geometric [26] Zou C, Zhang Y, Falcoz Q, Neveu P, Zhang C, Shu W, et al. Design and opti-
parameters on thermal performance for a cylindrical solar receiver using a 3D mization of a high-temperature cavity receiver for a solar energy cascade
numerical model. Energy Convers Manag 2017;149:293e302. https://doi.org/ utilization system. Renew Energy 2017;103:478e89.
10.1016/j.enconman.2017.06.088. [27] Li S, Xu G, Luo X, Quan Y, Ge Y. Optical performance of a solar dish concen-
[20] Prakash M, Kedare SB, Nayak JK. Numerical study of natural convection loss trator/receiver system: influence of geometrical and surface properties of
from open cavities. Int J Therm Sci 2012;51:23e30. https://doi.org/10.1016/ cavity receiver. Energy 2016;113:95e107. https://doi.org/10.1016/
J.IJTHERMALSCI.2011.08.012. j.energy.2016.06.143.
[21] Harris JA, Lenz TG. Thermal performance of solar concentrator/cavity receiver [28] Xiao G, Yang T, Ni D, Cen K, Ni M. A model-based approach for optical per-
systems. Sol Energy 1985;34:135e42. https://doi.org/10.1016/0038-092X(85) formance assessment and optimization of a solar dish. Renew Energy
90170-7. 2016;100:103e13. https://doi.org/10.1016/j.renene.2016.05.076.
[22] Xie WT, Dai YJ, Wang RZ. Numerical and experimental analysis of a point [29] Fang JB, Tu N, Wei JJ. Numerical investigation of start-up performance of a
focus solar collector using high concentration imaging PMMA Fresnel lens. solar cavity receiver. Renew Energy 2013;53:35e42. https://doi.org/10.1016/
Energy Convers Manag 2011;52:2417e26. https://doi.org/10.1016/ j.renene.2012.10.053.
j.enconman.2010.12.048. [30] Chu S, Bai F, Zhang X, Yang B, Cui Z, Nie F. Experimental study and thermal
[23] Wang J, Yang S, Jiang C, Yan Q, Lund PD. A novel 2-stage dish concentrator analysis of a tubular pressurized air receiver. Renew Energy 2018;125:
with improved optical performance for concentrating solar power plants. 413e24. https://doi.org/10.1016/j.renene.2018.02.125.
Renew Energy 2017;108:92e7. https://doi.org/10.1016/j.renene.2017.02.059.